SAND2026-16588 SANDIA REPORT Unlimited Release Printed January 28, 2026 SIERRA Code Coupling Module: Arpeggio User Manual – Version 5.28 SIERRA Thermal/Fluid Development Team Prepared by Sandia National Laboratories Albuquerque, New Mexico 87185 Livermore, California 94550 Issued by Sandia National Laboratories, operated for the United States Department of Energy by National Technology & Engineering Solutions of Sandia, LLC. NOTICE: This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government, nor any agency thereof, nor any of their employees, nor any of their contractors, subcontractors, or their employees, make any warranty, express or implied, or assume any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represent that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government, any agency thereof, or any of their contractors or subcontractors. The views and opinions expressed herein do not necessarily state or reflect those of the United States Government, any agency thereof, or any of their contractors. Printed in the United States of America. This report has been reproduced directly from the best available copy. Available to DOE and DOE contractors from U.S. Department of Energy Office of Scientific and Technical Information P.O. Box 62 Oak Ridge, TN 37831 Telephone: (865) 576-8401 Facsimile: (865) 576-5728 E-Mail: reports@osti.gov Online ordering: http://www.osti.gov/scitech Available to the public from U.S. Department of Commerce National Technical Information Service 5301 Shawnee Road Alexandria, VA 22312 Telephone: (800) 553-6847 Facsimile: (703) 605-6900 E-Mail: orders@ntis.gov Online order: https://classic.ntis.gov/help/order-methods NT OF E ME N RT ER A DEP GY • • IC A U NIT ER ED ST M A TES OF A 2 ABSTRACT The SNL Sierra Mechanics code suite is designed to enable simulation of complex multiphysics scenarios. The code suite is composed of several specialized applications which can operate either in standalone mode or coupled with each other. Arpeggio is a supported utility that enables loose coupling of the various Sierra Mechanics applications by providing access to Framework services that facilitate the coupling. More importantly Arpeggio orchestrates the execution of applications that participate in the coupling. This document describes the various components of Arpeggio and their operability. The intent of the document is to provide a fast path for analysts interested in coupled applications via simple examples of its usage. 3 4 CONTENTS Contents 5 List of Figures 8 1. Introduction 9 1.1. Coupled Physics Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 1.2. Sierra Mechanics Coupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 1.3. Communication of Data (Transfer Services) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 1.4. Solution Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 1.5. Coupling Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 1.6. Coupling with Arpeggio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 1.7. Outline of the Manual . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2. Getting Started 17 2.1. Setting The Environment-Users External to Sandia Labs . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.2. Setting The Environment-Users at Sandia Labs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.3. Running Arpeggio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.4. Arpeggio Environment Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 2.5. Overview of the Input File Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 2.6. Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 2.7. User Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 3. Model Definition 27 3.1. Model Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 3.2. Finite Element Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 3.3. Parameters For Block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 5 3.4. Global Constants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 3.5. Definition For Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 3.6. Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 3.7. Restart Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 4. Solution Control Reference 59 4.1. Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 4.2. Solution Control Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 4.3. System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 4.4. Transient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 4.5. Nonlinear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 4.6. Subcycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 4.7. Sequential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 4.8. Initialize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 4.9. Parameters For . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 5. Transfer Reference 91 5.1. Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 5.2. Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 6. Input Output Region Reference 111 6.1. Input_Output Region Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 6.2. Input_Output Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 7. Examples 119 7.1. One-Way Coupling From File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 7.2. One-Way Coupling Using Transfer From Different Mesh . . . . . . . . . . . . . . . . . . . . . . . . . 124 7.3. One-Way Coupling Using Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 7.4. Two-Way Coupling With Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 7.5. Two-Way Coupling Using Transfer of Element Death . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 7.6. estack Regression Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 6 7.7. tv Regression Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166 Bibliography 169 Index 171 Distribution 174 7 LIST OF FIGURES 1.1-1. Loose Coupling Schematic (Z Scheme). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 1.3-1. Sierra Mechanics Data Types. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 1.5-1. One-way Loose Coupling At Same Time Step. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 1.5-2. Deferred One-way Loose Coupling At Same Time Step. . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 1.5-3. One-way Loose Coupling with Subcycling Schematic. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 1.5-4. Two-way Loose Coupling Schematic. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 1.6-1. Thermal-Mechanical With Thermal One-Way Element Death. . . . . . . . . . . . . . . . . . . . . . . 16 1.6-2. Thermal-Mechanical With Two-Way Element Death. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2.4-1. Schematic UML class diagram for the Expression subsystem. . . . . . . . . . . . . . . . . . . . . . . . 19 5.1-1. Valid Transfer Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 5.1-2. Invalid Transfer Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 8 1. INTRODUCTION The SNL Sierra Mechanics code suite is designed to enable numerical simulations of complex multi-physics scenarios. The code suite is composed of specialized applications which can operate either in standalone mode or in a coupled mode with other Sierra Mechanics applications. Arpeggio is a supported utility that enables loose coupling of the various Sierra Mechanics applications by providing access to Framework services that facilitate application coupling. Utilizing these services Arpeggio is able orchestrate the execution of applications that participate in code coupling. This document describes the Framework services used by Arpeggio for coupling and the inter-operability of these services for coupling of Sierra SM and Sierra TF applications. Through the use of simple examples, the document also provides a resource for analysts interested performing in coupled-physics simulations. 1.1. COUPLED PHYSICS APPROACHES When modelling tightly-coupled physics, the numerical representation of all PDEs within a region of interest are often combined a single system matrix and solved using a nonlinear solution strategy specific to the application. This approach to solving coupled-physics problems is available for a limited set of physics in the Sierra Mechanics TF module. Relaxing the notion of tight-coupling one could alternatively obtain solutions for each of the physics independently and patch the individual solutions together in some prescribed manner, this is the essence of loosely-coupled physics simulations. The numerical analysis community has long recognized the need to include results from various physics in a single simulation. However, the fact that most application codes are often developed around single physics often limits the extent to which coupled-physics simulations can be achieved. Early approaches to coupled-physics simulations often simplified the coupling by level of importance by assigning primary physics and secondary physics roles. Here the primary physics depended upon secondary physics and the dependence of secondary physics upon primary physics was deemed less important. Under this assumption coupled physics simulations can be realized by first performing independent simulations of the secondary physics followed by a simulation of the primary physics utilizing results of the initial simulation. Figure 1.1-1 illustrates the coupling approach for a quasi-static solution step from a state tn to state tn+1 . Broadly speaking, loose-coupling strategies are classified as Z-methods, since a Z describes the basic pattern of data communication between the physics applications. The one-way view of loosely-coupled physics lends itself to file-based approaches where single state results are obtained on a common spatially meshed discretization. Here the problem solutions are generally obtained at cell vertices (nodes) or cell centers (elements). Quite often each physics simulation lends itself to a particular spatial discretization and this gave rise to the introduction of an intermediate mapping step whereby the secondary physics results were mapped onto the primary physics discretization as in the MAPVAR utility [1]. For transient coupled-physics simulations best results are obtained when sharing a common 9 tn t n+1 ∆t Secondary Physics Primary Physics Figure 1.1-1.. Loose Coupling Schematic (Z Scheme). time discretization but in many cases this is impractical and the coupling is based upon closest-time matched solutions or interpolations of solutions in time. 1.2. SIERRA MECHANICS COUPLING Sierra Mechanics physics applications deal with solving PDEs on a physical geometric domain, a Region. In defining a coupled physics problem, users configure one or more Regions corresponding to some particular physics. Each Region considers one or more PDEs to be solved on either the entire input mesh or on a portion of the mesh. When the Region physics are coupled one can elect whether to solve the physics in a tightly-coupled manner in a single application or by loosely coupling individual Region results. Here we note that for loose coupling the physical geometry and spatial discretization must overlap but need not be identical in each of the participating Regions. In the context of Sierra Mechanics, loosely-coupled physics nonlinear solutions are obtained on each of the Regions and then combined to form an overall coupled solution. Not surprisingly there are numerous ways one can approach loose-coupling since different strategies are appropriate to different problem sets. That is, the solution for one Region may depend strongly upon the solution in another Region but not vice-versa (one-way coupled), or the the solution for each Region may depend upon the solution the other Region (two-way coupling). The goal of Sierra Mechanics is to provide services which enable one to easily perform variants of a multi-physics coupling. Some considerations which are relevant to the loose-coupling solution strategies include • Communication of data from one Region to another Region (Transfer), • Initialization of the individual Regions, • Solution for the individual Regions (Advance), • Time stepping or pseudo-time stepping for the individual Regions, • Time synchronization of participating Regions, • Conditional convergence, 10 • Drive mesh adaptivity, • Sequencing for all of the above. Within Sierra Mechanics communication of data between application Regions is handled by the Framework Transfer service and all aspects of solution behavior are managed by another Framework service, Solution Control. Mesh Adaptivity is managed through the Percept library. 1.3. COMMUNICATION OF DATA (TRANSFER SERVICES) In Sierra Mechanics application data is generally associated with nodes, elements, faces or edges of a meshed discretization as shown in Figure 1.3-1. A loose-coupling between applications implies the dependence of one application on data supplied from some external source. Since the physical location of data on the external source may or may not map geometrically onto the the other application solution, provisions must be made to perform this data mapping in a flexible manner. It is important to note that these mappings can be accomplished both for the case of different mesh and different element types. Within Sierra Mechanics this responsibility is handled by Framework Transfer services. Here it is important to note that Framework Transfer services enable the external source data to include element, face and edge data as well as nodal data. Node Element Face Edge Figure 1.3-1.. Sierra Mechanics Data Types. 1.4. SOLUTION CONTROL The Solution Control subsystem controls the execution of coupled multi-physics applications. Solution control provides two basic operations for controlling the solution of a multi-physics system by defining the order for object execution and by setting parametric values on the controlled objects at the proper time. For transient problems this approach enables the applications to easily transition through the designated time periods. The same system can also service steady-state simulations by treating them in a sequential manner. The solution controllers are able to initialize Region data, set parameter values, advance Regions, execute transfers, call events and send notifications based on the input file specifications. 11 1.4.1. Region Initialization When beginning execution, all applications require some baseline initialization operations at the Region level. When performing some loose-coupling simulations the dependence of data may may require that initialization of data be performed in some specific manner. Here the manner in which initialization occurs is determined by how the application solution variables are defined and the application code implementation of initialization. As an example for a thermal-mechanical coupling one might initialize the reference temperature state in the solid before any temperature change in the solid were allowed to occur. Solution Control provides a means for performing various types of non-standard data initialization. 1.4.2. Solution Each set of coupled physics represents a System of equations which must be solved. While participating in loose-coupling an application physics will attempt to advance its solution to a later state. In the parlance of Solution Control this step is known as an advance event. Here the details of code operations associated with advancing the solution are controlled entirely by the physics application. Additionally, because the advance can occur conditionally it provides flexibility in how the coupling is performed. 1.4.3. Time Stepping Within Sierra Mechanics each application is allowed to define its own notion of solution time. The Solution Control time step controller probes the individual application solution time and uses that information to determine how time should be advanced for the coupled physics. For couplings of transient simulations with quasi-static applications, the time step controller manages a unified notion of pseudo-time and physical time seamlessly, even when the time step selection is adaptive. 1.4.4. Conditional Events In loosely-coupled simulations the need often arises to perform some high level operations conditionally. Here Solution Control is able to probe the application for current states or variables to determine whether whether some coupling action should occur. These conditionals can be applied to both the data transfer or advance Solution Control events. Examples of conditional events are included in Chapter 4. 1.5. COUPLING STRATEGIES Using the Solution Control one can easily define loose couplings between two or more Regions. For example, some or all of a solution from one Region may be transferred to another Region where it is treated as a constant, external field. The aggregate nonlinear problem including the contributions from all of the Regions may be iterated to convergence. The details of which physics are solved in each 12 Region and the nonlinear solution strategy used within and between Regions is completely specified through the input file. Furthermore, a Sierra Mechanics user may pick a simple, minimal algorithm without needing to fit it into an overly-generalized worst-case scenario that represents the union of all possible algorithms. Dynamically-specified loose coupling has many potential advantages that users may leverage to obtain solutions. First, the resulting linear system is considerably smaller than a fully-coupled system and contains far fewer off-diagonal contributions which can significantly improve the performance of linear solvers. Furthermore the resulting linear system may have a more desirable mathematical properties, such as being symmetric positive-definite, this permits the use of tailored iterative solutions techniques. Other extensions to loose coupling include subcycling of transient simulations where each Region may advance in time with its own time step size and in-core coupling to other applications based upon the Sierra framework. The simplest loose-coupling strategy is a one-way coupling between two applications, App I and App II, is shown schematically in Figure 1.5-1. Here it is assumed that information (data) from App I is needed by App II but App I is independent (decoupled) from App II. Furthermore it is assumed that the applications can proceed at the same time step. In this case the solution for each application can proceed in locked step. tn t n+1 t n+2 t n+3 ∆t ∆t ∆t App I transfer App II Figure 1.5-1.. One-way Loose Coupling At Same Time Step. A variant of the simplest loose-coupling would be the case where the dependence of App II solutions on App I data is such that update of the App I data can be deferred for several steps. This type coupling behavior can be enforced using a conditional advance event in Solution Control. As an example, a data transfer event defined for every two time steps of each application is shown schematically in Figure 1.5-2. In some couplings the temporal response of one application physics, App I, is much faster than that of another physics, App II. Here one may wish to advance the App I physics many time steps before requiring an update of its contribution to the App II information, Figure 1.5-3. Here Solution Control provides a facility denoted as subcycling to invoke this behavior. When the coupling between App I and App II is circular in nature, (i.e. App I solutions depend upon App II and vice versa) the coupling can be achieved by adding an additional Transfer step to the one-way coupling approach. However, if the coupling dependency is fairly strong it may be prudent to ascertain a converged solution between the physics models before advancing to the solution step. Here the conditional event aspect of Solution Control can be employed to iterate App I and App II until a 13 tn t n+1 t n+2 t n+3 ∆t ∆t ∆t App I transfer App II Figure 1.5-2.. Deferred One-way Loose Coupling At Same Time Step. tn t n+1 t n+2 ∆ t t n+3 ∆t I ∆t I I App I transfer App II ∆ t II ∆ t II ∆ t II Figure 1.5-3.. One-way Loose Coupling with Subcycling Schematic. solution of the desired quality is obtained. The strategy is depicted in Figure 1.5-4 and is supported as the Nonlinear option within Solution Control. 14 tn t n+1 t n+2 t n+3 ∆t ∆t ∆t App I transfer App II Figure 1.5-4.. Two-way Loose Coupling Schematic. 1.6. COUPLING WITH ARPEGGIO While the previous sections have described the component utilities needed to enable coupled physics simulations but little has been said of existing tools composed of these utilities. Previous efforts in the development of Sierra Mechanics focused upon thermal-mechanical coupling of the Calore and Adagio applications with the Calagio utility to analyze problems of thermal stress. Here Sierra Mechanics utilities were used to solve the temperature state, then initializing the reference temperature state in Adagio followed by subsequent solves and transfer of the temperature state to Adagio to obtain a thermal stress state in the deformed configuration. Within the Calagio utility extra efforts were made to obscure the use of Framework utilities lying outside the realm of Calore and Adagio. Early one-way coupling efforts were later followed by two-way couplings where the deformed configuration was communicated to Calore and the heat transfer problem could be solved in the updated geometry. Although couplings with Calagio were largely successful it was recognized that incremental improvements in coupling capability came with a high price in terms of code development effort both to alter the predefined coupling strategies and to hide the underlying implementation from the analyst within the application code. Moreover, the predefined coupling strategy approach prevented the analyst from fully exploiting the resources available within Sierra Mechanics and the applications themselves. These shortcomings provided a motivation for creation of the Arpeggio utility in which the analyst fully specifies details of the coupling strategy. 1.6.1. Coupling Including Element Death Coupling strategies in predecessors of Arpeggio precluded the possibility of simulations that required syncronization of the meshed discretization such as element death. Here the transfer capability in conjunction with a consistent notion of an application code indicator of element death (Death_Status) enables coupled simulations that include element death. Prevalent uses of this capability are one-way coupled thermal-mechanical simulations with thermally-driven element death Figure 1.6-1 and two-way coupled thermal-mechanical simulations with element death driven by either application Figure 1.6-2. For both types of coupling the mechanical code behavior is essentially the same as for a two-way coupling. On the other hand, in the case of two-way coupling the one-way coupling 15 thermal invocation of a death criteria test is altered by the addition of an Aria Region level command line, Transfer Element Death, to trigger marking of dead elements upon transfer. Each application manages element death by applying/testing user defined death criteria to mark elements as dead with a death_status variable. Couplings including element death will involve transfer of a source application’s death_status variable to an intermediate field (user variable) on the target application. This intermediate field will then be used in evaluation of death criteria on the target application. Thermal Mechanical Jacobian Test Temperature Elem Death Test Solve Nonlinear − Death_Status Death_Status Elem Death Test Jacobian Test − Thermal Criteria Solve Nonlinear Figure 1.6-1.. Thermal-Mechanical With Thermal One-Way Ele- ment Death. Mechanical Thermal Temperature Elem Death Test Elem Death Test − Death_Status − Death_Status − Temperature Criteria Jacobian Test Death_Status Jacobian Test Solve Nonlinear Solve Nonlinear Displacement Elem Death Test − Structural Criteria Figure 1.6-2.. Thermal-Mechanical With Two-Way Element Death. 1.7. OUTLINE OF THE MANUAL Chapter 2 discusses the overall Sierra Mechanics environment for running Arpeggio, including the layout for the Arpeggio input file. Sierra Mechanics users familiar with the overall environment need only browse the input file structure and move directly to the sections describing Framework Transfer 5 and Solution Control 3.7. Experienced Sierra Mechanics users may opt to move directly to examples of coupling in Chapter 7 16 2. GETTING STARTED 2.1. SETTING THE ENVIRONMENT-USERS EXTERNAL TO SANDIA LABS To access Sierra/Arpeggio one will likely need to setup the user environment. This setup will differ upon location and the local system administrator can provide information on setting up your local environment. 2.2. SETTING THE ENVIRONMENT-USERS AT SANDIA LABS The environment for using Arpeggio is the same as for individual Sierra applications and can be configured by module files. The modules ensure that the look and feel of running Sierra applications is the same across a multitude of compute platforms. To obtain the proper environment for code execution one simply runs: % module load sierra 2.3. RUNNING ARPEGGIO This section includes some very simple examples of how to run Arpeggio. For more information on running on some of Sandia’s clusters, etc. see [2]. In its simplest form, Arpeggio can be run like this: % sierra arpeggio -i myrun.i In this example, myrun.i is the Arpeggio input file. The output – nonlinear iterations, time step information, etc. – will be written to a file called myrun.log. So, you can monitor the progress of the simulation by watching the log file. Alternatively, you can have all of the output sent to the display by using the -l logfile command line option. If you set the log file to be - (a single “minus” character) all of the output will be sent to the standard output (usually your display): % sierra arpeggio -i myrun.i -l - 17 If you would like to use aprepro in your input file, add the -a command line option to have your input file automatically processed: % sierra arpeggio -i myrun.i -l - -a Oftentimes we want to run Arpeggio remotely or locally in a batch mode, save any standard output and perhaps even logout from a session. Unfortunately, termination of the session through either voluntary (interactive) or involuntary (timeout) logout may in effect terminate the Arpeggio job. In this case one can prevent the job from terminating by using the Unix nohup command in conjunction with the standard execution command line. % nohup sierra arpeggio -i myrun.i -l YourLogFile -a If one wishes to run the job in a background mode the nohup command should be terminated with & at the end of the command line. 2.4. ARPEGGIO ENVIRONMENT OVERVIEW The Sierra Mechanics code suite is composed of several specialized applications which can operate either in standalone mode or coupled with each other. The various application models and algorithms are integrated into the Sierra framework through the architecture illustrated in Figure 2.4-1. A Sierra-based application has four layers of code: Domain, Procedure, Region, and Model/Algorithm. The outermost layer of an application is the Domain, or “main” program of the application. This domain layer is implemented by the Sierra Framework to manage the startup/shutdown of an application, and to orchestrate the execution of an application-proved set of procedures. Code at the Procedure level is responsible for evolving one or more loosely coupled set of physics through a sequence of steps. This sequence may be a set of time steps, nonlinear solver iterations, or some combinations of these or other types of steps. An application mauy define multiple procedures to implement hand-off coupling between physics within the same main program. In hand-off coupling the first (or preceding) procedure completes execution, mesh and field data is transferred to a succeeding procedure, and the succeeding procedure continues the simulation with a different set of physics. For example, the first thermal procedure could calculate a temperature distribution inside a differentially heated fluid, and the second procedure could simulate natural convection of the fluid due to the density gradients set up by the resulting temperature field. Code at the Region level is responsible for evolving a tightly coupled set of physics throug a single step. Loose coupling of Regions is supported by the advanced transfer services provided by the Sierra framework. Each Region owns (1) a set of models or algorithms that implement its tightly coupled set of physics and solvers and (2) an in-memory parallel distributed mesh and field database. This mesh and field data is fully distributed among parallel processors via domain decomposition. 18 “Main” Procedure (time step control) Region A Region B Parallel (single time step) (single time step) Synchronous Parallel Models Models Asynchronous Mesh and Fields Mesh and Fields Parallel Distributed Transfer Figure 2.4-1.. Schematic UML class diagram for the Expression subsystem. 2.5. OVERVIEW OF THE INPUT FILE STRUCTURE An Arpeggio model is described by commands contained in an ASCII input file. The structure of the input file follows a nested hierarchy. The topmost level of this hierarchy is named the domain. Below the domain lies a level named procedure, followed by the region level as depicted in Figure 2.4-1. The domain level contains one or more procedures. At the domain level, one will also find commands associated with describing the finite element mesh, the linear solver set-up, material properties associated with a defined material, and user functions associated with source terms and boundary conditions that are added into Arpeggio’s intrinsic set of functions. The procedure level contains one or more regions. The procedure level is also used to specify the time stepping parameters, and interactions between regions, such as data transfers. Essentially at the procedure level, loose coupling algorithms are specified. Loose coupling here is defined within the context of Arpeggio’s implicitly full-coupled paradigm. Whenever an independent variables’s interaction with other variables in the solution procedure is not fully represented in the global matrix, the algorithm for loose coupling of that variable and its associated equation will be described at the procedure level. This loose coupling algorithm is known as a “solution control description”. The procedure level contains a command block specifying the solution control procedure. An analogy to this block in simpler codes would be top level loop. For example in time dependent applications, the solution control description block would involve a block to solve the time dependent problem repeated for each time step until the desired solution time is reached. 19 The Region level is used to specify details about the physics to be solved. Details related to the solve include boundary conditions and initial conditions, where materials models are applied, and where surface and volumetric source terms are applied. Here the meshed discretization and material properties described at the domain level are tied into the problem statement by virtue of their names. Global constraint equations are also specified at the region level. At the region level, specification of information written to the output file and the frequency at which output occurs. Additional post-processing associated with the output is specified. For example, additional volumetric fields which are functions of the independent variables may be specified to be added to the output file. There are two types of commands in the input file. The first type is referred to as a block command. A block command is a grouping mechanism. A block command contains a set of commands made up of other block commands and line commands. A line command is the second type of command. The domain, procedure, and region levels are all parsed as block commands. A block command is defined in the input file by a matching pair of Begin and End lines. For example, Begin SIERRA myJob ... block commands End SIERRA myJob A set of key words for the block command follows the “Begin” and “End” keywords. In most cases a user-specified name is added to the block commands. In the example above the keywords, SIERRA myJob, are added. Optionally, the keyword may be left off of the end of the block. The second type of command is the line command. A line command is used to specify parameters within a given block command. In the remaining chapters and sections of this manual, the scope of each block and line command is identified, along with summaries of the meanings. Note that the ordering of any commands within a command block is arbitrary. Thus, Begin Finite Element model fluid Database name is pipeflow2d.g Use Material water for block_1 End Finite Element model fluid will have the same effect as Begin Finite Element model fluid Use Material water for block_1 Database name is pipeflow2d.g End Finite Element model fluid And the ordering of command blocks within the domain/procedure/region blocks are arbitrary–allowing you conderable freedom to collect and arrange commands. Note that the terms “command block” and “block command” are interchangeable. The Sierra command block must contain a block for a procedure containing at least one Region. For a case where only an Aria Region is being used: 20 Begin procedure myProcedureName . Begin Aria region myRegionName . End Aria region myRegionName End procedure myProcedure and similarly for a case using both Aria and Adagio Regions: Begin procedure myProcedureName . Begin Adagio region myAdagioRegionName . End Adagio region myAdagioRegionName . Begin Aria region myAriaRegionName . End Aria region myAriaRegionName End procedure myProcedure The procedure command block is used to contain all of the application code commands that are associated with a solution procedure defined for a set of Regions. The myProcedureName and name keywords of the procedure and region blocks are left up to you. Note that the procedure command block must be present in the input file and must contain at least one application code Region command block. The procedure command block also contains other important command blocks such as the SOLUTION CONTROL block. 2.5.1. Syntax Conventions for Commands In this section we describe the conventions used in presenting all the command descriptions in the remainder of this manual. There are four basic kinds of tokens, or words, that an application code expects to find as it parses an input file. These are keywords, names, parameters and delimiters. 2.5.1.1. Keywords The words which distinguish one block command, or line command, from another we term keywords. Keywords are denoted in this manual in the monospaced font, for example, BOUNDARY CONDITION. 21 2.5.1.2. Names The word, or words, that you supply on the same line of the begin line of a block command, is the name. Many times you may need to supply this name as a character parameter in a separate line command. Names are denoted in italics, name , as are parameters. It is worth noting that the interpreter used to process standard input command lines is also used to process lines defining algebraic operations. This means that a "-" appearing within a name would be interpreted as a subtraction operation and as a consequence, the use of "-" within a name is not allowed. Thus instead of Begin Adagio region name-1 one could perhaps use Begin Adagio region name_1. 2.5.1.3. Parameters There are three types of input parameters one will need to supply to line commands: character strings, integers, and real numbers. These are denoted in the documentation as (C), (R), and (I), respectively. In most cases character strings may be specified in a free format. One exception to this paradigm is when a string begins a number. In this case the character string must be specified within quotation marks in order to be properly interpreted. Real numbers may be entered in decimal form or exponential form. For example 0.0001, .1E-3, 10.0d-5 are all equivalent. Furthermore, if a real(R) is expected, an integer can be used. Integer values (I) need not include a decimal point in their specification. 2.5.1.4. Multiple Parameters For the case when a list of one or more paremeters is allowed, or required, for a command, (C,...) denotes a list of character strings, (I,...) a list of integers, and (R, ...) a list of real numbers. For a list of character strings, the separator between the strings must be one or more spaces or tab characters. Therefore, phrases with multiple spaces and words in them are tokenized into multiple character parameters before being processed by the application. For a list of real or integer numbers the comma can also be used as a separator. 2.5.1.5. Enumerated Parameters Certain commands have predefined parameters, called enumerations, which are listed within {}. Each parameter in the list is separated using | . The default parameter for the list of parameters is enclosed by <>. 22 2.5.1.6. Delimiters The keywords of a line command are often required to be separated from the parameters by a delimiter. You have a choice of delimiters to use: the equal sign, =, or a word. In this manual, we denote the choices surrounded by {}, and separated by |. You may use any one of the delimiters from those listed. For example, the line command to specify the density within the Calore Material Block command is Density {= |IS} (R) Examples of valid forms you could write in the input file are Begin Property Specification for Calore Material water ... Density = 1.0E-3 # kg/m\^3 at 20C ... End and Begin Aria Material water ... Density is constant rho = 1.0E-3 # kg/m\^3 at 20C ... End 2.5.1.7. White Space Command keywords, names, and parameters and delimiters must have spaces around them. 2.5.1.8. Indentation All leading spaces and/or tab characters are ignored in the input file. Of course, we recommend that you use indentation to improve the readability for yourself and others that may need to see your files. 2.5.1.9. Case Sensitivity None of the command keywords, parameters, or delimiters read from the input file are case sensitive. For example, the following two lines are equivalent: Use Material water for block_1 . and 23 USE material wATer for blOCK_1 . The exception to this rule are file names used for input and output, because the current operating systems on which SIERRA applications are run are based on UNIX, where file names are case sensitive. 2.5.1.10. Comments and Line Continuation You may place comments in the input file starting with either the $ or # character. All further characters on a line following a comment character are ignored. You can continue a command in the input file to the next line by using the line continuation character $, or you may optionally following it with a comment#. All further characters on the same line following a line continuation character $ are ignored, and the characters on the following line are joined and parsing continues. An example is the line command used to specify the title of a thermal model: Begin SIERRA Job_Indentifier # This thermal model for Calore simulates a convective heat transfer Title The title command is used to set the analysis title $\ Convective heat transfer to a part. The analysis $\ makes use of conjugate heat transfer to account for $\ cooling of a part due to flowing water. ... End SIERRA Job_Indentifier 2.5.1.11. Checking the Syntax Errors in the input deck can be checked by adding the command, “–check-syntax” to the aria command line. For example, sierra arpeggio --check-syntax -i input.i This command will print the code echo of the input deck and any syntax errors within it to the display. 24 2.6. FIELDS Fields are defined as variables which are distributed on mesh objects (e.g. nodes, elements, faces or edges). The mesh object and Field data may be distribued among parallel processors via a domain decomposition algorithm. Each application registers Fields by name on its own Region. In a coupled-physics simulation Framework transfer services may be called on to communicate these Fields to another application. For example, the temperature Field in one application may be communicated to a solid mechanics application in order to perform a thermal-stress analysis. 2.7. USER FIELDS Situations often arise where one wishes to provide Field data storage so that data can be transferred into or out of the application. Each of the application codes provide some mechanism for enabling this type of data access. Additionally, User Fields are often used to as additional storage needed in user supplied subroutines. 25 26 3. MODEL DEFINITION 3.1. MODEL OVERVIEW Sierra Framework services provide overall control of input commands, discretization input data and output data, IO. Additionally they provide a directed interaction of Framework services at the so-called Domain level with with the application code at the Region level. This controlled interaction is enabled by commands that follow. The model discretization (mesh) and the mesh components to be used in the model are defined at the Domain level and are later referenced by the application at the Region level. The association of material properties with portions of the mesh are also defined here within the Finite Element Model command block/s. For some couplings using the same mesh a single Finite Element Model may be used but for most cases one will use separate Finite Element Model command blocks for each Region. A sample outline of a setup for coupling of a solid mechanics application sm and a thermal-fluid tf is shown below. . Begin Finite Element Model my_fem_model_sm . End . Begin Finite Element Model my_fem_model_tf . End Begin Global Constants . End . Model definition commands - Material definitions for sm - Function definitions for sm - Local Coordinate Systems for sm . - Material definitions for tf 27 - Function definitions for tf - Local Coordinate Systems for tf . 28 Begin Procedure My_Procedure . procedural commands - Solution Control Description - Transfer operations . Begin Adagio Region My_Adagio_Region . use Finite Element Model my_fem_model_sm . - sm Region level commands . End . Begin Aria Region My_Aria_Region . use Finite Element Model my_fem_model_tf . - tf Region level commands . End . End . Note that a given application may not support the entire set of available options available in the Finite Element Model command block, particularly in the Parameters for Block section. Rather than attempting to include the entire set of command lines available in the Finite Element command block, only a small subset of key command lines are shown here. One should consult documentation for the specific application to find a complete listing of the relevant Finite Element Model command lines. 3.2. FINITE ELEMENT MODEL Scope: Sierra Begin Finite Element Model Finite-Element-Model-Name Alias DatabaseName As InternalName Component Separator Character Option Separator Create GroupType NewSurfaceName Add SurfaceName... Coordinate System {= | are | is} CoordinateSystem Database Name {= | are | is} StreamName 29 Database Type {= | are | is} DatabaseTypes Decomposition Method {= | are | is} Method Omit Assembly AssemblyList... Omit Block BlockList... Omit Volume VolumeList... Time Scale Factor Option Scale Use Generic Names Use Material MaterialName For VolumeList... Begin Assembly Assembly_Name End Begin Block Blockname End Begin Parameters For Block Blockname End Begin Parameters For Phase Phase Name End Begin Parameters For Surface Surface_Name End End Summary Describes the location and type of the input stream used for defining a geometry model for the enclosing region. 30 3.2.1. Alias Alias DatabaseName As InternalName DatabaseName string – InternalName string – Summary Name the database entity "DatabaseName" as "InternalName" Description This "InternalName" may then be referenced in the data file in addition to the original name. 3.2.2. Component Separator Character Component Separator Character Option Separator Option {= | is} – Separator string – Summary The separator is the single character used to separate the output variable basename (e.g. "stress") from the suffices (e.g. "xx", "yy") when displaying the names of the individual variable components. For example, the default separator is "_", which results in names similar to "stress_xx", "stress_yy", ... "stress_zx". To eliminate the separator, specify an empty string ("") or NONE. 3.2.3. Create Create GroupType NewSurfaceName Add SurfaceName... GroupType {edgeset | elemset | faceset | nodeset | – sideset | surface} NewSurfaceName string – SurfaceName string... – Summary Create a new set (node, edge, face, element, side/surface) as the union of two or more existing sets. The sets must exist in the mesh database or have been created by a previous CREATE command. 31 3.2.4. Coordinate System Coordinate System {= | are | is} CoordinateSystem Parameter Value Default CoordinateSystem {axisymmetric | barycentric | cartesian – | cyclidic | cylindrical | polar | quadriplanar | skew | spherical | toroidal | trilinear} Summary The interpretation of the geometry data stored in this database. Optional. Defaults to Cartesian. 3.2.5. Database Name Database Name {= | are | is} StreamName Parameter Value Default StreamName string – Summary The base name of the database containing the output results. If the filename begins with the ’/’ character, it is an absolute path; otherwise, the path to the current directory will be prepended to the name. If this line is omitted, then a filename will be created from the basename of the input file with a ".g" suffix appended. 3.2.6. Database Type Database Type {= | are | is} DatabaseTypes Parameter Value Default DatabaseTypes {catalyst | catalyst_exodus | cgns | – dof | dof_exodus | exodus | exodusii | exonull | generated | genesis | null | parallel_exodus | textmesh} Summary The database type/format used for the mesh. 32 3.2.7. Decomposition Method Decomposition Method {= | are | is} Method Method {block | cyclic | external | geom_kway – | hsfc | kway | kway_geom | linear | map | metis_sfc | random | rcb | rib | variable} Summary The decomposition algorithm to be used to partition elements to each processor in a parallel run. 3.2.8. Omit Assembly Omit Assembly AssemblyList... AssemblyList string... – Summary Specifies that the element blocks that are in the assemblies in AssemblyList will be omitted from the analysis. Description If an assembly is used to omit an element block, then it is illegal to refer to that element block later in the file. Any of the element blocks omitted will be removed from any assembly that contains them. 3.2.9. Omit Block Omit Block BlockList... BlockList string... – Summary Specifies that the element blocks named in the blockList be omitted from the analysis. Description If an element block is omitted, then it is illegal to refer to it later in the input file e.g an initial condition may not be specified on an omitted element block. The elements, faces, etc are never created and it is as if the omitted element blocks did not exist in the mesh file. If a surface is completely determined by the omitted element block, then it is illegal to specify boundary conditions on that surface. However, if the surface spans multiple element blocks, boundary conditions may be applied on the portion of the surface supported by the element blocks that are not omitted. 33 3.2.10. Omit Volume Omit Volume VolumeList... VolumeList string... – Summary Specifies that the volumes named in the volumeList be omitted from the analysis. Description If a volume is omitted, then it is illegal to refer to it later in the input file e.g an initial condition may not be specified on an omitted volume. The elements, faces, etc are never created and it is as if the omitted volumes did not exist in the mesh file. If a surface is completely determined by the omitted volume, then it is illegal to specify boundary conditions on that surface. However, if the surface spans multiple volumes, boundary conditions may be applied on the portion of the surface supported by the volumes that are not omitted. 3.2.11. Time Scale Factor Time Scale Factor Option Scale Option {= | is} – Scale real – Summary The scale factor to be applied to the times on the mesh database. If the scale factor is 20 and the times on the mesh database are 0.1, 0.2, 0.3, then the application will see the mesh times as 2, 4, 6. 3.2.12. Use Generic Names Summary If this command is present then the name of all blocks and sets in the mesh will be of the form "type_"+id. For example, an element block with id=42 will be named "block_42"; a sideset with id 314 will be named "surface_314". If there are any names in the mesh file, those names will be aliases for the blocks and sets. If this command is not present, then if a name is in the mesh file, it will be used as the name and the generic generated name will be an alias. This is used as a workaround in codes that do not correctly handle named blocks and sets or as a workaround in meshes which contain non-user-specified names. 34 3.2.13. Use Material Use Material MaterialName For VolumeList... Parameter Value Default MaterialName string – VolumeList string... – Summary Associate the given volumes with the indicated material name. 3.3. PARAMETERS FOR BLOCK Begin Parameters For Block Blockname Active For Procedure ProcedureName During Periods PeriodNames... Bending Hourglass Option {= | are | is} Hgval Density Scale Factor {= | are | is} densityScaleFactor Deposit Specific Internal Energy Edep [ Over Time Tdep Starting At Time Tinit ] Effective Moduli Model {= | are | is} Option Element Numerical Formulation {= | are | is} Option Energy Iteration Tolerance {= | are | is} Eit Hourglass Option {= | are | is} Hgval Hourglass Option {= | are | is} Hgval Inactive For Procedure ProcedureName During Periods PeriodNames... Include All Blocks Inversion Aversion Exponent {= | are | is} ia_exponent Inversion Aversion Stiffness {= | are | is} ia_stiffness Inversion Aversion Transition Jacobian {= | are | is} transition_jacobian Linear Bulk Viscosity {= | are | is} Lbv Local Coordinate System {= | are | is} Mesh Entities Material MatName 35 Material = MatName Max Energy Iterations {= | are | is} Mei Membrane Hourglass Option {= | are | is} Hgval Minimum Effective Dilatational Moduli Ratio {= | are | is} minEffectiveModuliRatio Minimum Effective Shear Moduli Ratio {= | are | is} minEffectiveModuliRatio Model {= | are | is} ModelName Nonlocal Regularization Kmeans Cell Size {= | are | is} kmeans_cell_size Nonlocal Regularization Kmeans Maximum Iterations {= | are | is} kmeans_maximum_iterations Nonlocal Regularization Kmeans Tolerance {= | are | is} kmeans_tolerance Nonlocal Regularization On stateVariableName With Length Scale {= | are | is} lengthScale [ And Staggering ] Nonlocal Regularization Partitioning Scheme {= | are | is} PartitioningScheme Phase PhaseLabel {= | are | is} MaterialName Quadratic Bulk Viscosity {= | are | is} Qbv Remove Block {= | are | is} ExcludeBlockList... Section {= | are | is} SectionName Solid Mechanics Use Model ModelName Transverse Shear Hourglass Option {= | are | is} Hgval End Summary Specifies analysis parameters associated with each element block. 36 3.3.1. Active For Procedure Active For Procedure ProcedureName During Periods PeriodNames... ProcedureName string – PeriodNames string... – Summary Lists the solution periods during which the given BC, solver, preconditioner, etc. is active. Multiple uses of this line command within a single block will have a cumulative affect. 3.3.2. Bending Hourglass Bending Hourglass Option {= | are | is} Hgval Option {stiffness | viscosity} – Hgval real – Summary Supplies the hourglass stiffness and viscosity parameters for bending deformation in a shell element block. 3.3.3. Density Scale Factor Density Scale Factor {= | are | is} densityScaleFactor densityScaleFactor real – Summary Specifies a scale factor to apply to the density defined in the material. This value must be greater than zero. The default is 1.0 (no scaling). 37 3.3.4. Deposit Specific Internal Energy Deposit Specific Internal Energy Edep [ Over Time Tdep Starting At Time Tinit ] Edep real – Summary Defines the amount of specific (per unit mass) internal energy to be deposited in the material. The energy is deposited over time tdep, beginning at time tinit. The optional parameters tdep and tinit both default to zero, so the energy will be deposited instantaneously at time zero if they are not specified. The deposition is uniform in space, so each element in the block has the same amount edep added to its specific internal energy. 3.3.5. Effective Moduli Model Effective Moduli Model {= | are | is} Option Option {elastic | probed | pronto} – Summary Specifies the method used to determine the effective moduli. This choice can have a significant effect on the resulting hourglassing behavior. The models are: * elastic: use the initial elastic moduli * pronto: use the old PRONTO method for computing elastic moduli this approach is straight out of PRONTO, PRESTO’s predecessor. This is a bounded tangent method. * probe: Use a pronto-like method, but pass in a an artificial probe strain rate rather than the actual strain. 3.3.6. Element Numerical Formulation Element Numerical Formulation {= | are | is} Option Option {new | old} – Summary Specifies which element numerical formulation to use for this block. 38 3.3.7. Energy Iteration Tolerance Energy Iteration Tolerance {= | are | is} Eit Parameter Value Default Eit real – Summary Specifies the tolerance criteria for exiting the iterative solve of the implicit internal energy update equation. Applicable when using EOS material models with extracted energy updates. 3.3.8. Hourglass Hourglass Option {= | are | is} Hgval Parameter Value Default Option {stiffness | viscosity} – Hgval real – Summary Supplies the hourglass stiffness and viscosity parameters for this element block. 3.3.9. Hourglass Hourglass Option {= | are | is} Hgval Parameter Value Default Option {exponent | transition strain} – Hgval real – Summary Supplies the hourglass stiffness and viscosity parameters for this element block. 3.3.10. Inactive For Procedure Inactive For Procedure ProcedureName During Periods PeriodNames... Parameter Value Default ProcedureName string – PeriodNames string... – 39 Summary Lists the solution periods during which the given BC, solver, preconditioner, etc. is inactive. Multiple uses of this line command within a single block will have a cumulative affect. 3.3.11. Include All Blocks Summary Use this parameters definition for all blocks. When using this option within the FINITE ELEMENT MODEL command block the PARAMETERS FOR BLOCK will not use a Blockname. 3.3.12. Inversion Aversion Exponent Inversion Aversion Exponent {= | are | is} ia_exponent Parameter Value Default ia_exponent integer 5 Summary Sets the exponent used to compute the smooth approximate nodal jacobian ratio. A higher exponent results in a more-accurate approximation to the ratio. This is only active for uniform gradient elements. Default = 5. 3.3.13. Inversion Aversion Stiffness Inversion Aversion Stiffness {= | are | is} ia_stiffness Parameter Value Default ia_stiffness real 1.e5 Summary Sets a stiffness parameter for the inversion aversion penalty. This is only active for uniform gradient elements. Default = 1.0e5. 40 3.3.14. Inversion Aversion Transition Jacobian Inversion Aversion Transition Jacobian {= | are | is} transition_jacobian Parameter Value Default transition_jacobian real 0 Summary Sets the critical relative nodal Jacobian ratio for inversion aversion. If this value is nonzero, an additional recoverable energy term is added which penalizes further element distortion. This energy is only active for uniform gradient elements. 3.3.15. Linear Bulk Viscosity Linear Bulk Viscosity {= | are | is} Lbv Parameter Value Default Lbv real – Summary Supplies the linear coefficient for the bulk viscosity computations. 3.3.16. Local Coordinate System Local Coordinate System {= | are | is} Mesh Entities Parameter Value Default Mesh Entities string – Summary Associate coordinate system with mesh entity. Description Specify the local coordinate system to be used in conjunction with given element blocks. 3.3.17. Material Material MatName Parameter Value Default MatName string – Summary Associates this element block with its material properties. 41 3.3.18. Material = Material = MatName Parameter Value Default MatName string – Summary Associates this element block with its material properties. 3.3.19. Max Energy Iterations Max Energy Iterations {= | are | is} Mei Parameter Value Default Mei integer – Summary Specifies the maximum number of iterations to take in solving the implicit internal energy update equation. Applicable when using EOS material models with extracted energy updates. 3.3.20. Membrane Hourglass Membrane Hourglass Option {= | are | is} Hgval Parameter Value Default Option {stiffness | viscosity} – Hgval real – Summary Supplies the hourglass stiffness and viscosity parameters for membrane deformation in a shell or membrane element block. 3.3.21. Minimum Effective Dilatational Moduli Ratio Minimum Effective Dilatational Moduli Ratio {= | are | is} minEffectiveModuliRatio Parameter Value Default minEffectiveModuliRatio real – 42 Summary Specifies a minimum effective DILATATIONAL moduli ratio. This value keeps the effective moduli from dropping below minEffectiveModuliRatio*ElasticModulus. This can aid in keeping the corresponding time step and bulk viscosity terms dropping to zero 3.3.22. Minimum Effective Shear Moduli Ratio Minimum Effective Shear Moduli Ratio {= | are | is} minEffectiveModuliRatio minEffectiveModuliRatio real – Summary Specifies a minimum effective SHEAR moduli ratio. This value keeps the effective moduli from dropping below minEffectiveModuliRatio*ElasticModulus. This can aid in keeping the corresponding hourglass stiffness terms dropping to zero 3.3.23. Model Model {= | are | is} ModelName ModelName string – Summary Associates a solid mechanics material model with this element block. The material parameters for this block are specified in the material denoted by the MATERIAL command. 3.3.24. Nonlocal Regularization Kmeans Cell Size Nonlocal Regularization Kmeans Cell Size {= | are | is} kmeans_cell_size kmeans_cell_size real – Summary This line command specifies the cell size used to construct the background grid for the computation of the centroidal Voronoi tessellation for the Kmeans partitioning scheme. 43 3.3.25. Nonlocal Regularization Kmeans Maximum Iterations Nonlocal Regularization Kmeans Maximum Iterations {= | are | is} kmeans_maximum_iterations kmeans_maximum_iterations integer – Summary This line command specifies the maximum number of iterations to perform for Lloyd’s algorithm for the computation of the centroidal Voronoi tessellation for the Kmeans partitioning scheme. 3.3.26. Nonlocal Regularization Kmeans Tolerance Nonlocal Regularization Kmeans Tolerance {= | are | is} kmeans_tolerance kmeans_tolerance real – Summary This line command specifies the relative tolerance for Lloyd’s algorithm. Iterations continue until the maximum number is reached or the L2 norm of a vector of all the center steps is less or equal than the tolerance times the cell size of the background grid. 3.3.27. Nonlocal Regularization On Nonlocal Regularization On stateVariableName With Length Scale {= | are | is} lengthScale [ And Staggering ] stateVariableName string – lengthScale real – Summary This line command will cause the mesh to be partitioned into sub domains where each sub domain volume is on the order of lengthScale3 and regularizes the governing PDE by averaging the material state variable stateVariableName over the sub domain. 44 3.3.28. Nonlocal Regularization Partitioning Scheme Nonlocal Regularization Partitioning Scheme {= | are | is} PartitioningScheme Parameter Value Default PartitioningScheme {kmeans | metis | zoltan_hypergraph | – zoltan_rcb | zoltan_rib} Summary This line command specifies the type of partitioning algorithm used to perform the domain decomposition for the nonlocal regularization method 3.3.29. Phase Phase PhaseLabel {= | are | is} MaterialName Parameter Value Default PhaseLabel string – MaterialName string – Summary Associate phase PhaseLabel with material Material_Name on this block. 3.3.30. Quadratic Bulk Viscosity Quadratic Bulk Viscosity {= | are | is} Qbv Parameter Value Default Qbv real – Summary Supplies the quadratic coefficient for the bulk viscosity computations. 3.3.31. Remove Block Remove Block {= | are | is} ExcludeBlockList... Parameter Value Default ExcludeBlockList string... – Summary List of blocks to exclude. 45 3.3.32. Section Section {= | are | is} SectionName Parameter Value Default SectionName string – Summary Specifies the section to use for this element block. 3.3.33. Solid Mechanics Use Model Solid Mechanics Use Model ModelName Parameter Value Default ModelName string – Summary Associates a solid mechanics material model with this element block. The material parameters for this block are specified in the material denoted by the MATERIAL command. 3.3.34. Transverse Shear Hourglass Transverse Shear Hourglass Option {= | are | is} Hgval Parameter Value Default Option {stiffness | viscosity} – Hgval real – Summary Supplies the hourglass stiffness and viscosity parameters for transverse shear deformation in a shell element block. 3.4. GLOBAL CONSTANTS Scope: Sierra Begin Global Constants empty Faradays Constant {= | are | is} Faraday Gravity Vector {= | are | is} Gravity1 Gravity2 Gravity3 Ideal Gas Constant {= | are | is} Sigma K-E Turbulence Model Parameter Param {= | are | is} Value 46 K-W Turbulence Model Parameter Param {= | are | is} Value Les Turbulence Model Parameter Param {= | are | is} Value Light Speed {= | are | is} LightSpeed Planck Constant {= | are | is} PlanckConstant Stefan Boltzmann Constant {= | are | is} Sigma Turbulence Model Param Number {= | are | is} Value End Summary Set of universal constants for a simulation. 3.4.1. Faradays Constant Faradays Constant {= | are | is} Faraday Faraday real – Summary Faraday’s Constant. NOTE: Another Faraday’s constant value can be specified while using certain code capabilities. This global constants value will be discarded for any other specified Faraday’s constant values. 3.4.2. Gravity Vector Gravity Vector {= | are | is} Gravity1 Gravity2 Gravity3 Gravity real_1 real_2 real_3 – Summary Gravity constant in vector form, acceleration components. 47 3.4.3. Ideal Gas Constant Ideal Gas Constant {= | are | is} Sigma Parameter Value Default Sigma real – Summary Ideal gas constant. NOTE: Another ideal gas constant value can be specified while using certain code capabilities. This global constants value will be discarded for any other specified ideal gas constant values. 3.4.4. K-E Turbulence Model Parameter K-E Turbulence Model Parameter Param {= | are | is} Value Parameter Value Default Param string – Value real – Summary k − ϵ RANS turbulence model parameters. 3.4.5. K-W Turbulence Model Parameter K-W Turbulence Model Parameter Param {= | are | is} Value Parameter Value Default Param string – Value real – Summary k − ω RANS turbulence model parameters. 3.4.6. Les Turbulence Model Parameter Les Turbulence Model Parameter Param {= | are | is} Value Parameter Value Default Param string – Value real – Summary LES turbulence model parameters. 48 3.4.7. Light Speed Light Speed {= | are | is} LightSpeed LightSpeed real – Summary Speed of Light. Depending on the units involved in the specific problem by the user, this value will differ. 3.4.8. Planck Constant Planck Constant {= | are | is} PlanckConstant PlanckConstant real – Summary Planck Constant. Depending on the units involved in the specific problem by the user, this value will differ. 3.4.9. Stefan Boltzmann Constant Stefan Boltzmann Constant {= | are | is} Sigma Sigma real – Summary Stefan-Boltzmann constant. Depending on the units involved in the specific problem by the user, this value will differ. 3.4.10. Turbulence Model Turbulence Model Param Number {= | are | is} Value Param string – Value real – Summary Turbulence model Schmidt and Prandtl numbers 49 3.5. DEFINITION FOR FUNCTION Scope: Sierra Begin Definition For Function FunctionName Abscissa {= | are | is} Name... Abscissa Offset {= | are | is} Abscissa_offset Abscissa Scale {= | are | is} Abscissa_scale At Discontinuity Evaluate To Option Column Titles Titles1 Titles2 ... Data File = filename [ X From Column xcol Y From Column ycol ] Debug {= | are | is} Option Differentiate Expression {= | are | is} Expr Evaluate Expression {= | are | is} Expr Evaluate From x0 To x1 By Dx Expression Variable: Expr = VarType source_var_name... [ State StateEnum ] Expression Variable: Expr Field Types Titles1 Titles2 ... Ordinate {= | are | is} Name... Ordinate Offset {= | are | is} Ordinate_offset Ordinate Scale {= | are | is} Ordinate_scale Scale By x Type {= | are | is} Type X Offset {= | are | is} X_offset X Scale {= | are | is} X_scale Y Offset {= | are | is} Y_offset Y Scale {= | are | is} Y_scale Begin Expressions empty End Begin Values empty End 50 End Summary Defines a function in terms of its type and values. 3.5.1. Abscissa Abscissa {= | are | is} Name... Name string... – Summary Specifies a string identifier for the independent variable. Optionally specify a scale and/or offset value which transforms the abscissa values into scaled_abscissa = scale * ( abscissa + abscissa_offset). 3.5.2. Abscissa Offset Abscissa Offset {= | are | is} Abscissa_offset Abscissa_offset real – Summary Alias for X OFFSET 3.5.3. Abscissa Scale Abscissa Scale {= | are | is} Abscissa_scale Abscissa_scale real – Summary Alias for X SCALE 51 3.5.4. At Discontinuity Evaluate To At Discontinuity Evaluate To Option Option {left | right} – Summary Control the behavior of a piecewise constant function when evaluated at a discontinuity (plus or minus a small tolerance). The default behavior is to take the value to the right of the discontinuity. If "Left" is specified, the value to the left of the discontinuity is taken instead. 3.5.5. Column Titles Column Titles Titles1 Titles2 ... Titles string_1 string_2... – Summary Name the columns (and also defined the expected number of columns) for Multicolumn Piecewise Linear tabular data. 3.5.6. Data File Data File = filename [ X From Column xcol Y From Column ycol ] filename string – Summary Function will read tabular data from an input file. Compatible with the piecewise linear function type. File must be of form like: --------------------------------------------- \# EXAMPLE FILE 1.099 1191 1.101 221 5.9011 133.1 --------------------------------------------- Lines headed by a # are considered comments and will be ignored. Data itself must by in tabular columns separated by whitespace or commas. 52 3.5.7. Debug Debug {= | are | is} Option Parameter Value Default Option {off | on} – Summary Prints functions to the log file. 3.5.8. Differentiate Expression Differentiate Expression {= | are | is} Expr Parameter Value Default Expr (expression) – Summary Specifies the expression of derivative of evaluation expression. 3.5.9. Evaluate Expression Evaluate Expression {= | are | is} Expr Parameter Value Default Expr (expression) – Summary Specifies the expression to evaluate. Description This will greatly help with manufactured solutions, and be useful for other purposes as well. This uses the STK expression evaluator to evaluate the provided string. See the STK user manual for details about valid syntax. begin definition for function pressure type is analytic evaluate expression is "x <= 0.0 ? 0.0 : (x < 0.5 ? x*200.0 : (x < 1.0 ? (x - 0.5) *50.0 + 100.00 : 150.0))" end definition for function pressure 53 3.5.10. Evaluate From Evaluate From x0 To x1 By Dx Parameter Value Default x0 real – x1 real – Dx real – Summary Specifies the range and evaluation interval. 3.5.11. Expression Variable: Expression Variable: Expr = VarType source_var_name... [ State StateEnum ] Parameter Value Default Expr string – VarType {element | element_sym_tensor | – element_tensor | element_vector | face | global | nodal | nodal_sym_tensor | nodal_tensor | nodal_vector} source_var_name string... – Summary Specifies what the arguments of an expression correspond to. For example: BEGIN DEFINITION FOR FUNCTION dx_shear TYPE = ANALYTIC EXPRESSION variable: mx = NODAL model_coordinates(x) EXPRESSION variable: my = NODAL model_coordinates(y) EXPRESSION variable: time = GLOBAL time EVALUATE EXPRESSION = "(time/{termTime})*({stretchx}*(mx - 0.0) + ( END Assuming the above expression is being evaluated on nodes the current values for x and y model coordinates would be placed into mx and my and current analysis time placed into time 54 3.5.12. Expression Variable: Expression Variable: Expr Expr string – Summary Specifies what the arguments of an expression exists, but does not define it correspond to. For example: BEGIN DEFINITION FOR FUNCTION dx_shear TYPE = ANALYTIC EXPRESSION variable: mx EXPRESSION variable: my EXPRESSION variable: time EVALUATE EXPRESSION = "(time/{termTime})*({stretchx}*(mx - 0.0) + ( END Call function must determine what each variable actually is is based off of the string name 3.5.13. Field Types Field Types Titles1 Titles2 ... Titles string_1 string_2... – Summary The field types (GLOBAL/NODE/ELEMENT) that correspond to the column titles listed for the multicolumn data. 3.5.14. Ordinate Ordinate {= | are | is} Name... Name string... – Summary Specifies a string identifier for the dependent variable. Optionally specify a scale and/or offset value which transforms the ordinate values into scaled_ordinate = scale * ( ordinate + ordinate_offset). 55 3.5.15. Ordinate Offset Ordinate Offset {= | are | is} Ordinate_offset Parameter Value Default Ordinate_offset real – Summary Alias for Y OFFSET 3.5.16. Ordinate Scale Ordinate Scale {= | are | is} Ordinate_scale Parameter Value Default Ordinate_scale real – Summary Alias for Y SCALE 3.5.17. Scale By Scale By x Parameter Value Default x real – Summary Specifies a scale factor to be applied. 3.5.18. Type Type {= | are | is} Type Parameter Value Default Type {analytic | constant | multicolumn – piecewise linear | piecewise analytic | piecewise constant | piecewise linear | piecewise multivariate | xtable} Summary Specifies the type of function. 56 3.5.19. X Offset X Offset {= | are | is} X_offset Parameter Value Default X_offset real – Summary Sets an offset for the x-axis 3.5.20. X Scale X Scale {= | are | is} X_scale Parameter Value Default X_scale real – Summary Sets a scale factor for the x-axis 3.5.21. Y Offset Y Offset {= | are | is} Y_offset Parameter Value Default Y_offset real – Summary Sets an offset for the y-axis 3.5.22. Y Scale Y Scale {= | are | is} Y_scale Parameter Value Default Y_scale real – Summary Sets a scale factor for the y-axis 57 3.6. VALUES Scope: Definition For Function Begin Values empty Xyvalues... End Summary Lists the values of the function. The values should be listed one pair per line, independent variable first, with whitespace or comma as a separator. 3.6.1. Scope: Values Xyvalues... Parameter Value Default Xyvalues real... – Summary For a piecewise linear function, lists an x-y pair for the nth interpolation point. 3.7. RESTART OVERVIEW Sierra Framework services provide convenient utilities for restarting an analysis from previous results. The most general capability supplements the results of a previous analysis with internal state variables to continue an analysis. In this case the input mesh is supplied from the Input Database Name from the Finite Element Model command block 3.1 and the restart information is obtained from the the Input Database Name from the Restart Data command block. Continuation of a job using restart data output is invoked using the command line which follows. 58 4. SOLUTION CONTROL REFERENCE 4.1. OVERVIEW Arpeggio uses the solution control (SC) library from the SIERRA Framework to orchestrate execution of simulations. All Arpeggio input files must include a Solution Control Description block in the Procedure section of the input file. This description contains directives for executing a steady-state (sequential) or transient analysis, either of which can include nested nonlinear iteration or subcycling. Within the description one selects a named solution control system where the details of execution are more clearly spelled out. Because there are similarities between the Sequential, Transient, Nonlinear Iteration and Subcycling many operations are shared between these directives. However, each of these segments must be uniquely named internally so they can be properly managed under solution control. Within each SC system, execution of a problem defined at the Region level corresponds to an Advance directive. Thus a steady-state analysis could conceivably be carried out with a single Advance directive. For transient analysis the system can contain several time blocks, each with a corresponding Advance directive. Examples of different control structures are first demonstrated, followed by syntactical descriptions of the solution control structures. 4.1.1. Steady Analysis As an example, the solution control command block for a steady-state Aria analysis would reflect the structure indicated below: . Begin Procedure myProcedure Advance myRegion End 59 End End Begin Aria Region myRegion . . ICs, BCs, equations, output instructions . . myRegion output . End Aria Region myRegion End Procedure myProcedure . Problems of fully-coupled physics are best solved within a single application. As an alternative, loose-coupling is often carried out by supplying local application variables that define the coupling to another application. Solution control provides various means of carrying out these analyses depending upon the strength of physics coupling within a solution step. When the different problem physics are weakly-coupled one often assumes that for each solution step it is sufficient to supply current values of variables involved in the coupling to the the other physics. In cases of stronger coupling one may chose to iterate on the exchange of information between the two physics until the interaction between physics has converged within a solution step before advancing. A solution control command block for steady-state analysis containing nonlinear iteration for Aria and Adagio would reflect the general structure indicated below. Here the nonlinear iteration will continue until user specified criteria, Parameters for Nonlinear Iteration satisfy the converged criteria and the solution will then advance. . Begin Procedure myProcedure Begin Nonlinear Iteration Advance myAriaRegion Advance myAdagioRegion End Nonlinear Iteration 60 End End End Begin Parameters for Nonlinear Iteration Converged when "myAriaRegion.MaxInitialNonlinearResidual(0) < 1E-1 && myAdagioRegion.MaxInitialNonlinearResidual(0) < 1E-1 " End Begin transfer adagio_to_aria . . transfer commands . End transfer adagio_to_aria Begin Aria Region myAriaRegion . . ICs, BCs, equations . myAriaRegion output . End Aria Region myAriaRegion Begin Adagio Region myAdagioRegion . . ICs, BCs, equations . myAdagioRegion output . End Adagio Region myAdagioRegion End Procedure myProcedure . 4.1.2. Transient Analysis In the case of transient analysis the solution control command block will contain specification of times for which the analysis will be carried out. Additionally, parameters defining the time integration for each Region must also be supplied by the user. Details concerning time integration parameters are included in the user manual for the application. 61 A simple example of transient analysis including two Aria Regions would resemble the structure shown below: . . Simulation Start Time = 0.0 Simulation Termination Time = 10.0 Simulation Max Global Iterations = 1000 Begin Transient Time_Block_1 Advance My_Aria_Region End Begin Transient Time_Block_2 Advance My_Aria_Region End End Begin Parameters For Transient Time_Block_1 Start Time = 0.0 Number of steps = 8 Begin Parameters For Aria Region My_Aria_Region Time Step Variation = Fixed Initial Time Step Size = 0.001 End End Begin Parameters For Transient Time_Block_2 Begin Parameters For Aria Region My_Aria_Region Time Step Variation = Adaptive Initial Time Step Size = 0.001 Predictor-Corrector Tolerance = 1e-3 Minimum Time Step Size = 1e-6 End End 62 End . End Procedure My_Aria_Procedure . A simple example of stronger coupling for transient analysis with nonlinear iteration would resemble the structure indicated below. Here time step for advancement to the next solution step is negotiated between coupled Regions based upon their respective Transient parameters. . . Simulation Start Time = 0.0 Simulation Termination Time = 10.0 Simulation Max Global Iterations = 1000 Begin Transient Time_Block Begin Nonlinear Iteration Advance myAriaRegion Advance myAdagioRegion End Nonlinear Iteration End End Begin Parameters for Nonlinear Iteration Converged when "myAriaRegion.MaxInitialNonlinearResidual(0) < 1E-1 && myAdagioRegion.MaxInitialNonlinearResidual(0) < 1E-1 " End Begin Parameters For Transient Time_Block_1 Start Time = 0.0 Termination Time = 10.0 Begin Parameters For Aria Region myAriaRegion Time Step Variation = adaptive 63 Initial Time Step Size = 0.001 Predictor-Corrector Tolerance = 1e-3 Minimum Time Step Size = 1e-6 End Begin Parameters For Adagio Region myAdagioRegion Time increment = 0.001 End End End Solution Control Description Begin transfer adagio_to_aria . transfer commands End transfer adagio_to_aria Begin transfer aria_to_adagio . transfer commands End transfer aria_to_adagio . End Procedure My_Aria_Procedure . Similarly, subcycled iterations in a two-way coupling between Aria and Presto could also be carried out in a transient analysis. In this case Presto subcycles at a small time step, Aria has a larger time step and the solution will advance when the two time applications arrive at the same solution time. In most application codes management of variable state is done using pointer-swaps and with SC one can manage the states available an application through appropriate choice of transfer to the subcycled application. Here transfer of temperature state new to both presto state old and new will allow the same temperature to be used by presto throughout the subcyle operation. . Simulation Start Time = 0.0 Simulation Termination Time = 10.0 Simulation Max Global Iterations = 1000 64 Begin Transient Time_Block_1 Transfer Presto_to_Aria Advance My_Aria_Region Transfer Aria_to_Presto Begin Subcycle PrestoSubcycle Advance PrestoRegion End End End Begin Parameters For Transient Time_Block_1 Start Time = 0.0 Number of steps = 8 Begin Parameters For Aria Region My_Aria_Region Time Step Variation = Fixed Initial Time Step Size = 0.001 End Begin Parameters for Presto Region PrestoRegion initial time step = 1.0e-6 # time step scale factor = 1.0 time step increase factor = 10. # step interval = 500 End End End . Begin transfer presto_to_aria . transfer commands End transfer presto_to_aria Begin transfer aria_to_presto Copy Volume Nodes From My_Aria_region to PrestoRegion Send Field solution->Temperature State New to Temperature State New Send Field solution->Temperature State New to Temperature State Old End transfer aria_to_presto . Begin Aria Region myAriaRegion . . End Aria Region myAriaRegion 65 Begin Presto Region myPrestoRegion . . End Presto Region myPrestoRegion End Procedure myProcedure . 4.1.3. Conditional Operations It is important to note that Solution Control can orchestrate the execution of one Region or the execution of many Regions. Within a loosely-coupled code analysis SC is also used to control the movement of data between the coupled codes using the Transfer subsystem. The outline views of various couplings include both Transfer and Advance events. In the examples above the event will always occur in the sequence specified. Alternatively one can specify that the event be carried out conditionally subject to criteria described syntactically as a "C" language [W hen − expression] where the expression criteria includes internal code variables or explicit evaluations. Here the input [W hen − expression] is parsed and transformed into an executable "C" statement. While some of the internal code variables used by a [W hen − expression] are intuitive (i.e. CURRENT_TIME and CURRENT_STEP) many others are application dependent. The most widely used explicit evaluations are measures of convergence based upon solution residuals adagio.norm(0.0) for solid mechanics applications and aria.MaxResidualNorm(0.0) for thermal-fluid applications. Other thermal-fluid evaluations used in convergence comparisons are aria.MaxInitialNonlinearResidual(0.0), aria.MaxInitialNonlinearCorrection(0.0) and aria.MaxInitialNonlinearCorrection(0.0). Several examples of [W hen − expression] are given below noting that the "C" expression must be enclosed in quotes within the input file. Convergence based upon comparison of application residuals: Begin parameters for nonlinear converge_step_p1 # following two lines shown must be a single input command line converged when $"(aria.MaxResidualNorm(0.0) < 1.e-6 && adagio.norm(0.0) < 1.e-6) || CURRENT_STEP > 2000" End parameters for nonlinear converge_step_p1 Transfer at first step and then every four steps: Transfer aria_to_adagio when "(CURRENT_STEP == 1) || (CURRENT_STEP % 4 == 0)" Advance the region at second step: 66 advance aria_region when "CURRENT_STEP == 2" Additionally, one may also use application specific global variables in the [W hen − expression] criteria. Global variables that are generally available for use are listed as such in the simulation log file. Unfortunately these variables may not be directly accessible to the user. Hence consultation with an application developer may be required in this regard. 4.1.4. Variable Initialization In the case of transient analysis it is sometimes necessary to initialize a distribution of values before the analysis actually begins. As an example, one may want to initialize a Field that will be transferred to another Region with a distribution of values with the goal of setting a reference state. For this purpose solution control provides a means for variable initialization, Initialize. . . Begin Initialize Transfer var1_Region_to_var2_My_Aria_Region End Initialize Use system Initialize Simulation Start Time = 0.0 Simulation Termination Time = 10.0 Simulation Max Global Iterations = 1000 Begin Transient Time_Block_1 Advance My_Aria_Region Advance var1_Region End End Begin Parameters For Transient Time_Block_1 Start Time = 0.0 Number of steps = 8 Begin Parameters For Aria Region var1_Region 67 . parameter commands End Begin Parameters For Aria Region My_Aria_Region . parameter commands End End End . . Var1_Region commands . . My_Aria_Region commands . End Procedure My_Aria_Procedure . 4.1.5. Mixed Physics Usage There are certains steps one will have to take when it is desired to just advance either the Aria or Adagio region in one of the transient blocks in an Arpeggio simulation. The example shown below displays an Adagio-only second transient block. All transferred fields from the disabled app to the still-enabled app need to be handled using the steps described below. With this example, the only relevant transferred field is the temperature. The first step taken was to switch new and old temperature states in Aria between the transient blocks. This will allow the Adagio region in the next transient block to use the latest temperature solution from Aria. The other notable syntax in the second transient block is to keep the aria to adagio transfer command in place. This input syntax will make sure the Adagio region uses a consistent temperature value while the region is being advanced. Begin Transient time_block Advance aria_region #begin subcycle Transfer aria_to_adagio Advance adagio_region #end subcycle Transfer adagio_to_aria 68 End Transient time_block Transfer T_switch Begin Transient time_block2 # Advance aria_region #begin subcycle Transfer aria_to_adagio Advance adagio_region #end subcycle End Transient time_block2 End System Main Begin Transfer aria_to_adagio Copy Volume Nodes from aria_region to adagio_region Send Field solution->TEMPERATURE State NEW to Temperature State NEW End Transfer aria_to_adagio Begin Transfer adagio_to_aria Copy Volume Nodes from adagio_region to aria_region Send Field DISPLACEMENT State New to Solution->Mesh_Displacements State New End Transfer adagio_to_aria Begin Transfer T_switch Copy Volume Nodes From Aria_region to Aria_region Send Field solution->Temperature State Old to solution->Temperature State Ne Send Field solution->Temperature State New to solution->Temperature State Ol End 4.2. SOLUTION CONTROL DESCRIPTION Scope: Procedure Begin Solution Control Description Name Use System Name Begin Adaptiveloop Name End Begin Initialize Name 69 End Begin Parameters For End Begin System Name End End Summary Contains the commands needed to execute an analysis using the arpeggio procedure that uses Solver Control. 4.2.1. Use System Scope: Solution Control Description Use System Name Parameter Value Default Name string – Summary This set the name of which system to use. 4.3. SYSTEM Scope: Solution Control Description Begin System Name Adapt Mesh For AdaptRegionName Using AdaptBlockName [When WhenExpression ] Event Name... [ When When-expression ] Output Name [ When When-expression ] Postprocess Aria Region RegionName [Equation System EquationSystemName | When WhenExpression ] Simulation Max Global Iterations {= | are | is} Number Simulation Start Time {= | are | is} Number Simulation Termination Time {= | are | is} Number Transfer Name [ When When-expression ] Use Initialize Name 70 Begin Adaptiveloop Name End Begin Sequential Name End Begin Transient Name End End Summary This block wraps a solver system for a given name. The NAME parameter is the name used to define the system. There can be more than one system block in the Solver Control Description block. The "use system NAME" line command controls which one is to be used. 4.3.1. Adapt Mesh Adapt Mesh For AdaptRegionName Using AdaptBlockName [When WhenExpression ] AdaptRegionName string – AdaptBlockName string – WhenExpression (expression) – Summary Adapt the mesh using the adaptive command block name 4.3.2. Event Event Name... [ When When-expression ] Name string... – 71 4.3.3. Output Output Name [ When When-expression ] Name string – 4.3.4. Postprocess Aria Region Postprocess Aria Region RegionName [Equation System EquationSystemName | When WhenExpression ] RegionName string – EquationSystemName string – WhenExpression (expression) – Summary Run Aria Region postprocessors. If specified, only the single equation system’s PPs are run. 4.3.5. Simulation Max Global Iterations Simulation Max Global Iterations {= | are | is} Number Number integer – Summary The Total number of Solves. 4.3.6. Simulation Start Time Simulation Start Time {= | are | is} Number Number real – Summary Simulation starting time. (by default 0.0) 72 4.3.7. Simulation Termination Time Simulation Termination Time {= | are | is} Number Parameter Value Default Number real – Summary The drop dead time. 4.3.8. Transfer Transfer Name [ When When-expression ] Parameter Value Default Name string – Summary A Solver Control Transfer line command which executes all transfers defined from the specified region. All transfers with a send region of ’name’ will be executed. 4.3.9. Use Initialize Use Initialize Name Parameter Value Default Name string – Summary This set the name of which initialization to use. 4.4. TRANSIENT Begin Transient Name Advance Name... [ When When-expression ] Adapt Mesh For AdaptRegionName Using AdaptBlockName [When WhenExpression ] Event Name... [ When When-expression ] Output Name [ When When-expression ] 73 Postprocess Aria Region RegionName [Equation System EquationSystemName | When WhenExpression ] Transfer Name [ When When-expression ] Begin Adaptiveloop Name End Begin Nonlinear Name End Begin Subcycle Name End End Summary This block is used to wrap a time loop. 4.4.1. Advance Advance Name... [ When When-expression ] Name string... – Summary Used within a Solver Control block to indicate a single step that advances the solution. 4.4.2. Adapt Mesh Adapt Mesh For AdaptRegionName Using AdaptBlockName [When WhenExpression ] AdaptRegionName string – AdaptBlockName string – WhenExpression (expression) – Summary Adapt the mesh using the adaptive command block name 74 4.4.3. Event Event Name... [ When When-expression ] Name string... – 4.4.4. Involve Name string – Summary Specify a physics participant to a coupled problem solved using matrix-free nonlinear. 4.4.5. Output Output Name [ When When-expression ] Name string – 4.4.6. Postprocess Aria Region Postprocess Aria Region RegionName [Equation System EquationSystemName | When WhenExpression ] RegionName string – EquationSystemName string – WhenExpression (expression) – Summary Run Aria Region postprocessors. If specified, only the single equation system’s PPs are run. 75 4.4.7. Transfer Transfer Name [ When When-expression ] Parameter Value Default Name string – Summary A Solver Control Transfer line command which executes all transfers defined from the specified region. All transfers with a send region of ’name’ will be executed. 4.5. NONLINEAR Begin Nonlinear Name Advance Name... [ When When-expression ] Adapt Mesh For AdaptRegionName Using AdaptBlockName [When WhenExpression ] Event Name... [ When When-expression ] Output Name [ When When-expression ] Postprocess Aria Region RegionName [Equation System EquationSystemName | When WhenExpression ] Transfer Name [ When When-expression ] Begin Subcycle Name End End Summary This block is used to wrap a nonlinear solve loop. 76 4.5.1. Advance Advance Name... [ When When-expression ] Name string... – Summary Used within a Solver Control block to indicate a single step that advances the solution. 4.5.2. Adapt Mesh Adapt Mesh For AdaptRegionName Using AdaptBlockName [When WhenExpression ] AdaptRegionName string – AdaptBlockName string – WhenExpression (expression) – Summary Adapt the mesh using the adaptive command block name 4.5.3. Event Event Name... [ When When-expression ] Name string... – 4.5.4. Involve Name string – Summary Specify a physics participant to a coupled problem solved using matrix-free nonlinear. 77 4.5.5. Output Output Name [ When When-expression ] Name string – 4.5.6. Postprocess Aria Region Postprocess Aria Region RegionName [Equation System EquationSystemName | When WhenExpression ] RegionName string – EquationSystemName string – WhenExpression (expression) – Summary Run Aria Region postprocessors. If specified, only the single equation system’s PPs are run. 4.5.7. Transfer Transfer Name [ When When-expression ] Name string – Summary A Solver Control Transfer line command which executes all transfers defined from the specified region. All transfers with a send region of ’name’ will be executed. 78 4.6. SUBCYCLE Begin Subcycle Name Advance Name... [ When When-expression ] Adapt Mesh For AdaptRegionName Using AdaptBlockName [When WhenExpression ] Event Name... [ When When-expression ] Output Name [ When When-expression ] Postprocess Aria Region RegionName [Equation System EquationSystemName | When WhenExpression ] Transfer Name [ When When-expression ] End Summary This block is used to wrap a subcycle time loop. 4.6.1. Advance Advance Name... [ When When-expression ] Name string... – Summary Used within a Solver Control block to indicate a single step that advances the solution. 4.6.2. Adapt Mesh Adapt Mesh For AdaptRegionName Using AdaptBlockName [When WhenExpression ] AdaptRegionName string – AdaptBlockName string – WhenExpression (expression) – Summary Adapt the mesh using the adaptive command block name 79 4.6.3. Event Event Name... [ When When-expression ] Name string... – 4.6.4. Involve Name string – Summary Specify a physics participant to a coupled problem solved using matrix-free nonlinear. 4.6.5. Output Output Name [ When When-expression ] Name string – 4.6.6. Postprocess Aria Region Postprocess Aria Region RegionName [Equation System EquationSystemName | When WhenExpression ] RegionName string – EquationSystemName string – WhenExpression (expression) – Summary Run Aria Region postprocessors. If specified, only the single equation system’s PPs are run. 80 4.6.7. Transfer Transfer Name [ When When-expression ] Parameter Value Default Name string – Summary A Solver Control Transfer line command which executes all transfers defined from the specified region. All transfers with a send region of ’name’ will be executed. 4.7. SEQUENTIAL Begin Sequential Name Advance Name... [ When When-expression ] Adapt Mesh For AdaptRegionName Using AdaptBlockName [When WhenExpression ] Event Name... [ When When-expression ] Output Name [ When When-expression ] Postprocess Aria Region RegionName [Equation System EquationSystemName | When WhenExpression ] Transfer Name [ When When-expression ] Begin Adaptiveloop Name End Begin Nonlinear Name End End Summary This block is used to wrap a sequential solution. It is used to wrap a sequence of Non-Linear or pseudo time solve step solves. 81 4.7.1. Advance Advance Name... [ When When-expression ] Name string... – Summary Used within a Solver Control block to indicate a single step that advances the solution. 4.7.2. Adapt Mesh Adapt Mesh For AdaptRegionName Using AdaptBlockName [When WhenExpression ] AdaptRegionName string – AdaptBlockName string – WhenExpression (expression) – Summary Adapt the mesh using the adaptive command block name 4.7.3. Event Event Name... [ When When-expression ] Name string... – 4.7.4. Involve Name string – Summary Specify a physics participant to a coupled problem solved using matrix-free nonlinear. 82 4.7.5. Output Output Name [ When When-expression ] Name string – 4.7.6. Postprocess Aria Region Postprocess Aria Region RegionName [Equation System EquationSystemName | When WhenExpression ] RegionName string – EquationSystemName string – WhenExpression (expression) – Summary Run Aria Region postprocessors. If specified, only the single equation system’s PPs are run. 4.7.7. Transfer Transfer Name [ When When-expression ] Name string – Summary A Solver Control Transfer line command which executes all transfers defined from the specified region. All transfers with a send region of ’name’ will be executed. 83 4.8. INITIALIZE Scope: Solution Control Description Begin Initialize Name Advance Name... [ When When-expression ] Event Name... [ When When-expression ] Transfer Name [ When When-expression ] End Summary This block wraps a initializer for a given name. The NAME parameter is the name used to define the initialization block. There can be more than one initialize block in the Solver Control Description block. The "use initialize NAME" line command controls which one is to be used. 4.8.1. Advance Scope: Initialize Advance Name... [ When When-expression ] Name string... – Summary Used within a Solver Control block to indicate a single step that advances the solution. 4.8.2. Event Scope: Initialize Event Name... [ When When-expression ] Name string... – Summary Used within a Solver Control block to indicate a single step that has no time associated 84 4.8.3. Involve Scope: Initialize Name string – Summary Specify a physics participant to a coupled problem solved using matrix-free nonlinear. 4.8.4. Transfer Scope: Initialize Transfer Name [ When When-expression ] Name string – Summary A Solver Control Transfer line command which executes all transfers defined from the specified region. All transfers with a send region of ’name’ will be executed. 4.9. PARAMETERS FOR Scope: Solution Control Description Begin Parameters For Converged When Convergence-expression Incremental Number Of Steps {= | are | is} Number Initial Deltat {= | are | is} Number Number Of Adaptivity Steps {= | are | is} Number Number Of Steps {= | are | is} Number Reinitialize Transient Start Time {= | are | is} Number Suppress Output From Nonlinear Loop Termination Time {= | are | is} Number Time Step Quantum {= | are | is} TimeStepQuantum Time Step Style TimeStepStyle... Total Change In Time {= | are | is} Number Begin Parameters For Aria Region RegionName 85 End End Summary A Solver Control PARAMETERS block to set up control data for the SC_type parameter. Inside this block one sets the time step parameters or nonlinear parameters. 4.9.1. Converged When Converged When Convergence-expression Parameter Value Default Convergence-expression (expression) – Summary Set the convergence expression. 4.9.2. Incremental Number Of Steps Incremental Number Of Steps {= | are | is} Number Parameter Value Default Number integer – Summary The incremental number steps to run the time for nonlinear loop. Number of time steps to run after restarting. NUMBER OF STEPS is total number of steps to run 4.9.3. Initial Deltat Initial Deltat {= | are | is} Number Parameter Value Default Number real – Summary Assign an initial delta T 86 4.9.4. Number Of Adaptivity Steps Number Of Adaptivity Steps {= | are | is} Number Parameter Value Default Number integer – Summary The number steps to run the time or nonlinear loop 4.9.5. Number Of Steps Number Of Steps {= | are | is} Number Parameter Value Default Number integer – Summary The number steps to run the time for nonlinear loop 4.9.6. Reinitialize Transient Summary Reset time and re-initialize regions each step of the adaptivity loop. 4.9.7. Start Time Start Time {= | are | is} Number Parameter Value Default Number real – Summary Assign a start time. 4.9.8. Suppress Output From Nonlinear Loop Summary Specify that the nonlinear loop will not output. Output will be handled by calls outside of the nonlinear loop (such as an additional advance region call). 87 4.9.9. Termination Time Termination Time {= | are | is} Number Parameter Value Default Number real – Summary Assign a final time to stop 4.9.10. Time Step Quantum Time Step Quantum {= | are | is} TimeStepQuantum Parameter Value Default TimeStepQuantum real – Summary Set the time stepping quantum time for SNAP style stepping. 4.9.11. Time Step Style Time Step Style TimeStepStyle... Parameter Value Default TimeStepStyle {clip | noclip | nosnap | snap} CLIP NOSNAP Summary Set the time stepping style. When CLIP is specified, the time step size will be clipped at the last step of the transient loop so that it ends at the transient loop’s end time. If clip is not specified, the last time is allowed to exceed to the transient loop’s end time and the following transient loop will start at the exceeded end time. When SNAP is specified, the time step is broken down into "quantum" time units. By default this quantum time is 12 orders of magnitude down from the difference between the start and end time for the transient loop. This value can be overridden using the TIME STEP QUANTUM line command. All time values are "snapped" to multiples of the quantum time by rounding to the nearest quantum multiple. 88 4.9.12. Total Change In Time Total Change In Time {= | are | is} Number Parameter Value Default Number real – Summary Use this number and the initial time to compute termination time. 89 90 5. TRANSFER REFERENCE 5.1. OVERVIEW Recall that Sierra Mechanics supports application data associated with nodes, elements, faces or edges of a meshed discretization as in Figure 1.3-1. The Sierra Transfer utility provides the means by which to communicate data between two Sierra application Regions. Generally speaking the same type of data is most often communicated but data movement need not be for the same type, e.g. nodal data can be communicated to element data and vice-versa. The Transfer utility is fairly flexible as it provides the ability to move data directly onto another problem domain either by direct copy or by interpolation. Analysts without prior experience with transfer are often uncertain as to which type of transfer to use. The two capabilities function exactly as their names imply but understanding which method to use requires a basic understanding of how each method works. Copy transfer assumes that the discretization for applications involved in the transfer are identical. Moreover, copy transfer also assumes that the mesh is identical so that global IDs of nodes and elements within each mesh are the same. Under these assumptions a geometric search of source to destination locations is not necessary and a simple algorithm is able to perform the data transfer in a straightforward manner. Interpolation transfer is much more general than copy transfer since it assumes only that data from one application must be geometrically mapped for use in another application. A mathematical definition of this mapping is made possible using the results from a geometric search of points on the destination mesh and their image on the sending mesh. With regard to code performance copy transfer will always more efficient than interpolation transfer but is rarely applicable in mainstream simulations. Interpolate transfer is designed to deal with complications that arise in mapping data from one application to the other and is more reliable. As a rule, one should always use interpolation transfer and not copy transfer. At the same time an analyst should strategize model construction so as to offset some of the performance costs of interpolation transfer. Even with a basic understanding of transfer users of what transfer operations should be defined. Several proper transfer source and destinations are illustrated in Figure 5.1-1, here the numbers on the figures correspond to the ExodusII global IDs of nodes or elements. Problematic transfer source and destination configurations are illustrated in Figure 5.1-2. Once again the numbers on the figures correspond to the ExodusII global IDs of nodes or elements. In using the transfer utility one must clearly define the sending region (where the data resides) and the the receiving region (the data destination). Additionally one must also specify the general geometric 91 1 5 9 13 1 5 9 13 2 6 10 14 2 6 10 14 3 7 11 15 3 7 11 15 4 8 12 16 NODE 4 8 12 16 COPY 10 1 11 12 7 4 25 12 17 6 4 9 13 2 10 23 8 11 14 8 5 7 NODE 3 2 16 15 INTERPOLATE 24 15 9 3 13 4 65 22 45 1 8 4 2 12 16 34 2 23 16 ELEMENT 3 22 13 INTERPOLATE Figure 5.1-1.. Valid Transfer Operations location of data sender and receiver based upon existing mesh entities (blocks or surfaces). Sender and receiver need not be of same topology but the source and target destinations should overlap geometrically. Clearly the definition mesh entities influences time spent in the geometric search process and should be a key consideration in model construction. The following section outlines the commands to be used in setting up transfer operations. Special attention should be paid to the syntax of the SEND command line since it differs between COPY and INTERPOLATION transfer. Since several different uses of transfer can arise and several of those examples for steady problems are included below. The same basic setup of transfer would apply to transient problems as well. A skeleton outline of one-way transfer from Region_1 to Region_2 in a steady-state problem would be: Begin Sierra . Begin Transfer my_transfer . transfer commands for first_region to second_region . End . 92 1 5 9 13 1 4 7 10 13 2 6 10 14 2 5 8 11 14 3 7 11 15 4 8 12 16 NODE 3 6 9 12 15 COPY 1 4 7 10 13 5 9 7 10 13 2 5 8 11 14 2 1 8 3 4 3 6 9 12 15 12 6 14 11 15 NODE COPY 1 4 7 1 5 6 2 5 8 2 7 9 3 6 9 8 4 3 ELEMENT COPY Figure 5.1-2.. Invalid Transfer Operation . Advance first_Region transfer my_transfer Advance second_Region End End End Begin Aria Region first_region . eq energy for temperature On block_1 using q1 with lumped_mass diff . End Begin Aria Region second_region . 93 eq energy for temperature On block_1 using q1 with xfer . End End . End Sierra 94 A skeleton outline of two-way transfer between Region_1 to Region_2 in a steady-state problem would be: Begin Sierra . Begin Transfer my_first_transfer . transfer commands for first_region to second_region . End . Begin Transfer my_second_transfer . transfer commands for second_region to first_region . End . . Advance first_Region transfer my_first_transfer Advance second_Region transfer my_second_transfer End End End Begin Aria Region first_region . eq energy for temperature On block_1 using q1 with diff eq species_3 for temperature On block_1 using q1 with xfer . End Begin Aria Region second_region . eq energy for temperature On block_1 using q1 with xfer eq species_3 for species_3 On block_1 using q1 with diff . End 95 End . End Sierra 96 Assume an input mesh for an Input_Output Region 6.1 contains a nodal variable ConvCoeff. In this case a skeleton outline for one-way transfer of ConvCoeff to to Region_2 in a steady-state problem would be: Begin Sierra . Begin Transfer my_first_transfer . transfer commands for input_output_region to second_region . SEND field hNd state none TO ConvCoeff state none . End . . Advance first_Region transfer my_first_transfer Advance second_Region End End End Begin Input_Output io_region USE FINITE ELEMENT MODEL my_input_transfer End Begin Aria Region second_region . USER FIELD REAL NODE SCALAR ConvCoeff on surface_1 . End End . End Sierra 5.2. TRANSFER Scope: Procedure 97 Begin Transfer Transfer_name Abort If Field Not Defined On Copy Transfer Send Or Receive Object Abort If Search Object Outside Of Tolerance All Fields Copy Option1 Option2 From From_region_name To To_region_name Distance Function Is Closest Receive Node To Send Centroid Exclude Ghosted From Option1 To Option2 Gauss Point Integration Order {= | are | is} Order Geometric Tolerance {= | are | is} Geometric_tolerance Inspect With File {= | are | is} File_name Ids {= | are | is} ID_list... Interpolate Option1 Option2 From From_region_name To To_region_name Interpolation Function Function_Name Nodes Outside Region {= | are | is} Option Parametric Tolerance {= | are | is} Parametric_tolerance Patch Recovery Evaluation {= | are | is} Option Search Coordinate Field Source_field_name State Option1 To Destination_field_name State Option2 Search Geometric Tolerance {= | are | is} Geometric_tolerance Search Surface Gap Tolerance {= | are | is} Surface_gap_tolerance [ Or Less ] Search Type {= | are | is} [ Option1 Option2 Option3 ] Select One Receiver For Each Send Object Select One Unique Receiver For Each Send Object Send Predefined-transfer Fields Send Block From_blocks... To To_blocks... Send Field Source_field_name State Option1 To Destination_field_name State Option2 [ Lower Bound Lower_bound Upper Bound Upper_bound ] 98 Toggle Search Warnings {= | are | is} Option Use Centroid For Geometric Proximity Begin Receive Blocks End Begin Send Blocks End End Summary Transfer region/mesh information. The mechanics/variables information will get sorted out by the calling procedure. 5.2.1. Abort If Field Not Defined On Copy Transfer Send Or Receive Object Summary For testing purposes only. Normally mesh objects in the send or receive mesh which do not have the specified field defined on them are just ignored. This line command allows the construction of tests in which it is known that every mesh object should have the specified field defined on it and to abort if that field is not found. 5.2.2. Abort If Search Object Outside Of Tolerance Summary For debugging purposes only. Abort transfer if search object lie outside of specified geometric tolerance. This command is deprecated. Use "Nodes outside region = abort" instead. 5.2.3. All Fields Summary Select all fields for transfer that have same name and state for source and destination regions. 99 5.2.4. Copy Copy Option1 Option2 From From_region_name To To_region_name Option1 {surface | volume} – Option2 {constraints | elements | nodes} – From_region_name string – To_region_name string – Summary Copy transfer elements, nodes or constraints from one region to another. The copy transfer is very specific in that the sending and receiving mesh parts must have identical global ids for every element to be copied. The copy transfer works by iterating over all the mesh objects in the receiving mesh and using the global id of the receiving mesh object to find a mesh object in the sending mesh with the same global id. The field to transfer is then copied from the sending to receiving objects. There is no interpolation and the actual coordinates of the sending and receiving objects are not used and could be very different. The copy transfer is used in very special cases where the same mesh was read into both the sending and receiving meshes, there was no element death and there was no adaptivity. In this special case, a copy transfer can be much faster than an interpolation transfer. 5.2.5. Distance Function Is Closest Receive Node To Send Centroid Summary To be used in conjunction with "SELECT ONE UNIQUE RECEIVER FOR EACH SEND OBJECT". This helped in the case where the sending and receiving element blocks did not overlap and an element transfer was using element centroids for the distance computation. The elements were very distorted so that a centroid of a surface element could be far from the surface. It was wanted that the receiving element be the one close to the surface of the block and close to the sending element in the adjacent block. Using the corner nodes was enough since it was a tet mesh with plane faces. In this particular and unusual case this alternative method of matching sending and receiving elements was useful, but it is not expected to be used often or maybe never again. 5.2.6. Exclude Ghosted Summary exclude ghosted nodes from a copy transfer 100 5.2.7. From From Option1 To Option2 Option1 {constraints | elements | nodes} – Option2 {constraints | elements | faces | – gauss_points | nodes} Summary Allows the send/receive mesh objects to be different. For a volume element field transfer (e.g. shell) to a face rank field on a surface, use ’from elements to faces’. 5.2.8. Gauss Point Integration Order Gauss Point Integration Order {= | are | is} Order Order integer – Summary Integration order to use when transferring to Gauss points. 5.2.9. Geometric Tolerance Geometric Tolerance {= | are | is} Geometric_tolerance Geometric_tolerance real – Summary This is the dimensional tolerance applied during the initial (coarse) search. If specified, all the coarse search boxes are padded by this value. The default behavior is for this to be 1e-9 times the problem bounding box size (diagonal) plus a 10 percent relative expansion per element. If a value is specified for this, it is used as a padding in place of the default with no relative box expansion. During the interpolation transfer there is a geometric search based on the coordinates of the send and receive objects. As part of this search, an axis aligned bounding box is contracted for each sending object and GEOMETRIC TOLERANCE is used to make this box bigger than just a tight bounding box. Lists of receiving points are then quickly found within these axis aligned boxes. If all points in the receiving mesh are within at least one box, no additional searching needs to be done and the search algorithm is fast. If there are still points in the receiving mesh that were outside of EVERY box, then a warning message will be issued about an 101 "expensive search for extrapolation" for these points. This ’expensive search" can be very costly if a large number of receiving objects fall into this category and this line command is provided for those special cases. 5.2.10. Inspect With File Inspect With File {= | are | is} File_name Ids {= | are | is} ID_list... File_name string – ID_list integer... – Summary STK Transfer inspection tool that allows user to output the search results for a specified list of entity ids on the receive mesh. The rank of the entities is deduced from the type of transfer being set up. 5.2.11. Interpolate Interpolate Option1 Option2 From From_region_name To To_region_name Option1 {surface | volume} – Option2 {constraints | elements | nodes} – From_region_name string – To_region_name string – Summary Interpolate will transfer elements, nodes or constraints from one mesh to another. The interpolation transfer is very general in that the field values to transfer will be interpolated from the sending to receiving mesh based on the coordinates of the sending and receiving mesh objects. Many line commands can be used to modify the behavior of the interpolation transfer but the basic algorithm is straightforward. Every mesh object in the receiving mesh is converted into a point. For elements this is the average of the nodal coordinates. An element in the sending mesh containing this point is found. If the field to transfer is nodal, the element shape functions are used to interpolate the nodal field to the receiving point. If the field to transfer is elemental, a bi-linear least squares fit based upon neighboring elements is first performed and then used to define the interpolation of the element field at the receiving point. 102 5.2.12. Interpolation Function Interpolation Function Function_Name Function_Name string – Summary Allows an application defined subroutine to be used for the interpolation. Normally the interpolation transfer will determine the best type of interpolation to use for node or element entities: Basis functions for nodal fields and a neighborhood least squares fit for element fields. This command line can be used to override the default behavior if needed. Available interpolation functions are Sum, Copy, Master_Element, Element_Centroid_Constant and Element_Centroid_Linear. Sum - Sum surrounding entity values on source and send to destination. Copy - Copy entity values directly from source to destination. Master_Element - Send interpolated nodal values from nearest source element to destination. Element_Centroid_Constant - For nodal field transfer send computed centroid value from the nearest source element to destination. For element field transfer send the element field value from the nearest source element to the destination Element_Centroid_Linear - Send values of interpolated centroid values obtained using patch recovery to destination. Patch recovery involves a least squares reconstruction and evaluation using a patch of elements surrounding the nearest source element. For linear and higher-order reconstructions one should consult the Patch Recovery Evaluation option. 5.2.13. Nodes Outside Region Nodes Outside Region {= | are | is} Option Option {abort | extrapolate | ignore | project | – truncate} Summary This line command defines what to do when a receiving point is outside the scope of the sending mesh. IGNORE - The receiving mesh object can be ignored and will receive no value. This is almost never a good idea as it can cause mesh objects just outside to have a zero value when the nodes just inside the mesh might have very large values. This can result in a discontinuous receiving field. 103 EXTRAPOLATE - This is the default behavior. The sending field is extrapolated beyond the bounds of the sending mesh. This can lead to extrapolation error, such as when a large gradient at the surface causes a negative values when only positive values are acceptable. If this happens to the upper and lower bounds that can be placed on the fields to be transferred with the SEND FIELD command. TRUNCATE - The receiving coordinate is projected back to the surface of the sending mesh to determine a value. This ensures that the receiving value is not outside of the field values in the sending mesh. PROJECT - This option is similar to TRUNCATE in which the receiving coordinate is projected back to the surface of the sending mesh to determine a value. In this case more effort is made to make sure that the projection is normal to the surface in the sending mesh. Sometimes gives a better result than Truncate but is a little more expensive to compute. If the PROJECT option is used in transferring of surface values, the sending mesh should envelope the receiving mesh. Failure to satisfy this condition will generally result in failure of the transfer. ABORT - If any receiving point is outside the sending mesh by more than the geometric tolerance, abort the simulation. Do not attempt to project, extrapolate, or otherwise handle the point. 5.2.14. Parametric Tolerance Parametric Tolerance {= | are | is} Parametric_tolerance Parametric_tolerance real – 5.2.15. Patch Recovery Evaluation Patch Recovery Evaluation {= | are | is} Option Option {linear least squares | linear moving – least squares | quadratic least squares | quadratic moving least squares} Summary This line command defines the available choices for the patch recovery and evaluation algorithm when using interpolation for element variables. The default option is Linear Least Squares. 104 5.2.16. Search Coordinate Field Search Coordinate Field Source_field_name State Option1 To Destination_field_name State Option2 Parameter Value Default Source_field_name string – Option1 {new | nm1 | nm2 | nm3 | nm4 | none | old} – Destination_field_name string – Option2 {new | nm1 | nm2 | nm3 | nm4 | none | old} – Summary Normally the interpolation transfers use the default coordinate field to determine geometry information. This line command can be used to specify an alternate field. 5.2.17. Search Geometric Tolerance Search Geometric Tolerance {= | are | is} Geometric_tolerance Parameter Value Default Geometric_tolerance real – 5.2.18. Search Surface Gap Tolerance Search Surface Gap Tolerance {= | are | is} Surface_gap_tolerance [ Or Less ] Parameter Value Default Surface_gap_tolerance real – Summary This is the dimensional tolerance applied during the initial (coarse) search. If specified, all the coarse search boxes are padded by this value. The default behavior is for this to be 1e-9 times the problem bounding box size (diagonal) plus a 10 percent relative expansion per element. If a value is specified for this, it is used as a padding in place of the default with no relative box expansion. This is a tricky parameter best ignored, let it default to something based on the problem size. During the interpolation transfer there is a geometric search based on the coordinates of the send and receive objects. As part of this search, an axis aligned bounding box is contracted for each sending object and SEARCH GAP TOLERANCE is used to make this box bigger than just a tight bounding box. Lists of receiving points are then quickly found within these axis aligned boxes. If all points in the receiving mesh are within at least one box, no additional searching needs to be done and the search algorithm is fast. If there are still points in the receiving 105 mesh that were outside of EVERY box, then a warning message will be issued about an "expensive search for extrapolation" for these points. This ’expensive search" can be very costly if a large number of receiving objects fall into this category and this line command is provided for those special cases. The OR LESS optional parameter is used when the tolerance must be set to large value for one part of the mesh but much of the mesh needs a much smaller value. In some cases it is necessary for the tolerance to be set to the actual largest surface gap tolerance which may be far too large a gap for the rest of the mesh. Setting OR LESS allows the search tolerance to be reduced in areas of the mesh thus resulting in a faster search. 5.2.19. Search Type 5.2.20. Select One Receiver For Each Send Object Summary This option will cause each sending object to be used once and only once. This will have the side effect of some receiving objects not getting any value at all. If you use this option, you will also want to set NODES OUTSIDE REGION IGNORE The example which necessitated this option was a case in which there was a delta function defined on an element in the sending mesh. It was desirable that the delta functions be summed into the receiving mesh such that the total value of the sending was conserved. It was better to have only a single element on the receiving side have a non-zero value that was the sum of sending values and not worry about how close the receiving element was to the sending element. A check that this option is working is to use Encore to computer the sum of the values of the sending and receiving fields to make sure the total sum is the same. 5.2.21. Select One Unique Receiver For Each Send Object Summary An unusual flag to get around an odd problem. Normally each receive object transfers from the nearest sending object so it is almost always the case that a send object will be used multiple times to define a receiving value. This option will cause each sending object to be used only once. This will have the side effect of some receiving objects not getting any value at all. If you use this option, you will also want to set 106 NODES OUTSIDE REGION IGNORE or else the uniqueness will be lost for nodes outside the sending region. The example which necessitated this option was a case in which there was a delta function defined on an element in the sending mesh. It was desirable that the delta function be defined on the receiving mesh for only a single element in the neighborhood of the sending element. The analysis was more sensitive to the number of delta functions on the receiving side than the location. So it was better to have only a single element on the receiving side have a non-zero value and not worry about how close the receiving element was to the sending element. 5.2.22. Send Send Predefined-transfer Fields Parameter Value Default Predefined-transfer {} – Summary Use predefined transfer semantics provided by the specified name. 5.2.23. Send Block Send Block From_blocks... To To_blocks... Parameter Value Default From_blocks string... – To_blocks string... – Summary Add element blocks to a particular same mesh element copy transfer operator. The copy transfer can have multiple of these lines to define many blocks, but each line sends a single block to a single block: SEND BLOCK block_1 TO block_1 SEND BLOCK block_101 TO block_101 The interpolation transfer can have only a single SEND BLOCK line, but can define many from/to blocks: SEND BLOCK block_3 block_5 block_6 TO block_3 block_5 107 5.2.24. Send Field Send Field Source_field_name State Option1 To Destination_field_name State Option2 [ Lower Bound Lower_bound Upper Bound Upper_bound ] Parameter Value Default Source_field_name string – Option1 {new | nm1 | nm2 | nm3 | nm4 | none | old} – Destination_field_name string – Option2 {new | nm1 | nm2 | nm3 | nm4 | none | old} – Summary Specifies the mapping between source and destination field names. Vector and tensor fields can be subscripted using parenthesis and 1’s based or brackets and 0 based. Notes on subscripting: • Does not work for COPY transfers, only INTERPOLATION type transfers. • If the field name itself actually contains either parenthesis or brackets then we are in trouble and an error is going to be thrown due to a syntax error in index specification. • Only a single subscript is allowed so vectors of vectors or higher order tensors can not use double subscripts. But it should be possible to determine the correct offset within the field and pick out the correct value with a little effort. • Once subscripted, only a single value will be transferred. It is not possible to transfer multiple values starting at a certain index, instead multiple line commands must be used, as shown above. • The indexes can be 0 based with brackets or 1 based when using parenthesis. Although this could be very confusing if mixed within a single line command. • Both the from and to fields can be subscripted independently on the same line. example SEND FIELD velocity TO velocity SEND FIELD temp TO temperature lower bound 0 SEND FIELD x TO y lower bound 10 upper bound 100 SEND FIELD A(2) TO B(3) lower bound 10 upper bound 100 SEND FIELD A[1] TO B[2] lower bound 10 upper bound 100 108 5.2.25. Toggle Search Warnings Toggle Search Warnings {= | are | is} Option Option {false | no | off | on | true | yes} – Summary Specify whether warnings about entities outside of the search domain should be printed. The default behavior is to always print these warning messages. 5.2.26. Use Centroid For Geometric Proximity Summary STK Transfer option to trigger the use of centroid based proximity comparison for selecting the best interpolating entity when receive entities lie outside of the domain. Default geometric proximity comparison is based on geometric projection which is more expensive but accurate. The use of this option is a way to reduce computational cost especially for meshes that are fairly regular. However, there is no guarantee of accuracy. 109 110 6. INPUT OUTPUT REGION REFERENCE 6.1. INPUT_OUTPUT REGION OVERVIEW For some coupled simulations one can approximate part of the problem physics independent of the entire problem physics. In order to facilitate this type of loose application coupling the Sierra Framework provides the ability to input datasets that include the output of other simulations. An application can then make requests of information from these datasets. In fullfilling these requests, data can be extracted from these datasets and be copied or interpolated to another problem domain. Moreover these requests can be satisfied by data interpolated through time. The mechanism provided to achieve this end goal is known as the Input_Output Region and its usage is described in what follows. The input_output region works in tandem with transfer 5.1 and solution control 4. Here transfer carries out the communication of data and solution control provides synchronization of the data transfer. Note that just like other Sierra Regions the input_output region must have its own Finite Element model command block defined. As an example, let us assume that an input mesh for an Input_Output Region contains a nodal variable ConvCoeff that we wish to use in another Region. In this case an outline for one-way transfer of ConvCoeff to to a Region, second_region, in a steady-state problem would be: Begin Sierra . Begin Finite Element Model input_transfer . End . Begin Transfer my_first_transfer . transfer commands for input_output_region to second_region . SEND field hNd state none TO ConvCoeff state none . End . . 111 Advance io_region transfer my_first_transfer Advance second_Region End End End Begin Input_Output io_region USE FINITE ELEMENT MODEL my_input_transfer End Begin Aria Region second_region . use Finite Element Model input_transfer . USER FIELD REAL NODE SCALAR ConvCoeff on surface_1 . End End . End Sierra 6.2. INPUT_OUTPUT REGION Scope: Procedure Begin Input_Output Region Parameter_block_name Create Element Field Field_name Of Type Option And Dimension Dimension [ Value {= | are | is} Number... ] Create Nodal Field Field_name Of Type Option And Dimension Dimension [ Value {= | are | is} Number... ] Fixed Time [ {= | are | is} Fixed_time ] Offset Time {= | are | is} Period_offset_time Periodicity Time {= | are | is} Periodicity_time Start Time {= | are | is} Start_time Time Interpolation Method {= | are | is} Method 112 Timestep Adjustment Interval {= | are | is} Nsteps Use Finite Element Model ModelName [ Model Coordinates Are Nodal_variable_name ] Begin Heartbeat Label End Begin Heartbeat Output Label End Begin History Output Label End Begin Restart Data Label End Begin Results Output Label End End Summary An input output region can load data from an exodus file and transfer it to other regions, either constant in time or varying. It can be used for setting initial conditions from prior data, defining boundary conditions, or any other applications of custom field values. For output it can be used to visualize data on a reduced set of the domain - for example one could transfer to an input output region that had a 2D slice through a 3D domain, or a coarser 3D mesh. 113 6.2.1. Create Element Field Create Element Field Field_name Of Type Option And Dimension Dimension [ Value {= | are | is} Number... ] Parameter Value Default Field_name string – Option {asym_tensor_03 | complex | – full_tensor_22 | full_tensor_36 | integer | long_integer | matrix_22 | matrix_33 | real | sym_tensor_21 | sym_tensor_31 | sym_tensor_33 | unsigned_integer | unsigned_integer_64 | vector_2d | vector_3d} Dimension integer – Summary Creates a Element Field name field_name on the region. 6.2.2. Create Nodal Field Create Nodal Field Field_name Of Type Option And Dimension Dimension [ Value {= | are | is} Number... ] Parameter Value Default Field_name string – Option {asym_tensor_03 | complex | – full_tensor_22 | full_tensor_36 | integer | long_integer | matrix_22 | matrix_33 | real | sym_tensor_21 | sym_tensor_31 | sym_tensor_33 | unsigned_integer | unsigned_integer_64 | vector_2d | vector_3d} Dimension integer – Summary Creates a Nodal Field name field_name on the region. 6.2.3. Fixed Time 114 Summary The line specifies that the database will be read for a single, fixed time. Specifying the actual time is optional. If the time is not specified, the final time plane in the database will be read. NOTE: • This option takes precedence over the periodic specifications given by START TIME, PERIODICITY TIME, and OFFSET TIME. if FIXED TIME is specified then if FIXED TIME value is given then (eg., FIXED TIME is 1.) DATABASE TIME = FIXED TIME else (eg., FIXED TIME) DATABASE TIME = last time in database else if PERIODICITY TIME greater than 0 then if APPLICATION TIME less than or equal to START TIME then DATABASE TIME = APPLICATION TIME else DATABASE TIME = START TIME + (APPLICATION TIME - START TIME) modulo PERIODICITY TIME else DATABASE TIME = APPLICATION TIME now add OFFSET TIME to the computed DATABASE TIME 6.2.4. Offset Time Offset Time {= | are | is} Period_offset_time Parameter Value Default Period_offset_time real – Summary This value is added to the application time to determine what database time slice to input. If OFFSET TIME were 15 than at application time 0 database time slice 15 would be read from the file and used for the initial values. At application time 1, database time slice 16 would be read. NOTES: • The OFFSET TIME is added in after the START TIME and PERIODICITY TIME are used. • The FIXED TIME option takes precedence over this option. 115 if FIXED TIME is specified then if FIXED TIME value is given then (eg., FIXED TIME is 1.) DATABASE TIME = FIXED TIME else (eg., FIXED TIME) DATABASE TIME = last time in database else if PERIODICITY TIME greater than 0 then if APPLICATION TIME less than or equal to START TIME then DATABASE TIME = APPLICATION TIME else DATABASE TIME = START TIME + (APPLICATION TIME - START TIME) modulo PERIODICITY TIME else DATABASE TIME = APPLICATION TIME now add OFFSET TIME to the computed DATABASE TIME 6.2.5. Periodicity Time Periodicity Time {= | are | is} Periodicity_time Parameter Value Default Periodicity_time real – Summary START TIME and PERIODICITY TIME taken together give the time frame from the input database to use to initialize the application values. If START TIME is 25 and PERIODICITY TIME is 10, then time slices from 25 to 35 will be used over and over again as the application time runs from 0 to whatever. In general DATABASE TIME is (APPLICATION TIME - START TIME) modulo PERIODICITY T after the application time reaches the START TIME. NOTES: • The OFFSET TIME is added in after the START TIME and PERIODICITY TIME are used. • The FIXED TIME option takes precedence over this option. if FIXED TIME is specified then if FIXED TIME value is given then (eg., FIXED TIME is 1.) DATABASE TIME = FIXED TIME else (eg., FIXED TIME) DATABASE TIME = last time in database 116 else if PERIODICITY TIME greater than 0 then if APPLICATION TIME less than or equal to START TIME then DATABASE TIME = APPLICATION TIME else DATABASE TIME = START TIME + (APPLICATION TIME - START TIME) modulo PERIODICITY TIME else DATABASE TIME = APPLICATION TIME now add OFFSET TIME to the computed DATABASE TIME 6.2.6. Start Time Start Time {= | are | is} Start_time Parameter Value Default Start_time real – Summary The time in which to start applying PERIODICITY TIME. If PERIODICITY TIME is not specified then START TIME is ignored. NOTES: • The OFFSET TIME is added in after the START TIME and PERIODICITY TIME are used. • The FIXED TIME option takes precedence over this option. if FIXED TIME is specified then if FIXED TIME value is given then (eg., FIXED TIME is 1.) DATABASE TIME = FIXED TIME else (eg., FIXED TIME) DATABASE TIME = last time in database else if PERIODICITY TIME greater than 0 then if APPLICATION TIME less than or equal to START TIME then DATABASE TIME = APPLICATION TIME else DATABASE TIME = START TIME + (APPLICATION TIME - START TIME) modulo PERIODICITY TIME else DATABASE TIME = APPLICATION TIME now add OFFSET TIME to the computed DATABASE TIME 117 6.2.7. Time Interpolation Method Time Interpolation Method {= | are | is} Method Method {closest | linear} – Summary This line specifies how interpolation in time in the database will be handled. If linear (the default option) is specified, quantities at a given point are linearly interpolated from the bounding known time points. If the closest option is selected, then the closest known time point will be taken. 6.2.8. Timestep Adjustment Interval Timestep Adjustment Interval {= | are | is} Nsteps Nsteps integer – Summary Specify the number of steps to ’look ahead’ and adjust the timestep to ensure that the specified output times or simulation end time will be hit ’exactly’. 6.2.9. Use Finite Element Model Use Finite Element Model ModelName [ Model Coordinates Are Nodal_variable_name ] ModelName string – Summary Associates a predefined finite element model with this region. 118 7. EXAMPLES Sierra application code couplings with Arpeggio can be carried out in a variety of ways. In this chapter a few simple problems are used to demonstrate some of the coupling approaches. Here we note that success in performing the coupling hinges upon defining a proper setup for each of the application codes participating in the coupling. Understandably the coupling becomes more straightforward if one begins by first setting up each of the independent application code problems (i.e. an application Region) and later unites the Regions under Arpeggio. The purpose of the examples is simply to demonstrate the basics of how the problem setup will differ for various use cases. The examples given here illustrate the use cases most likely to occur: • One-way coupling of TF with Adagio from file on same mesh 7.1, • One-way coupling of TF with Adagio from file on different mesh 7.2, • One-way coupling of TF with Adagio on same mesh using transfer 7.3, • Two-way coupling of TF with Adagio on same mesh 7.4, • Two-way coupling of TF with Adagio on same mesh including element death7.5, • One way coupling of TF with another TF, same mesh 7.7, • One way coupling of TF with Presto on same mesh with subcycling 7.6, Analysts are often confused by the usage of Field states, i.e. state new, state old, state none in transfer operations. Generally speaking, state new and state old are associated with solution variables whereas state none refers to variables that have a single state. The use of Field states is illustrated in various examples which follow. During a solution step, state old often serves as a reference or initial state and state new is simply the latest solution. In a thermal-mechanical problem the temperature increment can be represented using both state old and state new. Hence initial assignment of state old and state new to the same temperature defines an initial state with no temperature increment. Multi-state variables are often managed by an application by swapping the location of the state. As an example, advance of the solution from one solution to the next step is accomplished by changing the reference of the previous solution to state old and the current solution to state new. Hence assignment of the same value to state old and state new in a transient simulation would mean that the variable is unchanged even though time may be advancing. This is the desired behavior when performing subcycling, variables at the outer scope are unchanged for the solution step while solution variables at the inner scope continue to evolve during the subcycle. 119 7.1. ONE-WAY COUPLING FROM FILE In many problems of coupled physics one of the physics (primary) is dependent upon the other physics (secondary) but not vice-versa. In this case the coupling is considered to be one-way and can be accomplished simply by supplying a secondary physics solution to the primary physics simulation. In the context of problem solutions one would first solve the secondary physics problem and then communicate the solution to a primary physics simulation. Perhaps the easiest way to carry out such a simulation is to supply the secondary physics solution to the primary physics via file. The following example describes the process as it might be carried out in Arpeggio. 7.1.1. Problem Statement Consider a one-way coupled thermal structural analysis problem in which a body is free to expand as a response to gradual temperature change in time. Although the problem geometry is changing due to the structural deformation, the geometry change is assumed to have minimal effect upon heat transfer in the body. For each time step, a heat conduction problem was solved for the temperature distribution using the Aria code and the results were written to file. The Aria output file is then used as the input file for Adagio where the temperatures are read into Adagio. Adagio subsequently solves for mechanical equilibrium which includes calculation of thermal strains due to changing temperatures. Here we note that the thermal solution file time planes need not correspond to the Adagio time planes as the thermal solution will be interpolated in time to match the Adagio solution time. Furthermore, in this problem, an Aria results file is the Adagio input discretization so the problems correspond to the same mesh. Here it is important that the input Aria discretization contain the nodesets and sidesets needed to carry out the Adagio simulation. Problems in which one might wish to solve the Adagio problem on a different discretization can also be dealt with but in a slightly different manner. 7.1.2. Input File begin sierra barOneWayCouple begin function analytic_sigma_zz type is analytic evaluate expression = "lambda=5.769231e5; mu = 3.84615e5; Delta = 25; alpha = 1e-4; -((3*lam end begin function THERMAL_STRAIN ordinate is strain abscissa is temperature 200.0 0.0 400.0 0.02 120 end function THERMAL_STRAIN define direction x with vector 1.0 0.0 0.0 define direction y with vector 0.0 1.0 0.0 define direction z with vector 0.0 0.0 1.0 begin material linear_elastic density = 0.1 thermal log strain function = THERMAL_STRAIN youngs modulus = 1.e6 poissons ratio = 0.3 begin parameters for model elastic_plastic youngs modulus = 1.e6 poissons ratio = 0.3 yield stress = 1.0e6 hardening modulus = 10.0 beta is 0.999999 end parameters for model elastic_plastic end material linear_elastic begin finite element model mesh_arpeggio Database Name = 3dbar_temp.g Database Type = exodusII model = elastic_plastic end finite element model mesh_arpeggio begin procedure Arpeggio_Procedure $========================================================= $ Add in solver control parameters $========================================================= begin transient mytransient 121 end transient mytransient begin parameters for transient mytransient start time = 0.0 termination time = 2.0 Number of Steps = 2 time increment = 1.0 end end $========================================================= $ End ofsolver control parameters $========================================================= # coupling is one_way using temperature distribution from file $============================================================ $ Define the Adagio region $============================================================ begin adagio region adagio use finite element model mesh_arpeggio begin user output compute global analytic_sigma_zz as function analytic_sigma_zz compute global sigma_zz as max of element stress(zz) compute at every step end begin solution verification skip times = 0.0 to 1.0 completion file = VerifSigmaZZ verify global sigma_zz = function analytic_sigma_zz tolerance = 1 end begin prescribed temperature 122 read variable = temperature end ### definition of BCs ### surface = surface_10 components = z surface = surface_20 components = z ### ------------------### ### Solver definition ### ### ------------------### Begin cg Target relative Residual = 1.0e-11 Maximum Iterations = 30 Minimum Iterations = 1 begin full tangent preconditioner automatic smoothing factor = 0.1 end end end ### output description ### begin Results Output output_adagio Database Name = barOneWayCoupleFromFile_mech.e Database Type = exodusII At Step 0, Increment = 1 nodal Variables = temperature as temperature nodal Variables = velocity as vel nodal Variables = displacement as displ element Variables = stress as stress global Variables = timestep as TIMESTEP global variables = external_energy as ExternalEnergy global variables = internal_energy as InternalEnergy global variables = kinetic_energy as KineticEnergy global variables = momentum as Momentum end adagio region adagio 123 end procedure Arpeggio_Procedure end sierra barOneWayCouple 7.2. ONE-WAY COUPLING USING TRANSFER FROM DIFFERENT MESH In some coupled physics one of the physics (primary) is dependent upon the other physics (secondary) but not vice-versa. In this case the coupling is considered to be one-way and can be accomplished simply by supplying a secondary physics solution to the primary physics simulation. In the context of problem solutions one would first solve the secondary physics problem and then communicate the solution to a primary physics simulation. As previously demonstrated one way to carry out such a simulation is to supply the secondary physics solution to the primary physics via file 7.1. However, in some cases the secondary physics solution is available on a vastly different geometry. In this case the secondary physics solution must be interpolated onto the primary physics as needed. In Sierra Mechanics the communication step of such an analysis is carried out using Solution Control and Transfer operations. Here Transfer describes the information and Solution Control ensures sequencing of information to the primary physics. The following example describes the solution process to perform a coupled analysis using a precomputed thermal solution and Adagio. 7.2.1. Problem Statement Consider a one-way coupled thermal structural analysis problem in which a body is free to expand as a response to gradual temperature change in time. Althought the problem geometry is changing due to the structural deformation, the geometry change is assumed to have minimal effect upon heat transfer in the body. For this situation a reasonable approach may be to precompute the heat transfer solution and then supply it to the mechanical simulation. Here a transient heat conduction problem on a full geometry was solved for the temperature distribution using the Aria code and the results were saved to file. Later on the previously computed temperature distribution was supplied to Adagio for solution of mechanical equilibrium which includes calculation of thermal strains due to changing temperatures. In this particular case the Adagio problem could be solved by invoking symmetry conditions so the model geometry is a subset of the thermal model geometry. In this particular case the Adagio problem could be solved by invoking symmetry conditions so the model geometry is a subset of the thermal model geometry. During the simulation the transient thermal solution is read from file these results are then communicated to Adagio using a transfer operation. Once the Aria values are received by Adagio the structural problem is then solved. Since the thermal and structural model geometries are different, it is necessary to use the transfer INTERPOLATE operation. Note that the problem advances with the two applications lock stepped in time with the thermal solution is being interpolated in both space and time. 124 7.2.2. Input File # # Temperature to the thermal stress problem is transferred to Adagio from # a file containing the temperature-time history on a dissimilar mesh. # Here the Adagio problem exploits quarter-section symmetry. # begin sierra barOneWayCouple begin function THERMAL_STRAIN ordinate is strain abscissa is temperature 200.0 0.0 400.0 0.02 end function THERMAL_STRAIN define direction x with vector 1.0 0.0 0.0 define direction y with vector 0.0 1.0 0.0 define direction z with vector 0.0 0.0 1.0 begin material linear_elastic density = 0.1 thermal engineering strain function = THERMAL_STRAIN youngs modulus = 1.e6 poissons ratio = 0.3 begin parameters for model elastic_plastic youngs modulus = 1.e6 poissons ratio = 0.3 yield stress = 1.0e6 hardening modulus = 10.0 beta is 0.999999 end parameters for model elastic_plastic end material linear_elastic begin finite element model mesh_arpeggio Database Name = quarter_model.g Database Type = exodusII 125 model = elastic_plastic end finite element model mesh_arpeggio begin finite element model input_transfer_temperature Database Name = full_model_temperature.e Database Type = exodusII end finite element model input_transfer_temperature begin procedure Arpeggio_Procedure $========================================================= $ Add in solver control parameters $========================================================= begin transient mytransient advance transfer_input_region transfer temperature_to_adagio advance adagioRegion end transient mytransient begin parameters for transient mytransient start time = 0.0 termination time = 2.0 Number of Steps = 2 begin parameters for adagio region adagioRegion time increment = 1.0 end end $========================================================= $ End ofsolver control parameters $========================================================= # coupling is one_way using temperature distribution from IORegion # here the temperature history is obtained from a dissimilar mesh 126 begin transfer temperature_to_adagio interpolate volume NODES FROM transfer_input_region TO adagioRegion SEND BLOCK block_1 TO block_1 SEND field T state none TO Temperature state none parametric TOLERANCE IS 0.01 geometric TOLERANCE IS 0.01 NODES OUTSIDE REGION IS IGNORE end transfer temperature_to_adagio $============================================================ $ Define the Adagio region $============================================================ begin adagio region adagioRegion use finite element model mesh_arpeggio ### definition of BCs ### surface = surface_10 components = z surface = surface_20 components = z # symmetry conditions surface = surface_40 components = x surface = surface_30 components = y Begin Initial temperature magnitude=273. End 127 ### ------------------### ### Solver definition ### ### ------------------### Begin cg Target relative Residual = 1.0e-11 Maximum Iterations = 30 Minimum Iterations = 1 begin full tangent preconditioner automatic smoothing factor = 0.1 end end end ### output description ### begin Results Output output_adagio Database Name = barOneWayCoupleFromDifferentMesh.e Database Type = exodusII At Step 0, Increment = 1 nodal Variables = temperature as temperature nodal Variables = velocity as vel nodal Variables = displacement as displ element Variables = stress as stress global Variables = timestep as TIMESTEP global variables = external_energy as ExternalEnergy global variables = internal_energy as InternalEnergy global variables = kinetic_energy as KineticEnergy global variables = momentum as Momentum end adagio region adagioRegion $============================================================ $ Define the IORegion $============================================================ BEGIN INPUT_OUTPUT REGION Transfer_Input_Region USE FINITE ELEMENT MODEL input_transfer_temperature END INPUT_OUTPUT REGION Transfer_Input_Region end procedure Arpeggio_Procedure end sierra barOneWayCouple 128 7.3. ONE-WAY COUPLING USING TRANSFER In many problems of coupled physics one of the physics (primary) is dependent upon the other physics (secondary) but not vice-versa. In this case the coupling is considered to be one-way and can be accomplished simply by supplying a secondary physics solution to the primary physics simulation. In the context of problem solutions one would first solve the secondary physics problem and then communicate the solution to a primary physics simulation. One way to carry out such a simulation is to supply the secondary physics solution to the primary physics via file 7.1. However, in many instances it is more convenient to carry out both simulations simultaneously and directly communicate the secondary physics solution to the primary physics as needed. In Sierra Mechanics the communication step of such an analysis is carried out using Solution Control and Transfer operations. Here Transfer describes the information and Solution Control ensures sequencing of information to the primary physics. The following example describes the solution process to perform a coupled analysis using Aria and Adagio. 7.3.1. Problem Statement Consider a one-way coupled thermal structural analysis problem in which a body is free to expand as a response to gradual temperature change in time. Althought the problem geometry is changing due to the structural deformation, the geometry change is assumed to have minimal effect upon heat transfer in the body. For each time step, a heat conduction problem was solved for the temperature distribution using the Aria code. Once the thermal solution has been obtained the temperature solution is communicated to Adagio via Transfer and Adagio then solves for mechanical equilibrium which includes calculation of thermal strains due to changing temperatures. Note that the problem advances with the two applications lock stepped in time. In this problem the Aria input discretization is identical to that of Adagio. During the simulation an Aria solution is performed and Aria results are then communicated to Adagio using a transfer COPY operation. Once the Aria values are received by Adagio the structural problem is then solved. Problems in which one might wish to solve the Aria and Adagio problems on different discretizations can dealt with by making simple modifications to the input replacing the transfer COPY operation with a INTERPOLATE operation. 7.3.2. Input File begin sierra barOneWayCouple begin function analytic_sigma_zz type is analytic evaluate expression = "lambda=5.769231e5; mu = 3.84615e5; Delta = 25; alpha = 1e-4; -((3*lam end begin function THERMAL_STRAIN 129 ordinate is strain abscissa is temperature 200.0 0.0 400.0 0.02 end function THERMAL_STRAIN define direction x with vector 1.0 0.0 0.0 define direction y with vector 0.0 1.0 0.0 define direction z with vector 0.0 0.0 1.0 begin material linear_elastic density = 0.1 thermal log strain function = THERMAL_STRAIN youngs modulus = 1.e6 poissons ratio = 0.3 begin parameters for model elastic_plastic youngs modulus = 1.e6 poissons ratio = 0.3 yield stress = 1.0e6 hardening modulus = 10.0 beta is 0.999999 end parameters for model elastic_plastic end material linear_elastic BEGIN TPETRA EQUATION SOLVER solve_temperature BEGIN CG SOLVER BEGIN JACOBI PRECONDITIONER END MAXIMUM ITERATIONS = 1000 CONVERGENCE TOLERANCE = 1.0e-06 RESIDUAL SCALING = r0 END begin ARIA MATERIAL linear_elastic thermal conductivity = constant k=401.0 specific heat = constant cp=385 density = constant rho=0.1 heat conduction = Fouriers_law end ARIA MATERIAL linear_elastic 130 begin finite element model mesh_arpeggio Database Name = 3dbar.g Database Type = exodusII model = elastic_plastic end finite element model mesh_arpeggio begin procedure Arpeggio_Procedure $========================================================= $ Add in solver control parameters $========================================================= begin initialize mytransient_init advance AriaRegion transfer TariatoTadagio_init end initialize mytransient_init use initialize mytransient_init begin transient mytransient advance AriaRegion transfer TariatoTadagio end transient mytransient begin parameters for transient mytransient start time = 0.0 termination time = 2.0 Number of Steps = 2 BEGIN PARAMETERS FOR ARIA REGION AriaRegion Initial Time Step Size = 1.0 Time Step Variation = Fixed END time increment = 1.0 131 end end $========================================================= $ End ofsolver control parameters $========================================================= #coupling type is one_way begin transfer TariatoTadagio_init copy volume nodes from AriaRegion to adagio send field solution->TEMPERATURE state new to temperature state old send field solution->TEMPERATURE state new to temperature state new end transfer TariatoTadagio_init begin transfer TariatoTadagio copy volume nodes from AriaRegion to adagio send field solution->TEMPERATURE state new to temperature state new end transfer TariatoTadagio $============================================================ $ Define the Aria region $============================================================ BEGIN ARIA REGION AriaRegion use linear solver solve_temperature nonlinear solution strategy = newton maximum nonlinear iterations = 10 nonlinear residual tolerance = 1.0e-6 nonlinear relaxation factor = 1.0 use finite element model mesh_arpeggio IC const at block_1 Temperature = 273.0 BC const dirichlet at nodelist_20 Temperature = 273.0 BC const dirichlet at nodelist_10 Temperature = 373.0 EQ ENERGY for TEMPERATURE on block_1 using Q1 with DIFF ### output description ### BEGIN RESULTS OUTPUT LABEL diffusion output database Name = barOneWayCoupleTransfer_therm.e 132 at step 0, increment = 1 title Aria cube test nodal variables = solution->TEMPERATURE as Temperature END RESULTS OUTPUT LABEL diffusion output END ARIA REGION AriaRegion $============================================================ $ Define the Adagio region $============================================================ begin adagio region adagio use finite element model mesh_arpeggio begin user output compute global analytic_sigma_zz as function analytic_sigma_zz compute global sigma_zz as max of element stress(zz) compute at every step end begin solution verification skip times = 0.0 to 1.0 completion file = VerifSigmaZZ verify global sigma_zz = function analytic_sigma_zz tolerance = 1 end ### definition of BCs ### surface = surface_10 components = z surface = surface_20 components = z ### ------------------### ### Solver definition ### ### ------------------### Begin cg Target relative Residual = 1.0e-11 133 Maximum Iterations = 30 Minimum Iterations = 1 begin full tangent preconditioner automatic smoothing factor = 0.1 end end end ### output description ### begin Results Output output_adagio Database Name = barOneWayCoupleTransfer_mech.e Database Type = exodusII At Step 0, Increment = 1 nodal Variables = temperature as temperature nodal Variables = velocity as vel nodal Variables = displacement as displ element Variables = stress as stress global Variables = timestep as TIMESTEP global variables = external_energy as ExternalEnergy global variables = internal_energy as InternalEnergy global variables = kinetic_energy as KineticEnergy global variables = momentum as Momentum end adagio region adagio end procedure Arpeggio_Procedure end sierra barOneWayCouple 7.4. TWO-WAY COUPLING WITH TRANSFER 7.4.1. Problem Statement This is a test of solving a simple one-dimensional thermal diffusion problem with Dirichlet BCs. The test problem is shown schematically in Figure. Although the problem is one-dimensional we solve the problem in a three-dimensional setting. Once the diffusion problem has been solved numerically the temperature result is postprocessed to obtain a comparison with the analytical result and the distribution of diffusive heat flux. This test input also demonstrates the use tabular function and Encore function material property specification in Aria. 7.4.2. Input File # Converted: gapClosure.i using c2a 0.12 134 begin sierra gapClosure begin definition for function THERMAL_STRAIN ordinate = strain abscissa = temperature 373.0 0.0 1373.0 12e-3 end definition for function THERMAL_STRAIN begin definition for function tempLeft 0.0 373 0.5 373 1.0 773 7.0 773 7.5 373 10.0 373 end end begin definition for function tempRight 0.0 373 1.0 373 2.0 623 3.0 873 4.0 1123 5.0 873 6.0 623 7.0 373 10.0 373 end end define direction x with vector 1.0 0.0 0.0 define direction y with vector 0.0 1.0 0.0 define direction z with vector 0.0 0.0 1.0 begin Aria material linear_elastic specific heat = constant cp = 1.0 density = constant rho = 0.1 heat conduction = basic 135 thermal conductivity = constant k = 250 end begin property specification for material linear_elastic density = 0.1 thermal log strain function = THERMAL_STRAIN youngs modulus = 1.e7 poissons ratio = 0.0 end property specification for material linear_elastic begin finite element model mesh_calagio Database Name = gapClosure.g model = elastic begin parameters for block block_2 model = elastic end parameters for block block_2 end finite element model mesh_calagio BEGIN TPETRA EQUATION SOLVER solve_temperature BEGIN gmres SOLVER BEGIN JACOBI PRECONDITIONER END MAXIMUM ITERATIONS = 1000 CONVERGENCE TOLERANCE = 1.0e-07 RESIDUAL SCALING = None END begin procedure Acca_Procedure Begin Initialize sys_init transfer aria_to_adagio_init End initialize sys_init 136 use Initialize sys_init Simulation Start Time = 0.0 Simulation Termination Time = 8.0 Begin Transient p1 Begin Nonlinear converge_step_p1 End Nonlinear converge_step_p1 End Transient p1 Begin Transient p2 Begin Nonlinear converge_step_p2 End Nonlinear converge_step_p2 End Transient p2 Begin Transient p3 Begin Nonlinear converge_step_p3 ### advance_aria_and_adagio End Nonlinear converge_step_p3 End Transient p3 Begin Transient p4 Begin Nonlinear converge_step_p4 End Nonlinear converge_step_p4 End Transient p4 Begin Transient p5 Begin Nonlinear converge_step_p5 End Nonlinear converge_step_p5 End Transient p5 End System Main 137 Begin parameters for nonlinear converge_step_p1 converged when "(aria.MaxResidualNorm(0.0) < 1.e-6 &&\$ adagio.norm(0.0) < 1.e-6) || CURRENT_STEP > 10" # many other choices in documentation End parameters for nonlinear converge_step_p1 Begin parameters for nonlinear converge_step_p2 converged when "(aria.MaxResidualNorm(0.0) < 1.e-6 &&\$ adagio.norm(0.0) < 1.e-6) || CURRENT_STEP > 10" # many other choices in documentation End parameters for nonlinear converge_step_p2 Begin parameters for nonlinear converge_step_p3 converged when "(aria.MaxResidualNorm(0.0) < 1.e-6 &&\$ adagio.norm(0.0) < 1.e-6) || CURRENT_STEP > 10" # many other choices in documentation End parameters for nonlinear converge_step_p3 Begin parameters for nonlinear converge_step_p4 converged when "(aria.MaxResidualNorm(0.0) < 1.e-6 &&\$ adagio.norm(0.0) < 1.e-6) || CURRENT_STEP > 10" # many other choices in documentation End parameters for nonlinear converge_step_p4 Begin parameters for nonlinear converge_step_p5 converged when "(aria.MaxResidualNorm(0.0) < 1.e-6 &&\$ adagio.norm(0.0) < 1.e-6) || CURRENT_STEP > 10" # many other choices in documentation End parameters for nonlinear converge_step_p5 begin parameters for transient p1 start time = 0.0 termination time = 1.0 initial time step size = 0.5 time step variation = fixed time integration method = first_order time increment = 1.0 end 138 begin parameters for transient p2 start time = 1.0 termination time = 3.0 initial time step size = 1.0 time step variation = fixed time integration method = first_order time increment = 1.0 end begin parameters for transient p3 start time = 3.0 termination time = 4.0 initial time step size = 0.5 time step variation = fixed time integration method = first_order time increment = 0.5 end begin parameters for transient p4 start time = 4.0 termination time = 7.0 initial time step size = 1.0 time step variation = fixed time integration method = first_order 139 time increment = 1.0 end begin parameters for transient p5 start time = 7.0 termination time = 8.0 initial time step size = 0.5 time step variation = fixed time integration method = first_order time increment = 0.5 end # default to: standard_aria_to_adagio_init begin transfer aria_to_adagio_init copy volume nodes from aria to adagio send field solution->temperature state old to temperature state old send field solution->temperature state new to temperature state new end # default to: standard_aria_to_adagio_advance begin transfer aria_to_adagio copy volume nodes from aria to adagio send field solution->temperature state new to temperature state new end begin transfer adagio_to_aria copy volume nodes from adagio to aria send field displacement state new to solution->mesh_displacements state new end begin aria region Aria EQ energy for temperature on all_blocks using Q1 with diff mass # src EQ mesh for mesh_displacements on all_blocks using Q1 with xfer 140 Begin Initial Condition myICBlock temperature = 373.0 add volume block_1 add volume block_2 $can have many of the above and below extent specifiers $nodeset = nodelist_1 $surface = surface_1 End Initial Condition myICBlock begin temperature boundary condition surface_14 add surface surface_14 temperature time function = tempLeft end begin temperature boundary condition surface_22 add surface surface_22 temperature time function = tempRight end begin contact definition mary contact surface surf_12 contains surface_12 contact surface surf_24 contains surface_24 begin interaction harry surfaces are surf_24 surf_12 normal tolerance = 10e-5 end begin enforcement larry gap conductance coefficient = constant value = 2000 enforcement for energy = gap_conductance end update search every 1 steps end use finite element model mesh_calagio model coordinates are model_coordinates nonlinear solution strategy = newton use dof averaged nonlinear residual accept solution after maximum nonlinear iterations = true use linear solver solve_temperature end Aria region Aria begin adagio region adagio use finite element model mesh_calagio model coordinates are model_coordinates ### output description ### begin Results Output output_adagio Database Name = gapClosure.e At Step 0, Increment = 1 141 nodal Variables = temperature as temp nodal Variables = velocity as vel nodal Variables = displacement as displ element Variables = stress as stress global Variables = timestep as timestep global variables = external_energy as ExternalEnergy global variables = internal_energy as InternalEnergy global variables = kinetic_energy as KineticEnergy global variables = momentum as Momentum ### definition of BCs ### surface = surface_14 surface_22 components = x surface = surface_11 surface_21 components = y # Contact begin contact definition search = dash contact surface surf_12 contains surface_12 contact surface surf_24 contains surface_24 begin interaction side A = surf_24 side B = surf_12 normal tolerance = 2.5e-4 end enforcement = al end ### ------------------### ### Solver definition ### ### ------------------### begin Loadstep predictor type = secant slip scale factor = 0.0 end 142 begin control contact target relative residual = 1.0e-7 target residual = 1.0e-12 maximum iterations = 10 minimum iterations = 0 end Begin cg target relative residual = 1.0e-10 target residual = 1.0e-15 Maximum Iterations = 500 Minimum Iterations = 0 end end end adagio region adagio end procedure Acca_Procedure end sierra gapClosure 7.5. TWO-WAY COUPLING USING TRANSFER OF ELEMENT DEATH In some problems of coupled physics the effect of one physics is manifest in substantial change of material in the other physics. One example of this is extreme softening leading to removal of material owing to an elevated temperature limit. In these instances it is convenient to carry out both simulations simultaneously and directly exchange solutions between the two physics and including a variable denoting material softening/removal (element death). In Sierra Mechanics the communication step of such an analysis is carried out using Solution Control and Transfer operations. Here Transfer describes the exchanged information and Solution Control ensures sequencing of information between the two physics. The following example describes the solution process to perform a coupled analysis using Aria and Adagio in which elements are removed during the simulation. 7.5.1. Problem Statement Consider a two-way coupled thermal structural analysis problem in which a body is free to expand as a response to gradual temperature change in time. The problem geometry is changing due to the structural deformation, and the geometry change is communicated back to the thermal problem. For each time step, a heat conduction problem was solved for the temperature distribution using the Aria code. Based upon the temperature solution, elements exceeding a threshold temperature are marked for removal with a death_status variable (1 active, 0 inactive). Once the thermal solution has completed the 143 temperature solution and the death_status are communicated to Adagio via Transfer and Adagio then solves for mechanical equilibrium including calculation of thermal strains due to changing temperatures but on a domain that no longer contains elements marked for death by Aria. The displacement solution is then communicated back to Aria where subsequent solves are carried out in the current configuration. It is important to note that each application has its own element death implementation. The exchange of death_status from one application to another is made indirectly to an intermediate variable that interfaces the other application’s element death implementation. Here the problem solution advances with the two applications lock stepped in time. In this problem the Aria input discretization is identical to that of Adagio. During the simulation an Aria solution is performed and Aria results are then communicated to Adagio using a transfer COPY operation. Once the Aria values are received by Adagio the structural problem is then solved and the entire process is then repeated. When mechanical deformation is small one might consider removal of the Aria mesh equation (EQ mesh) and ignoring the transfer of displacement back to Aria. Situations in which element death is driven by the mechanics solution are modeled in a similar manner as demonstrated here. Variants of this problems in which one might wish to solve the Aria and Adagio problems on different discretizations are usually much more difficult. A good starting point would be to make simple modifications to the input replacing the transfer COPY operation with a INTERPOLATE operation. 7.5.2. Input File begin sierra elemDeathSimpleAria begin function disp_x type = analytic expression variable: t = global time evaluate expression = "t*0.0002" end function disp_x begin Aria material linear_elastic specific heat = constant cp = 385.0 density = constant rho = 0.1 heat conduction = basic thermal conductivity = constant k = 1.0 end begin property specification for material linear_elastic density = 0.1 youngs modulus = 1.e6 poissons ratio = 0.0 end property specification for material linear_elastic 144 begin finite element model mesh_arpgio Database Name = elemDeathSimpleAria.g Database Type = exodusII begin parameters for block block_1 block_2 solid mechanics use model elastic end parameters for block block_1 block_2 end finite element model mesh_arpgio BEGIN TPETRA EQUATION SOLVER solve_temperature BEGIN CG SOLVER BEGIN JACOBI PRECONDITIONER END MAXIMUM ITERATIONS = 1000 CONVERGENCE TOLERANCE = 1.0e-07 RESIDUAL SCALING = None END begin procedure Arpgio_Procedure begin initialize two_way_couple_init Advance Aria Transfer Aria_Adagio_Initial transfer aria_adagio_element Advance Adagio end initialize two_way_couple_init use initialize two_way_couple_init simulation termination time = 20.0 begin transient two_way_couple advance Aria transfer aria_adagio_element transfer Aria_adagio_default transfer adagio_Aria_default end transient two_way_couple begin parameters for transient two_way_couple 145 start time = 0.0 initial time step size = 1.0 time step variation = fixed time increment = 1.0 end parameters for transient two_way_couple begin aria region Aria EQ energy for temperature on all_blocks using Q1 with diff mass # lumped_mass mass src EQ mesh for mesh_displacements on all_blocks using Q1 with xfer Begin Initial Condition myICBlock temperature = 273.0 add volume all_blocks End Initial Condition myICBlock begin temperature boundary condition surface_1 add surface surface_1 temperature = 273.0 end begin temperature boundary condition surface_2 add surface surface_2 temperature = 373.0 end use finite element model mesh_arpgio begin element death temperature_death add volume block_2 criterion is max nodal value of solution->temperature > 280.0 end element death temperature_death nonlinear solution strategy = newton use dof averaged nonlinear residual accept solution after maximum nonlinear iterations = true use linear solver solve_temperature 146 begin results output output_aria Database Name = elemDeathSimpleAria_aria.e Database Type = exodusII At Step 0, Increment = 1 nodal variables = solution->temperature nodal variables = solution->mesh_displacements as displacement element variables = death_status as death_status global Variables = timestep end results output output_aria end Aria region Aria begin adagio region adagio use finite element model mesh_arpgio ### output description ### begin results output output_adagio Database Name = elemDeathSimpleAria_adagio.e Database Type = exodusII At Step 0, Increment = 1 nodal Variables = temperature nodal Variables = velocity nodal Variables = displacement element Variables = stress element variables = death_status element variables = aria_death_status global Variables = timestep global variables = external_energy global variables = internal_energy global variables = kinetic_energy global variables = momentum ### definition of BCs ### node set = nodelist_1 components = x y z begin prescribed displacement node set = nodelist_2 component = x function = disp_x end prescribed displacement begin user variable aria_death_status 147 type = element real length = 1 initial value = 1.0 end user variable aria_death_status begin element death elemdeath block = block_2 criterion is element value of aria_death_status = 0.0 death method = deactivate element death steps = 1 end element death elemdeath ### ------------------### ### Solver definition ### ### ------------------### Begin cg target relative residual = 1.0e-9 Maximum Iterations = 50 Minimum Iterations = 0 preconditioner = probe end end end adagio region adagio begin transfer Aria_adagio_initial copy volume nodes from Aria to adagio send field solution->temperature state new to temperature state old send field solution->temperature state new to temperature state new end transfer Aria_adagio_initial begin transfer Aria_adagio_default copy volume nodes from Aria to adagio send field solution->temperature state new to temperature state new end transfer Aria_adagio_default begin transfer adagio_Aria_default copy volume nodes from adagio to Aria send field displacement state new to solution->mesh_displacements state new end transfer adagio_Aria_default begin transfer aria_adagio_element copy volume elements from Aria to adagio Send field death_status state none to aria_death_status state none end transfer aria_adagio_element 148 end procedure Arpgio_Procedure end sierra elemDeathSimpleAria 7.6. ESTACK REGRESSION TEST 7.6.1. Problem Statement This is a test of solving a simple one-dimensional thermal diffusion problem with Dirichlet BCs. The test problem is shown schematically in Figure. Although the problem is one-dimensional we solve the problem in a three-dimensional setting. Once the diffusion problem has been solved numerically the temperature result is postprocessed to obtain a comparison with the analytical result and the distribution of diffusive heat flux. This test input also demonstrates the use tabular function and Encore function material property specification in Aria. 7.6.2. Input File # Begin Sierra Aria_Presto_example Define Direction Y_Axis With Vector 0.0 1.0 0.0 Begin Definition For Function Delta Type is Piecewise Linear Ordinate is Displacement Abscissa is Time Begin Values 0.000E+00 0.00000E+00 1.400E-04 1.40098E-03 2.800E-04 2.80392E-03 4.200E-04 4.20883E-03 5.600E-04 5.61571E-03 7.000E-04 7.02456E-03 8.400E-04 8.43538E-03 9.800E-04 9.84818E-03 1.120E-03 1.12630E-02 1.260E-03 1.26797E-02 1.400E-03 1.40985E-02 1.540E-03 1.55192E-02 1.680E-03 1.69419E-02 1.820E-03 1.83666E-02 1.960E-03 1.97933E-02 2.100E-03 2.12221E-02 2.240E-03 2.26528E-02 149 2.380E-03 2.40855E-02 2.520E-03 2.55202E-02 2.660E-03 2.69569E-02 2.800E-03 2.83957E-02 2.940E-03 2.98364E-02 3.080E-03 3.12792E-02 3.220E-03 3.27240E-02 3.360E-03 3.41709E-02 3.500E-03 3.56197E-02 3.640E-03 3.70706E-02 3.780E-03 3.85235E-02 3.920E-03 3.99785E-02 4.060E-03 4.14354E-02 4.200E-03 4.28945E-02 4.340E-03 4.43556E-02 4.480E-03 4.58187E-02 4.620E-03 4.72838E-02 4.760E-03 4.87511E-02 4.900E-03 5.02204E-02 5.040E-03 5.16917E-02 5.180E-03 5.31651E-02 5.320E-03 5.46406E-02 5.460E-03 5.61181E-02 5.600E-03 5.75977E-02 5.740E-03 5.90794E-02 5.880E-03 6.05631E-02 6.020E-03 6.20489E-02 6.160E-03 6.35368E-02 6.300E-03 6.50268E-02 6.440E-03 6.65189E-02 6.580E-03 6.80131E-02 6.720E-03 6.95094E-02 6.860E-03 7.10077E-02 7.000E-03 7.25082E-02 7.140E-03 7.40107E-02 7.280E-03 7.55154E-02 7.420E-03 7.70222E-02 7.560E-03 7.85311E-02 7.700E-03 8.00421E-02 7.840E-03 8.15552E-02 7.980E-03 8.30704E-02 8.120E-03 8.45878E-02 8.260E-03 8.61073E-02 8.400E-03 8.76289E-02 8.540E-03 8.91526E-02 8.680E-03 9.06785E-02 8.820E-03 9.22065E-02 150 8.960E-03 9.37367E-02 9.100E-03 9.52690E-02 9.240E-03 9.68035E-02 9.380E-03 9.83401E-02 9.520E-03 9.98788E-02 9.660E-03 1.01420E-01 9.800E-03 1.02963E-01 9.940E-03 1.04508E-01 1.008E-02 1.06055E-01 1.022E-02 1.07605E-01 1.036E-02 1.09157E-01 1.050E-02 1.10711E-01 1.064E-02 1.12267E-01 1.078E-02 1.13825E-01 1.092E-02 1.15385E-01 1.106E-02 1.16948E-01 1.120E-02 1.18513E-01 1.134E-02 1.20080E-01 1.148E-02 1.21649E-01 1.162E-02 1.23220E-01 1.176E-02 1.24794E-01 1.190E-02 1.26370E-01 1.204E-02 1.27948E-01 1.218E-02 1.29528E-01 1.232E-02 1.31111E-01 1.246E-02 1.32695E-01 1.260E-02 1.34282E-01 1.274E-02 1.35871E-01 1.288E-02 1.37463E-01 1.302E-02 1.39056E-01 1.316E-02 1.40652E-01 1.330E-02 1.42250E-01 1.344E-02 1.43850E-01 1.358E-02 1.45453E-01 1.372E-02 1.47058E-01 1.386E-02 1.48665E-01 1.400E-02 1.50274E-01 1.414E-02 1.51885E-01 1.428E-02 1.53499E-01 1.442E-02 1.55115E-01 1.456E-02 1.56733E-01 1.470E-02 1.58354E-01 1.484E-02 1.59977E-01 1.498E-02 1.61602E-01 1.512E-02 1.63229E-01 1.526E-02 1.64859E-01 1.540E-02 1.66491E-01 151 1.554E-02 1.68125E-01 1.568E-02 1.69762E-01 1.582E-02 1.71400E-01 1.596E-02 1.73042E-01 1.610E-02 1.74685E-01 1.624E-02 1.76331E-01 1.638E-02 1.77979E-01 1.652E-02 1.79629E-01 1.666E-02 1.81282E-01 1.680E-02 1.82937E-01 1.694E-02 1.84594E-01 1.708E-02 1.86253E-01 1.722E-02 1.87915E-01 1.736E-02 1.89580E-01 1.750E-02 1.91246E-01 1.764E-02 1.92915E-01 1.778E-02 1.94586E-01 1.792E-02 1.96260E-01 1.806E-02 1.97936E-01 1.820E-02 1.99614E-01 1.834E-02 2.01295E-01 1.848E-02 2.02978E-01 1.862E-02 2.04663E-01 1.876E-02 2.06351E-01 1.890E-02 2.08041E-01 1.904E-02 2.09733E-01 1.918E-02 2.11428E-01 1.932E-02 2.13125E-01 1.946E-02 2.14825E-01 1.960E-02 2.16527E-01 1.974E-02 2.18231E-01 1.988E-02 2.19938E-01 2.002E-02 2.21647E-01 2.016E-02 2.23359E-01 2.030E-02 2.25072E-01 2.044E-02 2.26789E-01 2.058E-02 2.28507E-01 2.072E-02 2.30229E-01 2.086E-02 2.31952E-01 2.100E-02 2.33678E-01 2.114E-02 2.35406E-01 2.128E-02 2.37137E-01 2.142E-02 2.38870E-01 2.156E-02 2.40606E-01 2.170E-02 2.42344E-01 2.184E-02 2.44085E-01 2.198E-02 2.45828E-01 152 2.212E-02 2.47573E-01 2.226E-02 2.49321E-01 2.240E-02 2.51071E-01 2.254E-02 2.52824E-01 2.268E-02 2.54579E-01 2.282E-02 2.56337E-01 2.296E-02 2.58097E-01 2.310E-02 2.59859E-01 2.324E-02 2.61624E-01 2.338E-02 2.63392E-01 2.352E-02 2.65162E-01 2.366E-02 2.66934E-01 2.380E-02 2.68709E-01 2.394E-02 2.70487E-01 2.408E-02 2.72267E-01 2.422E-02 2.74049E-01 2.436E-02 2.75834E-01 2.450E-02 2.77621E-01 2.464E-02 2.79411E-01 2.478E-02 2.81204E-01 2.492E-02 2.82999E-01 2.506E-02 2.84796E-01 2.520E-02 2.86596E-01 2.534E-02 2.88399E-01 2.548E-02 2.90204E-01 2.562E-02 2.92011E-01 2.576E-02 2.93821E-01 2.590E-02 2.95634E-01 2.604E-02 2.97449E-01 2.618E-02 2.99267E-01 2.632E-02 3.01087E-01 2.646E-02 3.02910E-01 2.660E-02 3.04735E-01 2.674E-02 3.06563E-01 2.688E-02 3.08393E-01 2.702E-02 3.10226E-01 2.716E-02 3.12062E-01 2.730E-02 3.13900E-01 2.744E-02 3.15741E-01 2.758E-02 3.17584E-01 2.772E-02 3.19430E-01 2.786E-02 3.21279E-01 2.800E-02 3.23130E-01 2.814E-02 3.24983E-01 2.828E-02 3.26840E-01 2.842E-02 3.28699E-01 2.856E-02 3.30560E-01 153 2.870E-02 3.32424E-01 2.884E-02 3.34291E-01 2.898E-02 3.36160E-01 2.912E-02 3.38032E-01 2.926E-02 3.39907E-01 2.940E-02 3.41784E-01 2.954E-02 3.43664E-01 2.968E-02 3.45546E-01 2.982E-02 3.47431E-01 2.996E-02 3.49319E-01 3.010E-02 3.51209E-01 3.024E-02 3.53102E-01 3.038E-02 3.54998E-01 3.052E-02 3.56896E-01 3.066E-02 3.58797E-01 3.080E-02 3.60701E-01 3.094E-02 3.62607E-01 3.108E-02 3.64516E-01 3.122E-02 3.66428E-01 3.136E-02 3.68342E-01 3.150E-02 3.70259E-01 3.164E-02 3.72179E-01 3.178E-02 3.74101E-01 3.192E-02 3.76027E-01 3.206E-02 3.77954E-01 3.220E-02 3.79885E-01 3.234E-02 3.81818E-01 3.248E-02 3.83754E-01 3.262E-02 3.85692E-01 3.276E-02 3.87634E-01 3.290E-02 3.89578E-01 3.304E-02 3.91525E-01 3.318E-02 3.93474E-01 3.332E-02 3.95426E-01 3.346E-02 3.97381E-01 3.360E-02 3.99339E-01 3.374E-02 4.01299E-01 3.388E-02 4.03263E-01 3.402E-02 4.05229E-01 3.416E-02 4.07197E-01 3.430E-02 4.09169E-01 3.444E-02 4.11143E-01 3.458E-02 4.13120E-01 3.472E-02 4.15100E-01 3.486E-02 4.17082E-01 3.500E-02 4.19068E-01 3.514E-02 4.21056E-01 154 3.528E-02 4.23047E-01 3.542E-02 4.25040E-01 3.556E-02 4.27037E-01 3.570E-02 4.29036E-01 3.584E-02 4.31038E-01 3.598E-02 4.33043E-01 3.612E-02 4.35050E-01 3.626E-02 4.37061E-01 3.640E-02 4.39074E-01 3.654E-02 4.41090E-01 3.668E-02 4.43109E-01 3.682E-02 4.45131E-01 3.696E-02 4.47156E-01 3.710E-02 4.49183E-01 3.724E-02 4.51213E-01 3.738E-02 4.53246E-01 3.752E-02 4.55282E-01 3.766E-02 4.57321E-01 3.780E-02 4.59363E-01 3.794E-02 4.61407E-01 3.808E-02 4.63455E-01 3.822E-02 4.65505E-01 3.836E-02 4.67558E-01 3.850E-02 4.69614E-01 3.864E-02 4.71673E-01 3.878E-02 4.73735E-01 3.892E-02 4.75800E-01 3.906E-02 4.77867E-01 3.920E-02 4.79938E-01 3.934E-02 4.82011E-01 3.948E-02 4.84087E-01 3.962E-02 4.86167E-01 3.976E-02 4.88249E-01 3.990E-02 4.90334E-01 4.004E-02 4.92422E-01 4.018E-02 4.94512E-01 4.032E-02 4.96606E-01 4.046E-02 4.98703E-01 4.060E-02 5.00803E-01 4.074E-02 5.02905E-01 4.088E-02 5.05011E-01 4.102E-02 5.07119E-01 4.116E-02 5.09231E-01 4.130E-02 5.11345E-01 4.144E-02 5.13462E-01 4.158E-02 5.15583E-01 4.172E-02 5.17706E-01 155 4.186E-02 5.19832E-01 4.200E-02 5.21962E-01 4.214E-02 5.24094E-01 4.228E-02 5.26229E-01 4.242E-02 5.28367E-01 4.256E-02 5.30508E-01 4.270E-02 5.32653E-01 4.284E-02 5.34800E-01 4.298E-02 5.36950E-01 4.312E-02 5.39103E-01 4.326E-02 5.41260E-01 4.340E-02 5.43419E-01 4.354E-02 5.45581E-01 4.368E-02 5.47746E-01 4.382E-02 5.49915E-01 4.396E-02 5.52086E-01 4.410E-02 5.54261E-01 4.424E-02 5.56438E-01 4.438E-02 5.58619E-01 4.452E-02 5.60802E-01 4.466E-02 5.62989E-01 4.480E-02 5.65179E-01 4.494E-02 5.67371E-01 4.508E-02 5.69567E-01 4.522E-02 5.71766E-01 4.536E-02 5.73968E-01 4.550E-02 5.76173E-01 4.564E-02 5.78382E-01 4.578E-02 5.80593E-01 4.592E-02 5.82807E-01 4.606E-02 5.85025E-01 4.620E-02 5.87245E-01 4.634E-02 5.89469E-01 4.648E-02 5.91696E-01 4.662E-02 5.93926E-01 4.676E-02 5.96159E-01 4.690E-02 5.98395E-01 4.704E-02 6.00634E-01 4.718E-02 6.02877E-01 4.732E-02 6.05122E-01 4.746E-02 6.07371E-01 4.760E-02 6.09623E-01 4.774E-02 6.11878E-01 4.788E-02 6.14136E-01 4.802E-02 6.16398E-01 4.816E-02 6.18662E-01 4.830E-02 6.20930E-01 156 4.844E-02 6.23201E-01 4.858E-02 6.25475E-01 4.872E-02 6.27752E-01 4.886E-02 6.30033E-01 4.900E-02 6.32316E-01 4.914E-02 6.34603E-01 4.928E-02 6.36893E-01 4.942E-02 6.39186E-01 4.956E-02 6.41483E-01 4.970E-02 6.43783E-01 4.984E-02 6.46085E-01 4.998E-02 6.48392E-01 5.012E-02 6.50701E-01 5.026E-02 6.53014E-01 5.040E-02 6.55329E-01 5.054E-02 6.57648E-01 5.068E-02 6.59971E-01 5.082E-02 6.62296E-01 5.096E-02 6.64625E-01 5.110E-02 6.66957E-01 5.124E-02 6.69293E-01 5.138E-02 6.71631E-01 5.152E-02 6.73973E-01 5.166E-02 6.76318E-01 5.180E-02 6.78667E-01 5.194E-02 6.81019E-01 5.208E-02 6.83374E-01 5.222E-02 6.85732E-01 5.236E-02 6.88094E-01 5.250E-02 6.90459E-01 5.264E-02 6.92827E-01 5.278E-02 6.95199E-01 5.292E-02 6.97574E-01 5.306E-02 6.99952E-01 5.320E-02 7.02334E-01 5.334E-02 7.04719E-01 5.348E-02 7.07107E-01 5.362E-02 7.09498E-01 5.376E-02 7.11893E-01 5.390E-02 7.14292E-01 5.404E-02 7.16693E-01 5.418E-02 7.19098E-01 5.432E-02 7.21507E-01 5.446E-02 7.23919E-01 5.460E-02 7.26334E-01 5.474E-02 7.28752E-01 5.488E-02 7.31174E-01 157 5.502E-02 7.33600E-01 5.516E-02 7.36028E-01 5.530E-02 7.38461E-01 5.544E-02 7.40896E-01 5.558E-02 7.43335E-01 5.572E-02 7.45777E-01 5.586E-02 7.48223E-01 5.600E-02 7.50673E-01 5.614E-02 7.53125E-01 5.628E-02 7.55581E-01 5.642E-02 7.58041E-01 5.656E-02 7.60504E-01 5.670E-02 7.62970E-01 5.684E-02 7.65440E-01 5.698E-02 7.67913E-01 5.712E-02 7.70390E-01 5.726E-02 7.72871E-01 5.740E-02 7.75354E-01 5.754E-02 7.77842E-01 5.768E-02 7.80332E-01 5.782E-02 7.82826E-01 5.796E-02 7.85324E-01 5.810E-02 7.87825E-01 5.824E-02 7.90330E-01 5.838E-02 7.92838E-01 5.852E-02 7.95350E-01 5.866E-02 7.97865E-01 5.880E-02 8.00384E-01 5.894E-02 8.02906E-01 5.908E-02 8.05432E-01 5.922E-02 8.07962E-01 5.936E-02 8.10494E-01 5.950E-02 8.13031E-01 5.964E-02 8.15571E-01 5.978E-02 8.18115E-01 5.992E-02 8.20662E-01 6.006E-02 8.23212E-01 6.020E-02 8.25767E-01 6.034E-02 8.28325E-01 6.048E-02 8.30886E-01 6.062E-02 8.33451E-01 6.076E-02 8.36020E-01 6.090E-02 8.38592E-01 6.104E-02 8.41168E-01 6.118E-02 8.43747E-01 6.132E-02 8.46330E-01 6.146E-02 8.48917E-01 158 6.160E-02 8.51507E-01 6.174E-02 8.54101E-01 6.188E-02 8.56699E-01 6.202E-02 8.59300E-01 6.216E-02 8.61905E-01 6.230E-02 8.64513E-01 6.244E-02 8.67125E-01 6.258E-02 8.69741E-01 6.272E-02 8.72361E-01 6.286E-02 8.74984E-01 6.300E-02 8.77611E-01 6.314E-02 8.80241E-01 6.328E-02 8.82875E-01 6.342E-02 8.85513E-01 6.356E-02 8.88155E-01 6.370E-02 8.90800E-01 6.384E-02 8.93449E-01 6.398E-02 8.96102E-01 6.412E-02 8.98758E-01 6.426E-02 9.01418E-01 6.440E-02 9.04082E-01 6.454E-02 9.06750E-01 6.468E-02 9.09421E-01 6.482E-02 9.12096E-01 6.496E-02 9.14775E-01 6.510E-02 9.17457E-01 6.524E-02 9.20144E-01 6.538E-02 9.22834E-01 6.552E-02 9.25528E-01 6.566E-02 9.28225E-01 6.580E-02 9.30927E-01 6.594E-02 9.33632E-01 6.608E-02 9.36341E-01 6.622E-02 9.39054E-01 6.636E-02 9.41770E-01 6.650E-02 9.44491E-01 6.664E-02 9.47215E-01 6.678E-02 9.49943E-01 6.692E-02 9.52675E-01 6.706E-02 9.55410E-01 6.720E-02 9.58150E-01 6.734E-02 9.60893E-01 6.748E-02 9.63640E-01 6.762E-02 9.66391E-01 6.776E-02 9.69146E-01 6.790E-02 9.71905E-01 6.804E-02 9.74667E-01 159 6.818E-02 9.77434E-01 6.832E-02 9.80204E-01 6.846E-02 9.82978E-01 6.860E-02 9.85757E-01 6.874E-02 9.88539E-01 6.888E-02 9.91325E-01 6.902E-02 9.94114E-01 6.916E-02 9.96908E-01 6.930E-02 9.99706E-01 6.944E-02 1.00251E+00 6.958E-02 1.00531E+00 6.972E-02 1.00812E+00 6.986E-02 1.01094E+00 7.000E-02 1.01375E+00 End Values End Definition For Function Delta begin definition for function TEMPERATURE ordinate is temperature abscissa is time 0.00 1255.4 1.00 1255.4 end definition for function TEMPERATURE begin definition for function THERMAL_STRAIN ordinate is strain abscissa is temperature 200.0 0.0 3000.0 0.0 end definition for function THERMAL_STRAIN Begin Property Specification For Material Resistor density = 8.0E-4 thermal log strain function = THERMAL_STRAIN youngs modulus = 200.0E3 $ MPa poissons ratio = 0.305 End End Begin Aria Material Resistor 160 Electric Displacement = Basic Electrical Permittivity = Constant Kappa=3e-10 # N/V^2 Electrical Resistivity = Constant Rho=2e2 # Ohm-mm End Begin Property Specification For Material Metal density = 8.0E-4 thermal log strain function = THERMAL_STRAIN youngs modulus = 200.0E3 $ MPa poissons ratio = 0.305 End End Begin Aria Material Metal Electric Displacement = Basic Electrical Permittivity = Constant Kappa=1e-9 # N/V^2 Electrical Resistivity = Constant Rho=2e-5 # Ohm-mm End begin property specification for material dielectric density = 8.0E-4 thermal log strain function = THERMAL_STRAIN youngs modulus = 200.0E3 $ MPa poissons ratio = 0.305 end end Begin Aria Material Dielectric Electric Displacement = Basic Electrical Permittivity = Constant Kappa=3e-11 # N/V^2 Electrical Resistivity = Constant Rho=1e13 # Ohm-mm End begin finite element model mesh1 Database Name = estack.g Database Type = exodusII material resistor model = elastic hourglass stiffness = 0.05 hourglass viscosity = 0.03 161 begin parameters for block block_2 material metal model = elastic hourglass stiffness = 0.05 hourglass viscosity = 0.03 end parameters for block block_2 begin parameters for block block_3 material dielectric model = elastic hourglass stiffness = 0.05 hourglass viscosity = 0.03 end parameters for block block_3 begin parameters for block block_4 material metal model = elastic hourglass stiffness = 0.05 hourglass viscosity = 0.03 end parameters for block block_4 end finite element model mesh1 BEGIN TPETRA EQUATION SOLVER DIRECT_SOLVER BEGIN SUPERLU SOLVER END Begin Procedure The_Procedure # Define solution control for advancing and transferring forces and # displacements between Aria and Adagio Begin Transient MyTransient1 Transfer Presto_to_Aria Advance AriaRegion Begin Subcycle PrestoSubcycle Advance PrestoRegion End End Begin Transient MyTransient2 Transfer Presto_to_Aria Advance AriaRegion Begin Subcycle PrestoSubcycle2 162 Advance PrestoRegion End End simulation termination time = 0.07 End # Time-stepping parameters Begin Parameters for Transient MyTransient1 Start Time = 0.0 time step style noclip Termination Time = 0.035 # Parameters for Aria region: fluid mechanics region Begin Parameters for Aria Region AriaRegion Initial Time Step Size = 1e-3 Time Step Variation = Fixed End # Parameters for Adagio region: solid mechanics region Begin Parameters for Presto Region PrestoRegion initial time step = 1.0e-6 time step scale factor = 1.0 time step increase factor = 10. step interval = 1 print banner interval for solution control = 10 End End Begin Parameters for Transient MyTransient2 Start Time = 0.035 time step style noclip Termination Time = 0.07 # Parameters for Aria region: fluid mechanics region Begin Parameters for Aria Region AriaRegion Initial Time Step Size = 1e-3 Time Step Variation = Fixed End # Parameters for Adagio region: solid mechanics region Begin Parameters for Presto Region PrestoRegion time step scale factor = 1.0 time step increase factor = 10. step interval = 5 print banner interval for solution control = 5 End End End Solution Control Description Begin Transfer Presto_to_Aria 163 Copy Volume Nodes from PrestoRegion to AriaRegion Send Field displacement State New to Solution->Mesh_Displacements State New End Begin Presto Region PrestoRegion use finite element model mesh1 Begin Prescribed Temperature function = TEMPERATURE scale factor = 1.0 End ### output description ### begin Results Output output_presto Database Name = estack_presto.e Database Type = exodusII At Time 0.0, Increment = 1.0E-4 nodal Variables = force_external as f_ext nodal Variables = force_internal as f_int nodal Variables = velocity as vel nodal Variables = acceleration as accl nodal Variables = displacement as displ nodal Variables = temperature as temp element Variables = stress as stress global Variables = timestep as timestep global variables = external_energy as ExternalEnergy global variables = internal_energy as InternalEnergy global variables = kinetic_energy as KineticEnergy global variables = momentum as Momentum End ### definition of boundary conditions ### node set = nodelist_1 components = x End node set = nodelist_2 components = z End node set = nodelist_3 components = y 164 End Begin Prescribed Displacement Node set = nodelist_4 direction = y_axis function = delta scale factor = 1.0 End End Presto Region PrestoRegion Begin Aria Region AriaRegion Use Finite Element Model mesh1 Use Linear Solver Direct_Solver Nonlinear Solution Strategy = Newton Maximum Nonlinear Iterations = 10 Nonlinear Residual Tolerance = 1e-6 EQ Voltage for Voltage on all_blocks using Q1 with DIFF IC const at block_1 Voltage = 5 BC const dirichlet at nodelist_3 Voltage = 0 # ground BC const dirichlet at nodelist_4 Voltage = 10 # prescribed EQ Mesh for Mesh_Displacements on all_blocks using Q1 with XFER Begin Results Output The_Output Database Name = estack_aria.e At step 0, increment = 1 Title Aria-Presto electro-mechanical coupling Nodal Variables = solution->Voltage as V Nodal Variables = solution->Mesh_Displacements as Disp End End Aria Region AriaRegion End Procedure The_Procedure End Sierra Aria_Presto_example 165 7.7. TV REGRESSION TEST 7.7.1. Problem Statement This is a test of solving a simple one-dimensional thermal diffusion problem with Dirichlet BCs. The test problem is shown schematically in Figure. Although the problem is one-dimensional we solve the problem in a three-dimensional setting. Once the diffusion problem has been solved numerically the temperature result is postprocessed to obtain a comparison with the analytical result and the distribution of diffusive heat flux. This test input also demonstrates the use tabular function and Encore function material property specification in Aria. 7.7.2. Input File # # Begin Sierra Slump_Test Begin Aria Material Bar Thermal Conductivity = Constant k = 1.0 Electrical Conductivity = Constant sigma = 1.0 heat conduction = fouriers_law current density = ohms_law End BEGIN TPETRA EQUATION SOLVER DIRECT_SOLVER BEGIN UMFPACK SOLVER END Begin Finite Element Model The_Model Database Name = mesh3d.g Begin Parameters For Block block_1 Material Bar End End Begin Procedure The_Procedure begin sequential mysolveblk advance Voltage_Region 166 transfer VtoT advance Temperature_Region end begin transfer VtoT copy volume nodes from Voltage_Region to Temperature_Region send field solution->VOLTAGE state old to solution->VOLTAGE state new end transfer VtoT Begin Aria Region Voltage_Region Use Finite Element Model The_Model Use Linear Solver Direct_Solver Nonlinear Solution Strategy = Newton Maximum Nonlinear Iterations = 10 Nonlinear Residual Tolerance = 1.0e-6 Nonlinear Relaxation Factor = 1.0 # Sideset 1 : x = x_min # Sideset 2 : x = x_max # Sideset 3 : y = y_min # Sideset 4 : y = y_max # Sideset 5 : z = z_min # Sideset 6 : z = z_max # Sideset 10: y and z surfaces EQ Current For Voltage On Block_1 Using Q1 With Diff IC Const on block_1 Voltage = 0.0 BC Const Dirichlet at surface_1 Voltage = 10 # fat/low BC Const Dirichlet at surface_2 Voltage = 0 # Thin/High End Begin Results Output Label V_Output Database Name = v.e At Step 0, Increment is 1 Title TV Test - V Region Nodal Variables = solution->VOLTAGE as V End End Begin Aria Region Temperature_Region 167 Use Finite Element Model The_Model Use Linear Solver Direct_Solver Nonlinear Solution Strategy = Newton Maximum Nonlinear Iterations = 10 Nonlinear Residual Tolerance = 1.0e-6 Nonlinear Relaxation Factor = 1.0 # Sideset 1 : x = x_min # Sideset 2 : x = x_max # Sideset 3 : y = y_min # Sideset 4 : y = y_max # Sideset 5 : z = z_min # Sideset 6 : z = z_max # Sideset 10: y and z surfaces EQ Current For Voltage On Block_1 Using Q1 With xfer EQ Energy For Temperature On Block_1 Using Q1 With Diff Src IC Const on block_1 Temperature = 298 BC Flux for Energy on surface_3 = Nat_Conv H = 20 T_ref = 298 BC Flux for Energy on surface_4 = Nat_Conv H = 0.2 T_ref = 298 Source for Energy on Block_1 = Joule_Heating Begin Results Output Label T_Output Database Name = tv.e At Step 0, Increment is 1 Title TV Test - T Region Nodal Variables = solution->VOLTAGE as V Nodal Variables = solution->TEMPERATURE as T End End End End 168 BIBLIOGRAPHY [1] Gerald W. Wellman. Mapvar: a computer program to transfer solution data between finite element meshes. SAND 1999-0466, Sandia National Laboratories, Albuquerque, NM, USA, March 1999. 1.1 [2] The SNTools Project. SNTools SourceForge Project. Online. 2.3 169 170 INDEX A Definition For Function, 50 Abort If Field Not Defined On Copy Transfer Send Or Density Scale Factor, 35, 37 Receive Object, 98, 99 Deposit Specific Internal Energy, 35, 38 Abort If Search Object Outside Of Tolerance, 98, 99 Differentiate Expression, 50, 53 Abscissa, 50, 51 Distance Function Is Closest Receive Node To Send Abscissa Offset, 50, 51 Centroid, 98, 100 Abscissa Scale, 50, 51 Active For Procedure, 35, 37 E Adapt Mesh, 70, 71, 73, 74, 76, 77, 79, 81, 82 Effective Moduli Model, 35, 38 Adaptiveloop, 69, 71, 74, 81 elemDeathSimpleAria Advance, 73, 74, 76, 77, 79, 81, 82, 84 Input, 144 Alias, 29, 31 Statement, 143 All Fields, 98, 99 Element Numerical Formulation, 35, 38 Assembly, 30 Energy Iteration Tolerance, 35, 39 At Discontinuity Evaluate To, 50, 52 estack Input, 149 B Statement, 149 barOneWayCoupleDifferentMesh Evaluate Expression, 50, 53 Statement, 124 Evaluate From, 50, 54 barOneWayCoupleFromDifferentMesh Event, 70, 71, 73, 75–77, 79–82, 84 Input, 125 Exclude Ghosted, 98, 100 barOneWayCoupleFromFile Expression Variable:, 50, 54, 55 Input, 120 Expressions, 50 Statement, 120 barOneWayCoupleTransfer F Input, 129 Faradays Constant, 46, 47 Statement, 129 Field Types, 50, 55 Bending Hourglass, 35, 37 Finite Element Model, 29 Block, 30 Fixed Time, 112, 114 C From, 98, 101 Column Titles, 50, 52 Component Separator Character, 29, 31 G Converged When, 85, 86 gapClosure Coordinate System, 29, 32 Input, 134 Copy, 98, 100 Statement, 134 Create, 29, 31 Gauss Point Integration Order, 98, 101 Create Element Field, 112, 114 Geometric Tolerance, 98, 101 Create Nodal Field, 112, 114 Global Constants, 46 Gravity Vector, 46, 47 D Data File, 50, 52 H Database Name, 29, 32 Heartbeat, 113 Database Type, 30, 32 Heartbeat Output, 113 Debug, 50, 53 History Output, 113 Decomposition Method, 30, 33 Hourglass, 35, 39 171 I Output, 70, 72, 73, 75, 76, 78–81, 83 Ideal Gas Constant, 46, 48 Inactive For Procedure, 35, 39 P Include All Blocks, 35, 40 Parameters For, 70, 85 Incremental Number Of Steps, 85, 86 Parameters For Aria Region, 85 Initial Deltat, 85, 86 Parameters For Block, 30, 35 Initialize, 69, 84 Parameters For Phase, 30 Input_Output Region, 112 Parameters For Surface, 30 Inspect With File, 98, 102 Parametric Tolerance, 98, 104 Interpolate, 98, 102 Patch Recovery Evaluation, 98, 104 Interpolation Function, 98, 103 Periodicity Time, 112, 116 Inversion Aversion Exponent, 35, 40 Phase, 36, 45 Inversion Aversion Stiffness, 35, 40 Planck Constant, 47, 49 Inversion Aversion Transition Jacobian, 35, 41 Postprocess Aria Region, 70, 72, 74–76, 78–81, 83 Involve, 73, 75–77, 79–82, 84, 85 Q K Quadratic Bulk Viscosity, 36, 45 K-E Turbulence Model Parameter, 46, 48 K-W Turbulence Model Parameter, 47, 48 R Receive Blocks, 99 L Reinitialize Transient, 85, 87 Les Turbulence Model Parameter, 47, 48 Remove Block, 36, 45 Light Speed, 47, 49 Restart Data, 113 Linear Bulk Viscosity, 35, 41 Results Output, 113 Local Coordinate System, 35, 41 S M Scale By, 50, 56 Material, 35, 41 Search Coordinate Field, 98, 105 Material =, 36, 42 Search Geometric Tolerance, 98, 105 Max Energy Iterations, 36, 42 Search Surface Gap Tolerance, 98, 105 Membrane Hourglass, 36, 42 Search Type, 98, 106 Minimum Effective Dilatational Moduli Ratio, 36, 42 Section, 36, 46 Minimum Effective Shear Moduli Ratio, 36, 43 Select One Receiver For Each Send Object, 98, 106 Model, 36, 43 Select One Unique Receiver For Each Send Object, 98, 106 N Send, 98, 107 Nodes Outside Region, 98, 103 Send Block, 98, 107 Nonlinear, 74, 76, 81 Send Blocks, 99 Nonlocal Regularization Kmeans Cell Size, 36, 43 Send Field, 98, 108 Nonlocal Regularization Kmeans Maximum Iterations, 36, Sequential, 71, 81 44 Simulation Max Global Iterations, 70, 72 Nonlocal Regularization Kmeans Tolerance, 36, 44 Simulation Start Time, 70, 72 Nonlocal Regularization On, 36, 44 Simulation Termination Time, 70, 73 Nonlocal Regularization Partitioning Scheme, 36, 45 Solid Mechanics Use Model, 36, 46 Number Of Adaptivity Steps, 85, 87 Solution Control Description, 69 Number Of Steps, 85, 87 Start Time, 85, 87, 112, 117 Stefan Boltzmann Constant, 47, 49 O Subcycle, 74, 76, 79 Offset Time, 112, 115 Suppress Output From Nonlinear Loop, 85, 87 Omit Assembly, 30, 33 System, 70 Omit Block, 30, 33 Omit Volume, 30, 34 T Ordinate, 50, 55 Termination Time, 85, 88 Ordinate Offset, 50, 56 Time Interpolation Method, 112, 118 Ordinate Scale, 50, 56 Time Scale Factor, 30, 34 172 Time Step Quantum, 85, 88 Time Step Style, 85, 88 Timestep Adjustment Interval, 113, 118 Toggle Search Warnings, 99, 109 Total Change In Time, 85, 89 Transfer, 70, 73, 74, 76, 78, 79, 81, 83–85, 97, 98 Transient, 71, 73 Transverse Shear Hourglass, 36, 46 Turbulence Model, 47, 49 tv Input, 166 Statement, 166 Type, 50, 56 U Use Centroid For Geometric Proximity, 99, 109 Use Finite Element Model, 113, 118 Use Generic Names, 30, 34 Use Initialize, 70, 73 Use Material, 30, 35 Use System, 69, 70 V Values, 50, 58 X X Offset, 50, 57 X Scale, 50, 57 Y Y Offset, 50, 57 Y Scale, 50, 57 173 DISTRIBUTION Email—Internal Name Org. Sandia Email Address Technical Library 01911 sanddocs@sandia.gov 174 175 Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.