An adaptive scheme for the optimization of damping positions by decoupling controllability spaces in vibrational systems
J. Przybilla, M. Ugrica Vukojević, N. Truhar, P. Benner · 2025
Read An adaptive scheme for the optimization of damping positions by decoupling controllability spaces in vibrational systems by J. Przybilla, M. Ugrica Vukojević, N. Truhar, P. Benner (2025) 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-2506-20372.txt.
Introduction
Jennifer Przybilla and co-workers sped up the search for the best damper placements in shaking forms in a 2025 paper. Engineers place outer dampers to shrink how strongly inputs shake outputs. Judged by an energy response whose math work means solving large Lyapunov math rules. Trying many placements multiplies that cost painfully, so the team built a reduction that learns the relevant answer space once and reuses it.
An self-tuning scheme grows a basis covering Lyapunov answers for all option placements, starting from the undamped system's decoupled subspace. The reduced tuning problem then runs in fair time while still ranking placements truly. Decoupling controllability spaces cleanly separates what dampers add from what the bare form already does.
Catalog record
| CREATOR | J. Przybilla, M. Ugrica Vukojević, N. Truhar, P. Benner |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 90 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-20372
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
An adaptive scheme for the optimization of damping positions by decoupling controllability spaces in vibrational systems J. Przybilla, M. Ugrica Vukojević, N. Truhar, P. Benner Abstract In this work, the problem of optimizing damper positions in vibrational systems is investigated. The objective is to determine the positions of external dampers in such a way that the influence of the input on the output is minimized. The energy response serves as an optimization criterion, whose computation involves solving Lyapunov equations. Hence, in order to find the best positions, many of these equations need to be solved, and so the minimization process can have a high computational cost. To accelerate the process of finding the optimal positions, we propose a new reduction method. Our algorithm generates a basis spanning an approximation to the solution space of the Lyapunov equations for all possible positions of the dampers. We derive an adaptive scheme that generates the reduced solution space by adding the subspaces of interest, and then we define the corresponding reduced optimization problem that is solvable in a reasonable amount of time. We decouple the solution spaces of the problem to obtain a space that corresponds to the system without external dampers and serves as a starting point for the reduction of the optimization problem. In addition, we derive spaces corresponding to the different damper positions that are used to expand the reduced basis if needed. To evaluate the quality of the basis, we introduce an error indicator based on the space decomposition. Our new technique produces a reduced optimization problem of significantly smaller dimension that is faster to solve than the original problem, which we illustrate with some numerical examples. An adaptive scheme for the optimization of damping positions by decoupling controllability spaces in vibrational systems \novelty We propose a new method that reduces the optimization problem of finding the suboptimal damping positions in mechanical systems. For that, we invent a procedure that decouples the controllability spaces into parameter-independent and parameter-dependent subspaces. We derive an adaptive reduced basis method that uses these subspaces to define a reduced optimization problem and hence accelerates the optimization process. Also a new error indicator is invented that is used to evaluate the quality of the approximations. The performance of our new method is illustrated in several examples. 1 Introduction Structures such as buildings, bridges or damps are often disturbed by external forces, e.g. wind occurrences, earthquakes or pedestrian movements. These disturbances can cause vibrations, deflection or even damages in the constructions which can be prevented by adding external dampers. This paper is about determining the best positions for dampers in order to minimize the effects of external forces on the structure. We consider vibrational systems of the form \displaystyle\begin{split}\mathbf{M}\ddot{\mathbf{x}}(t)+\mathbf{D}(c,g)\dot{\mathbf{x}}(t)+\mathbf{K}\mathbf{x}(t)&=\mathbf{B}\mathbf{u}(t),\\ \mathbf{y}(t)&=\mathbf{C}\mathbf{x}(t),\end{split} (1) where \mathbf{M},\penalty\ \mathbf{D}(c,g),\penalty\ \mathbf{K}\in\mathbb{R}^{n\times n} are the mass matrix, the damping matrix, and the stiffness matrix, respectively, and where the damping matrix depends on parameters c\in\boldsymbol{\mathcal{D}}_{c} and g\in\boldsymbol{\mathcal{D}}_{g} . Moreover, the matrix \mathbf{B}\in\mathbb{R}^{n\times m} is the input matrix and \mathbf{C}\in\mathbb{R}^{p\times n} is the output matrix. The vectors \mathbf{u}(t)\in\mathbb{R}^{m} , \mathbf{x}(t)\in\mathbb{R}^{n} and \mathbf{y}(t)\in\mathbb{R}^{p} are respectively representing the input, the state and the output of the system. Naturally, the matrices \mathbf{M} , and \mathbf{K} are symmetric and positive definite, while \mathbf{D}(c,g) is symmetric and generally positive semi definite matrix. In our context, the damping matrix \mathbf{D}(c,g) is symmetric, positive definite, consisting of two parts: a parameter-independent internal damping \mathbf{D}_{\mathrm{int}} which is positive definite and a parameter-dependent external damping \mathbf{D}_{\mathrm{ext}}(c,g) , i.e., \mathbf{D}(c,g)=\mathbf{D}_{\mathrm{int}}+\mathbf{D}_{\mathrm{ext}}(c,g). The external damping \mathbf{D}_{\mathrm{ext}}(c,g) depends on two types of parameters. The first ones are the damping gains g=\begin{bmatrix}g_{1},\dots,g_{\ell}\end{bmatrix}\in\boldsymbol{\mathcal{D}}_{g}\subset\mathbb{R}^{\ell}_{+} , which represent the viscosities of the dampers. We assume that the viscosities g_{j} are fixed over time and lie in given intervals [g_{j}^{-},\,g_{j}^{+}] , for all j=1,\dots,\ell . We collect these bounds by setting g\in\boldsymbol{\mathcal{D}}_{g} , where the parameter set \boldsymbol{\mathcal{D}}_{g} contains all given conditions. The second set of parameters consists of the damping positions c\in\boldsymbol{\mathcal{D}}_{c}\subset\{1,\dots,n\}^{\ell} which are stored in the matrix \mathbf{F}(c)\in\mathbb{R}^{n\times\ell} . The structure of the matrix \mathbf{F}(c) depends on the damper type so that e.g. grounded dampers are described by unit vectors that are concatenated to build the matrix \mathbf{F}(c) . This is described in more detail for the different numerical examples. The resulting external damper is then given as \displaystyle\mathbf{D}_{\mathrm{ext}}(c,g):=\mathbf{F}(c)\mathbf{G}(g)\mathbf{F}(c)^{\mathrm{T}},\qquad\mathbf{G}(g):=\mathop{\mathrm{diag}}\left(g_{1},\dots,g_{\ell}\right). In addition to the external dampers, the system is internally damped. There are several different models for internal damping. In this work we use a small multiple of the critical damping that is \displaystyle\mathbf{D}_{\mathrm{int}}:=2\alpha\mathbf{M}^{\frac{1}{2}}\left(\mathbf{M}^{-\frac{1}{2}}\mathbf{K}\mathbf{M}^{-\frac{1}{2}}\right)^{\frac{1}{2}}\mathbf{M}^{\frac{1}{2}}, (2) for \alpha\ll 1 , which is a widely used convention and has been applied, for example, in [ 4 , 5 ] . However, our theory can be applied to all modal dampings, which includes, for example, Rayleigh damping defined in [ 40 , 19 ] . We assume that the number of external dampers \ell is significantly smaller than the dimension n . Because of this structure, these systems are also asymptotically stable, i.e. all eigenvalues \lambda(c,g) of the polynomial eigenvalue problem (\lambda(c,g)^{2}\mathbf{M}+\lambda(c,g)\mathbf{D}(c,g)+\mathbf{K})x(c,g)=0 have a negative real part. In order to simplify the computations in this work, we consider a system transformation, that was introduced, e.g. in [ 37 ] , where it is shown that there exists a transformation \boldsymbol{\Phi} such that \displaystyle\boldsymbol{\Phi}^{\mathrm{T}}\mathbf{M}\boldsymbol{\Phi}=\mathbf{I}_{n},\qquad\boldsymbol{\Phi}^{\mathrm{T}}\mathbf{K}\boldsymbol{\Phi}=\boldsymbol{\Omega}^{2}=\mathop{\mathrm{diag}}\left(\omega_{1}^{2},\dots,\omega_{n}^{2}\right),\qquad\boldsymbol{\Phi}^{\mathrm{T}}\mathbf{D}_{\mathrm{int}}\boldsymbol{\Phi}=2\alpha\boldsymbol{\Omega}, i.e. the matrix \boldsymbol{\Phi} simultaneously diagonalizes the mass matrix, the stiffness matrix and the internal damping where the transformed mass matrix is the identity matrix. The values \omega_{1},\dots,\omega_{n} are the eigenvalues of the undamped system and are called eigenfrequencies , see [ 38 ] . The transformed external damping matrix consists of low-rank factors \boldsymbol{\Phi}^{\mathrm{T}}\mathbf{D}_{\mathrm{ext}}(c,g)\boldsymbol{\Phi}=\widetilde{\mathbf{F}}(c)\mathbf{G}(g)\widetilde{\mathbf{F}}(c)^{\mathrm{T}},\qquad\widetilde{\mathbf{F}}(c):=\boldsymbol{\Phi}^{\mathrm{T}}\mathbf{F}(c). We summarize the transformed damping matrix by defining \widetilde{\mathbf{D}}(c,g):=2\alpha\boldsymbol{\Omega}+\widetilde{\mathbf{F}}(c)\mathbf{G}(g)\widetilde{\mathbf{F}}(c)^{\mathrm{T}} . Moreover, we define \widetilde{\mathbf{B}}:=\boldsymbol{\Phi}^{\mathrm{T}}\mathbf{B} and \widetilde{\mathbf{C}}:=\mathbf{C}\boldsymbol{\Phi} to obtain the transformed system \displaystyle\begin{split}\ddot{\widetilde{\mathbf{x}}}(t)+\widetilde{\mathbf{D}}(c,g)\dot{\widetilde{\mathbf{x}}}(t)+\boldsymbol{\Omega}^{2}\widetilde{\mathbf{x}}(t)&=\widetilde{\mathbf{B}}\mathbf{u}(t),\\ \mathbf{y}(t)&=\widetilde{\mathbf{C}}\widetilde{\mathbf{x}}(t),\end{split} (3) which is equivalent to the original system ( 1 ) in the sense that both systems have the same input-output behavior. Vibrational systems and the optimization of their contained dampers have been studied in a large number of papers in the last decades, for example in [ 2 , 9 , 12 , 15 , 18 , 39 , 23 , 21 , 16 ] , where external dampers are considered in systems that already contain internal damping of small magnitude. The aim is to choose external dampers that stabilize the system and shift eigenfrequencies in such a way that possible external loads do not lead to resonances. There are two variables that can be optimized: the damper positions and the damping gains. Moreover, depending on the application, different criteria are chosen to quantify the stability of systems and the response to external disturbances. If homogeneous systems are considered, i.e. systems ( 1 ) with \mathbf{B}\equiv 0 , then the spectral abscissa or the total average energy are used. Examples are described in [ 24 , 11 , 5 , 35 , 38 ] . In the non-homogeneous case, i.e. \mathbf{B}\neq 0 , as considered in this work, external disturbances are taken into account, which potentially play an important role in real-life scenarios. In these cases, the average displacement amplitude can be evaluated, which minimizes the square of the norm of the displacement \widetilde{\mathbf{x}}(t) averaged over a certain time period, see [ 20 , 36 ] . Another criterion, which is used in