-
Multi-level Optimal Control with Neural Surrogate Models
Authors:
Dante Kalise,
EstefanÃa Loayza-Romero,
Kirsten A. Morris,
Zhengang Zhong
Abstract:
Optimal actuator and control design is studied as a multi-level optimisation problem, where the actuator design is evaluated based on the performance of the associated optimal closed loop. The evaluation of the optimal closed loop for a given actuator realisation is a computationally demanding task, for which the use of a neural network surrogate is proposed. The use of neural network surrogates t…
▽ More
Optimal actuator and control design is studied as a multi-level optimisation problem, where the actuator design is evaluated based on the performance of the associated optimal closed loop. The evaluation of the optimal closed loop for a given actuator realisation is a computationally demanding task, for which the use of a neural network surrogate is proposed. The use of neural network surrogates to replace the lower level of the optimisation hierarchy enables the use of fast gradient-based and gradient-free consensus-based optimisation methods to determine the optimal actuator design. The effectiveness of the proposed surrogate models and optimisation methods is assessed in a test related to optimal actuator location for heat control.
△ Less
Submitted 12 February, 2024;
originally announced February 2024.
-
Extended Kalman filter based observer design for semilinear infinite-dimensional systems
Authors:
Sepideh Afshar,
Fabian Germ,
Kirsten A. Morris
Abstract:
In many physical applications, the system's state varies with spatial variables as well as time. The state of such systems is modelled by partial differential equations and evolves on an infinite-dimensional space. Systems modelled by delay-differential equations are also infinite-dimensional systems. The full state of these systems cannot be measured. Observer design is an important tool for esti…
▽ More
In many physical applications, the system's state varies with spatial variables as well as time. The state of such systems is modelled by partial differential equations and evolves on an infinite-dimensional space. Systems modelled by delay-differential equations are also infinite-dimensional systems. The full state of these systems cannot be measured. Observer design is an important tool for estimating the state from available measurements. For linear systems, both finite- and infinite-dimensional, the Kalman filter provides an estimate with minimum-variance on the error, if certain assumptions on the noise are satisfied. The extended Kalman filter (EKF) is one type of extension to nonlinear finite-dimensional systems. In this paper we provide an extension of the EKF to semilinear infinite-dimensional systems. Under mild assumptions we prove the well-posedness of equations defining the EKF. Local exponential stability of the error dynamics is shown. Only detectability is assumed, not observability, so this result is new even for finite-dimensional systems. The results are illustrated with implementation of finite-dimensional approximations of the infinite-dimensional EKF on an example.
△ Less
Submitted 15 February, 2022;
originally announced February 2022.
-
Closed-form H-infinity Optimal Control for a Class of Infinite-Dimensional Systems
Authors:
Carolina Bergeling,
Kirsten A. Morris,
Anders Rantzer
Abstract:
H-infinity optimal control and estimation are addressed for a class of systems governed by partial differential equations with bounded input and output operators. Diffusion equations are an important example in this class. Explicit formulas for the optimal state feedback controller as well as the optimal state estimator are given. Unlike traditional methods for H-infinity synthesis, no iteration i…
▽ More
H-infinity optimal control and estimation are addressed for a class of systems governed by partial differential equations with bounded input and output operators. Diffusion equations are an important example in this class. Explicit formulas for the optimal state feedback controller as well as the optimal state estimator are given. Unlike traditional methods for H-infinity synthesis, no iteration is needed to obtain the optimal solution. Moreover, the optimal performance for both the state feedback and state estimation problems are explicitly calculated. This is shown to be useful for problems of H-infinity optimal actuator and sensor location. Furthermore, the results can be used in testing and bench-marking of general purpose algorithms for H-infinity synthesis. The results also apply to finite-dimensional systems.
△ Less
Submitted 8 June, 2021;
originally announced June 2021.
-
Shape Optimization of Actuators over Banach Spaces for Nonlinear Systems
Authors:
M. Sajjad Edalatzadeh,
Dante Kalise,
Kirsten A. Morris,
Kevin Sturm
Abstract:
In this paper, optimal actuator shape for nonlinear parabolic systems is discussed. The system under study is an abstract differential equation with a locally Lipschitz nonlinear part. A quadratic cost on the state and input of the system is considered. The existence of an optimal actuator shape has been established in the literature. This paper focuses on driving the optimality conditions for act…
▽ More
In this paper, optimal actuator shape for nonlinear parabolic systems is discussed. The system under study is an abstract differential equation with a locally Lipschitz nonlinear part. A quadratic cost on the state and input of the system is considered. The existence of an optimal actuator shape has been established in the literature. This paper focuses on driving the optimality conditions for actuator shapes belonging to a Banach space. The application of the theory to the optimal actuator shape design for railway track model is considered.
△ Less
Submitted 17 February, 2020;
originally announced February 2020.
-
Hysteresis and Stabillity
Authors:
Amenda Chow,
Kristen A. Morris,
Gina Faraj Rabbah
Abstract:
Hysteresis can be defined from a dynamical systems perspective with respect to equilibrium points. Consequently, hysteresis naturally lends itself as a topic to illustrate and extend concepts in a dynamical systems course. A number of examples exhibiting hysteresis, most motivated by applications, are presented. Although the examples can be used to construct student exercises, specific questions a…
▽ More
Hysteresis can be defined from a dynamical systems perspective with respect to equilibrium points. Consequently, hysteresis naturally lends itself as a topic to illustrate and extend concepts in a dynamical systems course. A number of examples exhibiting hysteresis, most motivated by applications, are presented. Although the examples can be used to construct student exercises, specific questions are listed in an appendix. A brief extension on hysteresis in partial differential equations is also included.
△ Less
Submitted 23 May, 2022; v1 submitted 13 January, 2020;
originally announced January 2020.
-
Optimal Controller and Actuator Design for Nonlinear Parabolic Systems
Authors:
M. Sajjad Edalatzadeh,
Kirsten A. Morris
Abstract:
Many physical systems are modeled by nonlinear parabolic differential equations, such as the Kuramoto-Sivashinsky (KS) equation. In this paper, the existence of a concurrent optimal controller and actuator design is established for semilinear parabolic systems. Optimality equations are provided. The results are shown to apply to optimal controller/actuator design for the Kuramoto-Sivashinsky equat…
▽ More
Many physical systems are modeled by nonlinear parabolic differential equations, such as the Kuramoto-Sivashinsky (KS) equation. In this paper, the existence of a concurrent optimal controller and actuator design is established for semilinear parabolic systems. Optimality equations are provided. The results are shown to apply to optimal controller/actuator design for the Kuramoto-Sivashinsky equation and also nonlinear diffusion.
△ Less
Submitted 7 October, 2019;
originally announced October 2019.
-
Optimal actuator design for vibration control based on LQR performance and shape calculus
Authors:
M. Sajjad Edalatzadeh,
Dante Kalise,
Kirsten A. Morris,
Kevin Sturm
Abstract:
Optimal actuator design for a vibration control problem is calculated. The actuator shape is optimized according to the closed-loop performance of the resulting linear-quadratic regulator and a penalty on the actuator size. The optimal actuator shape is found by means of shape calculus and a topological derivative of the linear-quadratic regulator (LQR) performance index. An abstract framework is…
▽ More
Optimal actuator design for a vibration control problem is calculated. The actuator shape is optimized according to the closed-loop performance of the resulting linear-quadratic regulator and a penalty on the actuator size. The optimal actuator shape is found by means of shape calculus and a topological derivative of the linear-quadratic regulator (LQR) performance index. An abstract framework is proposed based on the theory for infinite-dimensional optimization of both the actuator shape and the associated control problem. A numerical realization of the optimality condition is presented for the actuator shape using a level-set method for topological derivatives. A Numerical example illustrating the design of actuator for Euler-Bernoulli beam model is provided.
△ Less
Submitted 18 March, 2019;
originally announced March 2019.
-
Stability and Well-posedness of a Nonlinear Railway Track Model
Authors:
M. Sajjad Edalatzadeh,
Kirsten A. Morris
Abstract:
Railway tracks rest on a foundation known for exhibiting nonlinear viscoelastic behavior. Railway track deflections are modeled by a semilinear partial differential equation. This paper studies the stability of solutions to this equation in presence of an input. With the aid of a suitable Lyapunov function, existence and exponential stability of classical solutions is established for certain input…
▽ More
Railway tracks rest on a foundation known for exhibiting nonlinear viscoelastic behavior. Railway track deflections are modeled by a semilinear partial differential equation. This paper studies the stability of solutions to this equation in presence of an input. With the aid of a suitable Lyapunov function, existence and exponential stability of classical solutions is established for certain inputs. The Lyapunov function is further used to find an a-priori estimate of the solutions, and also to study the input-to-state stability (ISS) of mild solutions.
△ Less
Submitted 3 July, 2018; v1 submitted 7 March, 2018;
originally announced March 2018.
-
Optimal Actuator Design for Semi-linear Systems
Authors:
M. Sajjad Edalatzadeh,
Kirsten A. Morris
Abstract:
Actuator location and design are important choices in controller design for distributed parameter systems. Semi-linear partial differential equations model a wide spectrum of physical systems with distributed parameters. It is shown that under certain conditions on the nonlinearity and the cost function, an optimal control input together with an optimal actuator choice exists. First-order necessar…
▽ More
Actuator location and design are important choices in controller design for distributed parameter systems. Semi-linear partial differential equations model a wide spectrum of physical systems with distributed parameters. It is shown that under certain conditions on the nonlinearity and the cost function, an optimal control input together with an optimal actuator choice exists. First-order necessary optimality conditions are derived. The results are applied to optimal actuator and controller design in a nonlinear railway track model as well as semi-linear wave models.
△ Less
Submitted 22 October, 2018; v1 submitted 15 February, 2018;
originally announced February 2018.
-
Zero Dynamics for Port-Hamiltonian Systems
Authors:
Birgit Jacob,
Kirsten A. Morris,
Hans Zwart
Abstract:
The zero dynamics of infinite-dimensional systems can be difficult to characterize. The zero dynamics of boundary control systems are particularly problematic. In this paper the zero dynamics of port-Hamiltonian systems are studied. A complete characterization of the zero dynamics for a port-Hamiltonian systems with invertible feedthrough as another port-Hamiltonian system on the same state space…
▽ More
The zero dynamics of infinite-dimensional systems can be difficult to characterize. The zero dynamics of boundary control systems are particularly problematic. In this paper the zero dynamics of port-Hamiltonian systems are studied. A complete characterization of the zero dynamics for a port-Hamiltonian systems with invertible feedthrough as another port-Hamiltonian system on the same state space is given. It is shown that the zero dynamics for any port-Hamiltonian system with commensurate wave speeds are well-defined, and are also a port-Hamiltonian system. Examples include wave equations with uniform wave speed on a network. A constructive procedure for calculation of the zero dynamics, that can be used for very large system order, is provided.
△ Less
Submitted 19 November, 2017;
originally announced November 2017.
-
Landau-Lifshitz Equation with Affine Control
Authors:
Amenda Chow,
Kirsten A. Morris
Abstract:
The Landau-Lifshitz equation is a coupled set of nonlinear partial differential equations that describes the dynamics of magnetization in a ferromagnet. This equation has an infinite number of stable equilibria. Steering the system from one equilibrium to another is a problem of both theoretical and practical interest. Since the objective is to steer between equilibria, approaches based on lineari…
▽ More
The Landau-Lifshitz equation is a coupled set of nonlinear partial differential equations that describes the dynamics of magnetization in a ferromagnet. This equation has an infinite number of stable equilibria. Steering the system from one equilibrium to another is a problem of both theoretical and practical interest. Since the objective is to steer between equilibria, approaches based on linearization are not appropriate. It is proven that affine proportional control can be used to steer the system from an arbitrary initial state, including an equilibrium point, to a specified equilibrium point. The second point becomes a globally asymptotically stable equilibrium of the controlled system. The control also removes hysteresis from the Landau-Lifshitz equation. These results are illustrated with simulations.
△ Less
Submitted 28 May, 2016;
originally announced May 2016.
-
Modeling and related results for current-actuated piezoelectric beams by including magnetic effects
Authors:
Kirsten A. Morris,
Ahmet Ozkan Ozer
Abstract:
Piezo-electric material can be controlled with current as the electrical variable, instead of voltage. The main purpose of this paper is to derive the governing equations for a current-controlled piezo-electric beam and to investigate stabilizability. Besides the consideration of current control, there are several new aspects to the model here. Most significantly, magnetic effects are included. Fo…
▽ More
Piezo-electric material can be controlled with current as the electrical variable, instead of voltage. The main purpose of this paper is to derive the governing equations for a current-controlled piezo-electric beam and to investigate stabilizability. Besides the consideration of current control, there are several new aspects to the model here. Most significantly, magnetic effects are included. For the electromagnetic part of the model, electrical potential and magnetic vector potential are chosen to be quadratic-through thickness to include the induced effects of the electromagnetic field. Two sets of decoupled system of partial differential equations are obtained; one for stretching motion and another one for bending motion. Hamilton's principle is used to derive a boundary value problem that models a single piezo-electric beam actuated by a charge (or current) source at the electrodes. Current or charge controllers at the electrodes can only control the stretching motion. Attention is therefore focused on control of the stretching equations in this paper. It is shown that the Lagrangian of the beam is invariant under certain transformations. A Coulomb-type gauge condition which is widely used in the electromagnetic theory is used here. This gauge condition decouples the electrical potential equation from the equations of the magnetic potential. A semigroup approach is used to prove that the Cauchy problem is well-posed. Unlike the voltage or charge actuation, a bounded control operator in the natural energy space is obtained in the current actuation case. The paper concludes with analysis of stabilizability and comparison with other actuation approaches and models.
△ Less
Submitted 7 March, 2016;
originally announced March 2016.
-
Control of the Landau-Lifshitz Equation
Authors:
Amenda Chow,
Kirsten A. Morris
Abstract:
The Landau--Lifshitz equation describes the dynamics of magnetization inside a ferromagnet. This equation is nonlinear and has an infinite number of stable equilibria. It is desirable to control the system from one equilibrium to another. A control that moves the system from an arbitrary initial state, including an equilibrium point, to a specified equilibrium is presented. It is proven that the s…
▽ More
The Landau--Lifshitz equation describes the dynamics of magnetization inside a ferromagnet. This equation is nonlinear and has an infinite number of stable equilibria. It is desirable to control the system from one equilibrium to another. A control that moves the system from an arbitrary initial state, including an equilibrium point, to a specified equilibrium is presented. It is proven that the second point is an asymptotically stable equilibrium of the controlled system. The results are illustrated with some simulations.
△ Less
Submitted 26 September, 2015; v1 submitted 18 September, 2015;
originally announced September 2015.
-
Hysteresis in the Linearized Landau-Lifshitz Equation
Authors:
Amenda Chow,
Kirsten A. Morris
Abstract:
The Landau-Lifshitz equation describes the behaviour of magnetization inside a ferromagnetic object. It is known that the Landau-Lifshitz equation has an infinite number of stable equilibrium points. The existence of multiple stable equilibria is closely related to hysteresis. This is a phenomenon that is often characterized by a looping behaviour; however, the existence of a loop is not sufficien…
▽ More
The Landau-Lifshitz equation describes the behaviour of magnetization inside a ferromagnetic object. It is known that the Landau-Lifshitz equation has an infinite number of stable equilibrium points. The existence of multiple stable equilibria is closely related to hysteresis. This is a phenomenon that is often characterized by a looping behaviour; however, the existence of a loop is not sufficient to identify hysteretic systems, but is defined more precisely as the presence of looping as the frequency of the input goes to zero. We describe these two approaches to identification of hysteresis and demonstrate that both the linear and nonlinear Landau-Lifshitz equations exhibit hysteresis. The presence of hysteresis in the linear Landau-Lifshitz equation, as well as in a simpler system also described here, indicates that nonlinearity is not necessary for hysteresis to exist.
△ Less
Submitted 18 September, 2015;
originally announced September 2015.
-
Modeling an elastic beam with piezoelectric patches by including magnetic effects
Authors:
A. O. Ozer,
K. A. Morris
Abstract:
Models for piezoelectric beams using Euler-Bernoulli small displacement theory predict the dynamics of slender beams at the low frequency accurately but are insufficient for beams vibrating at high frequencies or beams with low length-to-width aspect ratios. A more thorough model that includes the effects of rotational inertia and shear strain, Mindlin-Timoshenko small displacement theory, is need…
▽ More
Models for piezoelectric beams using Euler-Bernoulli small displacement theory predict the dynamics of slender beams at the low frequency accurately but are insufficient for beams vibrating at high frequencies or beams with low length-to-width aspect ratios. A more thorough model that includes the effects of rotational inertia and shear strain, Mindlin-Timoshenko small displacement theory, is needed to predict the dynamics more accurately for these cases. Moreover, existing models ignore the magnetic effects since the magnetic effects are relatively small. However, it was shown recently \cite{O-M1} that these effects can substantially change the controllability and stabilizability properties of even a single piezoelectric beam. In this paper, we use a variational approach to derive models that include magnetic effects for an elastic beam with two piezoelectric patches actuated by different voltage sources. Both Euler-Bernoulli and Mindlin-Timoshenko small displacement theories are considered. Due to the magnetic effects, the equations are quite different from the standard equations.
△ Less
Submitted 24 March, 2014;
originally announced March 2014.
-
Modeling and stabilizability of voltage-actuated piezoelectric beams with magnetic effects
Authors:
K. A. Morris,
A. O. Ozer
Abstract:
Models for piezoelectric beams and structures with piezoelectric patches generally ignore magnetic effects. This is because the magnetic energy has a relatively small effect on the overall dynamics. Piezoelectric beam models are known to be exactly observable, and can be exponentially stabilized in the energy space by using a mechanical feedback controller. In this paper, a variational approach is…
▽ More
Models for piezoelectric beams and structures with piezoelectric patches generally ignore magnetic effects. This is because the magnetic energy has a relatively small effect on the overall dynamics. Piezoelectric beam models are known to be exactly observable, and can be exponentially stabilized in the energy space by using a mechanical feedback controller. In this paper, a variational approach is used to derive a model for a piezoelectric beam that includes magnetic effects. It is proven that the partial differential equation model is well-posed. Magnetic effects have a strong effect on the stabilizability of the control system. For almost all system parameters the piezoelectric beam can be strongly stabilized, but is not exponentially stabilizable in the energy space. Strong stabilization is achieved using only electrical feedback. Furthermore, using the same electrical feedback, an exponentially stable closed-loop system can be obtained for a set of system parameters of zero Lebesgue measure. These results are compared to those of a beam without magnetic effects.
△ Less
Submitted 1 December, 2013;
originally announced December 2013.
