Djamel Eddine Chouaib Belkhiat

Robust Fault Detection Based-Observer for Linear Switched Systems

Abstract This paper deals with the design of a robust hybrid observer for switched linear systems with unknown inputs and modeling error. The proposed observer is synthesized for the task of robust fault detection and concerns the case when the active mode is unknown. The basic idea of our approach is to formulate the design of the robust fault detection observer as a H∞ model-matching problem. Then, some results of H∞ optimization are exploited and an LMI solution procedure is proposed to synthesize the desired robust observer.

A fault detection & isolation scheme for discrete time Switched Linear Systems

In this paper, hybrid observers based approach for faults detection and isolation, for a class of MIMO hybrid systems, is proposed. The faults herein considered can affect the sensors or the discrete trajectory of the considered system. The proposed approach is based on a Dedicated Switched Robust Observer Scheme (DSROS) using three blocks in interaction. A block for the mode signatures, a block for the fault signature a block for the diagnosis of the discrete events. Some LMIs conditions are provided to guarantee both of the robustness and the convergence of the proposed observers. Finally some simulation results illustrate the efficiency of the proposed approach.

Robust H ∞ tracking control design for a class of switched linear systems using descriptor redundancy approach

In this paper, the design of a switched Proportional-Derivative (PD) controller, for a class of Switched Linear Systems (SLS) subject to external disturbances, is investigated for the purpose of satisfying the robust H∞ output feedback tracking performance. The main idea of the proposed synthesis approach consists to use the descriptor redundancy formulation in order to decouple the crossing terms between the controllers and the switched systems matrices. Based on the multiple Lyapunov functional methods, sufficient conditions for the existence of a switched PD controller are formulated in terms of Linear Matrix Inequalities (LMI). The efficiency of the proposed synthesis procedure is illustrated by a numerical example.

Identification of Continuous Time Linear Parameter Varying models based on Reinitialized Partial Moments

In this paper we propose a practical identification solution for Continuous-Time (CT) Input-Output (IO) Linear Parameter Varying (LPV) systems. The main steps of the proposed method are: (a) CT-Reinitialized Partial Moments (RPM) is used as modeling tools; (b) the use of a Kronecker product framework to obtain linear regression form; (c) a least squares identification algorithm is used to estimate the parameters. Since the least squares technique does not give consistent estimates, the instrumental variables method is proposed to achieve this property. The model can represent both linear and non-linear behavior with polynomial scheduling parameters dependency. We assume that both the IO and the scheduling parameters are measured. Monte Carlo simulation analysis demonstrates the performance of the proposed identification approach.

Fuzzy quasi LPV model for a cross flow heat exchanger

This paper presents a new dynamic model called fuzzy quasi LPV model which simulate accurately the temperature and flow transients in a cross flow heat exchanger (CFHX). Motivated by the ability of fuzzy tools to approximate and management of nonlinearities, and based on the interpretation of parameters in Takagi-Sugeno (TS) models, we present a new procedure for identifying nonlinear models of linear parameter varying type. The fuzzy quasi LPV model is compared to a numerical model of a cross flow heat exchanger. Comparisons indicate that the developed fuzzy quasi LPV model is capable of predicting the transient performance of the heat exchangers satisfactorily.

Design of a robust observer-based FDI scheme for a class of switched systems subject to sensor faults

This paper concerns the Fault Detection and Isolation (FDI) problem for a class of Switched Linear Systems (SLS) subject to sensor faults and unknown bounded disturbances. The main work is based on the design of a generalized switched observer scheme. The FDI problems have been solved by using a robust control techniques. A suitable trade-off between the robustness to disturbances and the sensitivity to sensor faults was obtained. The main results are reformulated by using Linear Matrix Inequality (LMI) formulation. An example is included to illustrate the efficiency of the proposed approach.

Design of a Robust Tracking PD Controller for a Class of Switched Linear Systems with External Disturbances

This paper concerns the design of a robust output feedback tracking controller for a class of Switched Linear Systems (SLS) subject to external disturbances. The proposed synthesis approach, based on a descriptor redundancy formulation, allows to avoid of the crossing terms appearance between the switched Proportional-Derivative (PD) controllers and the switched systems matrices. Using the multiple Lyapunov functional methods, a robust H¥ output feedback tracking performance has been formulated in terms of set of Linear Matrix Inequality (LMI). The effectiveness of the proposed synthesis procedure has been illustrated by a numerical example