Computation of the Lyapunov matrix for periodic time-delay systems and its application to robust stability analysis
Introduction
Time-delay systems are known to be suitable for modeling a variety of physical phenomena (see, for instance, the books [1], [2], [3], [4]). A well known tool for analyzing their stability properties is the Lyapunov–Krasovskii approach. In the linear time-invariant case, the first results concerning the Lyapunov–Krasovskii functionals with negative prescribed derivatives were presented in [5]. Notable advances in this direction were then developed in [6] and [7]. In [8], this approach received a decisive impulse with the introduction of the so-called functionals of complete type. These functionals have two notable features: (i) they admit a quadratic lower bound if the system is stable and (ii) they are determined by the delay Lyapunov matrix (see [9] for a comprehensive study of this class of functionals).
Periodic delay systems have the complexity of both delay systems and time-varying systems. The stability properties of this class of equations have been analyzed by different methods, based mainly on the Floquet theory (see, for instance, [3] and [10]). A time-domain approach for studying periodic time-delay systems is presented in [11] by extending the Lyapunov–Krasovskii functionals of complete type framework developed for the invariant case [9]. This new perspective has made it possible to obtain robust stability conditions with respect to the parameters, exponential estimates [11], and more recently, to prove necessary stability conditions depending on the delay Lyapunov matrix [12].
These useful theoretical results depend on the knowledge of the delay Lyapunov matrix, which cannot be computed by the semianalytical method employed for time-invariant linear systems (see Section 2.10 in [9]). The computation of the delay Lyapunov matrix for periodic systems is a difficult task. In [11] and [13], it is shown that for the case of delay equal to the principal period, the Lyapunov matrix is a solution of a delay free partial differential equations (PDE) system that satisfies a set of boundary conditions. An iterative procedure based on the minimization of a function aiming at the computation of the initial conditions of a modified Runge–Kutta method is proposed. However, this methodology only works for some particular cases as it demands both the objective function to have a minimum and the minimization process to be convergent.
In this paper, we present an effective method for computing the delay Lyapunov matrix, which is based on the introduction of a new set of boundary conditions that are satisfied by the PDE system. This substantial modification enables us to propose a new procedure for computing the initial conditions. The main advantage of this new proposal, compared to the one presented in [11], is that it is not iterative and therefore it avoids the previously mentioned difficulties.
Having an algorithm for computing this matrix that defines the functional of complete type makes possible the robust stability analysis of periodic systems against time-varying uncertainties. In [11] and [13] this analysis is proposed by considering uncertain parameters. Here, the introduction of a new functional allows us to determine robust stability conditions for periodic systems with uncertain parameters, delay, and frequency.
Finally, we apply the results to the delayed Mathieu equation, which is a benchmark in the field. In particular, the delay Lyapunov matrix and robust stability bounds are obtained for the first time, up to our knowledge, for this equation.
The paper is organized as follows. The framework of Lyapunov–Krasovskii functionals is recalled in the next section. In Section 3, we present the new proposal to compute the delay Lyapunov matrix for time-delay periodic systems. A general robust stability analysis for this class of systems is carried out in Section 4. The robust stability results are applied to the well known delayed Mathieu equation in Section 5, and we end with some concluding remarks in Section 6.
Notation: represents the space of valued piecewise continuous functions on . The supremum norm is denoted by , where denotes the Euclidean norm for vectors. Inequality means that matrix is positive definite and and represent the minimum and maximum eigenvalue of matrix , respectively. Matrix denotes the identity matrix in .
Section snippets
Lyapunov–Krasovskii framework
We consider the following dynamical equation with one delay: where and are matrices of continuous coefficients with period , i.e. , , with range in , and is the delay. We denote the solution of system (1) with initial condition by , and its restriction to the interval by .
We say that system (1) is exponentially stable if there exist constants and such that
Computation of the delay Lyapunov matrix
In this section, a delay free PDE system for the computation of the delay Lyapunov matrix is presented. The introduction of a new set of boundary conditions allows us to give a procedure for computing the initial conditions that is simpler than the one in [11].
Robust stability analysis
In this section, we present a general robustness result for system (1), which covers uncertainties in the parameters, delays and frequency. We consider system (1) with a general perturbation term : Here, we assume that the functional is continuous with respect to both arguments, it is Lipschitz with respect to the second argument and it satisfies with , ,
Case study: delayed Mathieu equation
We consider the delayed Mathieu equation where . This interesting equation combines delays and the parametric excitation phenomenon, and it is the starting point for the stability analysis of some engineering models, such as the milling process [3].
The state space representation of Eq. (24) is given by system (1) with with the parameter numerical values
By constructing the stability map with the Matlab
Conclusion
We propose a new procedure for computing the delay Lyapunov matrix for periodic time-delay systems, and present a general robust stability analysis concerning uncertain parameters, delay and frequency by introducing a functional constructed by using the so-called functional of complete type. The new proposal for the computation of the delay Lyapunov matrix in combination with the robustness analysis enables us to compute robustness bounds on uncertainties of the well known delayed Mathieu
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This work was supported by the Project CONACYT, Mexico 180725 and Project SEP-CINVESTAV, Mexico 155.
References (19)
Quadratic Liapunov functionals for systems with delay
J. Appl. Math. Mech.
(1965)- et al.
A Liapunov functional for a matrix difference-differential equation
J. Differential Equations
(1978) - et al.
Lyapunov–Krasovskii approach to the robust stability analysis of time-delay systems
Automatica
(2003) - et al.
A numerical method for the construction of Lyapunov matrices for linear periodic systems with time delay
- et al.
Critical frequencies and parameters for linear delay systems: A Lyapunov matrix approach
Systems Control Lett.
(2013) - et al.
Introduction to the Theory and Applications of Functional Differential Equations
(1999) Delay Effects on Stability: A Robust Control Approach
(2001)- et al.
Semi-discretization for Time-delay Systems: Stability and Engineering Applications, Vol. 178
(2011) Introduction to Time-delay Systems: Analysis and Control
(2014)
Cited by (13)
Lyapunov stability tests for linear time-delay systems
2022, Annual Reviews in ControlCitation Excerpt :For a numerical approximation scheme and a detail smoothness analysis, see Jarlebring et al. (2011). The computation of the delay Lyapunov matrix is a challenging issue that is addressed in Gomez et al. (2019b) and Michiels and Gomez (2020). Next, we outline current ongoing work and topics we believe are promising directions of research in the present framework.
A hidden Markov model based control for periodic systems subject to singular perturbations
2021, Systems and Control LettersCitation Excerpt :As is well known that periodic behavior possesses the simplest time-varying character, which widely exists in many practical systems, as a result, periodic systems have been extensively implemented in many multiplexed systems, population dynamics, network communication protocol, for instance, [7]. In light of these reasons, many issues on periodic systems have been addressed, such as stability analysis, fault detection, optimal control/filtering [8,9]. Fruitful theories have been exhibited on the linear systems, but the periodic nonlinear systems remain interesting issues.
Closing the gap from nominal to robust asymptotic stabilization of time delay systems
2024, International Journal of Robust and Nonlinear ControlFault-alarm based hybrid control design for uncertain periodic piecewise time-varying systems with actuator constraints
2023, International Journal of Robust and Nonlinear ControlOn Asymptotic Stability of Discrete-Time Hybrid Systems
2023, IEEE Transactions on Circuits and Systems II: Express Briefs