Topology optimization for particle flow problems using Eulerian-Eulerian model with a finite difference method Journal: Elsevier
Chih-Hsiang Chen, Kentaro Yaji · 2024
Read Topology optimization for particle flow problems using Eulerian-Eulerian model with a finite difference method Journal: Elsevier by Chih-Hsiang Chen, Kentaro Yaji (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-19619.txt.
Introduction
Chih-Hsiang Chen and Kentaro Yaji wrote a paper from 2024 on shaping flow paths so that particles in a stream meet more drag. Plants that handle powders and grains need such control, but the dance between particles and fluid is hard to design by hand. Their method lets a computer grow the shape of the flow field step by step toward a set goal.
They model the mix as two fluids sharing one space. A fast grid method runs the model, and an auto-diff tool works out how each shape change moves the goal. The goal pairs a wide spread of drag with low waste of the fluid's power. First tests matched a known benchmark. Winding, snake-like channels then proved best at spreading the drag.
Catalog record
| CREATOR | Chih-Hsiang Chen, Kentaro Yaji |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 64 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-19619
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Topology optimization for particle flow problems using Eulerian-Eulerian model with a finite difference method Journal: Elsevier Chih-Hsiang Chen, Kentaro Yaji Abstract Particle flow processing is widely employed across various industrial applications and technologies. Due to the complex interactions between particles and fluids, designing effective devices for particle flow processing is challenging. In this study, we propose a topology optimization method to design flow fields that effectively enhance the resistance encountered by particles. Particle flow is simulated using an Eulerian-Eulerian model based on a finite difference method. Automatic differentiation is implemented to compute sensitivities using a checkpointing algorithm. We formulate the optimization problem as maximizing the variation of drag force on particles while reducing fluid power dissipation. Initially, we validate the finite difference flow solver through numerical examples of particle flow problems and confirm that the corresponding topology optimization produces a result comparable to the benchmark problem. Furthermore, we investigate the effects of Reynolds and Stokes numbers on the optimized flow field. The numerical results indicate that serpentine flow fields can effectively enhance the variation in particle drag force. 1 Introduction The process of handling particle flow plays a pivotal role in a wide range of industrial applications [ 1 , 2 ] . Among the factors governing these processes, the particle residence time—the duration particles spend in a specific region—serves as a fundamental factor of process efficiency, affecting the degree of reaction [ 3 , 4 ] , mixing quality [ 5 , 6 ] , and heat transfer effect [ 7 , 8 ] in industrial devices. A fundamental approach to adjusting particle residence time is to modify the resistance acting on the particles. This can be achieved by controlling operational conditions [ 9 , 10 ] , altering particle properties [ 11 , 12 , 13 ] , or designing flow paths to impose different levels of resistance [ 14 , 15 , 16 ] . Furthermore, it is beneficial to increase the resistance encountered by particles for applications that require extending particle residence time. For instance, in fluidized beds, baffles are used to control particle movement by reducing back-mixing and redistributing particles, which increases resistance and enhances mixing efficiency [ 17 , 18 , 19 ] . In microreactors, complex channel designs are employed to extend flow paths and ensure complete chemical reactions [ 20 , 21 , 22 ] . Additionally, porous media are incorporated within particle heating receivers to increase the resistance to particle flow, allowing the particles to absorb solar energy effectively and thereby improving thermal efficiency [ 23 , 24 ] . However, the complex interactions between particles and fluid make it challenging to design an effective flow path to increase particle flow resistance. Although size optimization is commonly employed for such requirements [ 25 ] , predefined shapes offer limited flexibility in determining optimized configurations at the conceptual design stage. Unlike size and shape optimization, topology optimization can generate optimized configurations without the designer’s intuition at the conceptual design stage. The homogenization design method for topology optimization, first proposed by Bendsøe and Kikuchi, processes heterogeneous materials into homogeneous forms to achieve optimized material distribution for enhanced structural performance [ 26 ] . Moreover, the density approach, also known as the SIMP (Solid Isotropic Material with Penalization) method, employs a continuous function to enable a smooth transition of design variables throughout the entire design domain in the optimization process [ 27 ] . The use of a penalization factor in the SIMP method eliminates intermediate values of design variables, improving material distribution from a manufacturing viewpoint. Based on the SIMP method, topology optimization for Stokes flow was pioneered by Borrvall and Petersson [ 28 ] . They introduced a fictitious force term to model porous domains, allowing for no-slip boundary conditions by setting a sufficiently high value of fictitious force. Subsequent research has broadened the scope of topology optimization in fluid dynamics. For instance, Guest et al. developed a methodology for optimizing creeping flow conditions [ 29 ] . Gersborg-Hansen et al. proposed a topology optimization method for laminar channel flow problems utilizing Navier-Stokes equations [ 30 ] . Topology optimization methods for unsteady flow problems have been extended by Kreissl et al. [ 31 ] and Deng et al. [ 32 ] . Owing to the high degree of design freedom, research in fluid topology optimization is not only limited to pure fluid issues but also includes a wide range of studies that consider interactions with various physical fields [ 33 , 34 , 35 , 36 , 37 ] . In the case of multi-physics problems, there has been growing interest in the topology optimization of particle flows in recent years. Andreasen proposed a topology optimization approach for optimizing inertial microfluidic devices for particle manipulation [ 38 ] . Yoon proposed studies focused on controlling particle trajectories in fluid [ 39 , 40 , 41 ] . These studies employ the Eulerian-Lagrangian model to explore the topology optimization of single or multiple particles in fluids. It is crucial to highlight that particle modeling can be conducted using either the Eulerian-Lagrangian approach or the Eulerian-Eulerian approach, each offering distinct advantages for particle flow analysis. In the Eulerian-Lagrangian approach, the fluid is treated as a continuum, and each particle is tracked individually within the Lagrangian framework, enabling the detailed simulation of particle trajectories and interactions with the fluid phase [ 42 ] . Conversely, the Eulerian-Eulerian approach treats both the gas and particles as continuous phases with their conservation equations, allowing for the investigation of behaviors of numerous particles, including sedimentation, fluidization, and particle residence time distribution [ 43 , 44 , 45 ] . Although prior studies have utilized the Eulerian-Lagrangian approach to optimize particle manipulations and trajectory problems, employing the Eulerian-Lagrangian approach to simulate the collective behavior of particles can lead to high consumption of computational resources due to the necessity of tracking each particle individually [ 46 ] . Therefore, it becomes important to optimize problems that address issues related to the behavior of numerous particles using the Eulerian-Eulerian approach. On the other hand, according to the literature review [ 47 ] , most studies on topology optimization employ the finite element method, while a few studies utilize the finite volume method [ 48 ] , lattice Boltzmann method [ 49 ] and particle-based method [ 50 ] . Although the finite element method is well-suited for structural problems with arbitrary shapes of objects, it requires relatively significant memory resources [ 51 ] and involves specific treatments for conservation law in fluid problems [ 52 , 53 ] . Despite its challenges with complex geometries, the finite difference method effectively handles fluid simulation [ 54 , 55 ] and supports high-order schemes for complex fluid behavior [ 56 , 57 ] . From the view of standard topology optimization methods, an optimized structure is typically determined by the distribution of design variables. There is no need to employ unstructured meshes to address issues of arbitrary structural shapes, which makes the finite difference method an appropriate choice for fluid topology optimization. In this paper, we present a topology optimization method for particle flow to increase the resistance acting on the particles