Mohamed Naceur Abdelkrim

Bond graphs for the diagnosis of chemical processes

Abstract The paper deals with the use of coupled bond graph as an integrated decision tool for health monitoring of chemical reactors. A bond graph model based diagnostic strategy is adopted to detect and isolate kinetic and thermodynamic drift of chemical reactors due to the appearance of secondary reactions. Furthermore, the causal and structural properties of the bond graph model are used not only for system dynamics simulation but also to design Fault Detection and Isolation (FDI) algorithms (i.e. the generation of formal fault indicators) for online supervision of faults affecting sensors, actuators, physico-chemical components and phenomena.

Reliable H ∞ control of multiple time scales singularly perturbed systems with sensor failure

This paper studies the reliable H ∞ control for linear time-invariant multiparameter singularly perturbed systems against sensor failures. By time-scale decomposition, the full-order system is decomposed into slow and fast subsystems. After designing a reliable H ∞ controller for the global system, three reduced reliable H ∞ sub-controllers based on the slow and fast problems are obtained through the manipulation of the algebraic Riccati equations. The resulting control systems are reliable in that they provide guaranteed asymptotic stability and H ∞ performance when all control components are operational as well as when sensor failures occur.

Model-based fault diagnosis of two-time scales singularly perturbed systems

The aim of this paper is to detect and to isolate sensor faults in singularly perturbed systems (SPS), with emphasis on the use of slow subsystem. Firstly, we considered the fast subsystem as a modelling error by the approximation of the SPS with the slow subsystem. Structured residuals were generated, using robust parity space, to the fault diagnosis of the SPS. A second approach was to design a set of observer-based residuals based on the slow subsystem. In order to validate the proposed approaches, we applied the results on a F-8 aircraft example.

UIO-based fault detection and isolation for a class of neutral system

The purpose of this paper is to provide new methods for neutral system stability and diagnosis analysis. Based on Kharitonov stability, a new theorem is proposed for delay neutral system with robust control. For stabilized neutral system, an unknown input observer with new structure is then investigated to diagnosis this type of system. Tow important conditions are presented to can apply this type of observer. Analysis of the simulation results reveals that the application of proposed theorem and approach provide a comprehensive result for fault neutral system stability and diagnosis.

Pseudo bond graph for fault detection and isolation of an industrial chemical reactor part II: FDI system design

This second part of the two series papers deals with the problem of bond graph model based fault diagnosis for a class of chemical reactions taking place in a jacketed reactor. A bond graph model based fault detection and isolation approach applied to design of a supervision system for a complex chemical reactor is presented. The nonlinearities and energetic couplings involved in the dynamics require a suitable supervision environment. Furthermore, the causal and structural properties of the bond graph tool are used to design Fault Detection and Isolation (FDI) algorithms (i.e. the generation of fault indicators) to determine the availability of the chemical process. Thus, a unified framework is developed to take into account faults affecting sensors, actuators and process. Structural analysis of the model is utilized to obtain the fault signatures for fault isolation. The decision procedure to detect and isolate the faults uses the residuals, evaluated online. A detailed kinetic modelling approach had been proposed in previous work to simulate chemical reactions and analyze the appearance of secondary reactions under industrial conditions in this present paper.

Robust fault diagnosis of chemical system by using uncertain bond graph model

In this paper, robust Fault Detection and Isolation (FDI) design of nonlinear uncertain dynamic system in presence of chemical and thermodynamic phenomenon is addressed. The methodology using a Bond Graph (BG) representation in linear fractional transformation (LFT) form is shown to be a valuable tool for developing dynamic threshold generators and achieving robustness against model uncertainty in combination with sensitivity to faults. Experimental results are given to illustrate the design procedure and to demonstrate the performance using to detect kinetic and thermodynamic drift due to appearance of secondary reaction via a real continuous reactor with heat exchange.

FDI of Chemical Process using Bond Graph Tool

Abstract In the chemical industry, faults can occur due to sensors failures, equipment failures or changes in process parameters. This paper presents a diagnosis methodology relying on a Bond Graph approach to detect kinetic and thermodynamic drift of chemical reactors due to appearance of secondary reaction. The Bond Graph model, used for supervision, incorporates the chemical kinetics of equilibrated reaction, heat and mass transport phenomena. Furthermore the causal and structural properties of the Bond Graph tool can be used to design FDI algorithms (i.e. the generation of fault indicators) to determine the availability of the chemical process. Structural analysis of the model is utilized to obtain the fault signatures for fault isolation. The diagnosis strategy is illustrated via a Continuous Stirred Tank Reactor (CSTR) example.

Attractive Force Based on Fractional Potential in Dynamic Motion Planning for Mobile Robot

In path planning, potential fields introduce force constraints to ensure curvature continuity of trajectories and thus to facilitate path-tracking design. In previous work, a path planning design by fractional (or generalized) repulsive potential has been developed to avoid fixed obstacles. A fractional road was determined by taking into account danger of each obstacle. Danger level of each obstacle is characterized by the fractional order of differentiation. Then a new attractive force based on fractional potential was introduced, with the aim to extend the method to dynamic obstacles characterized by fractional potential fields. In this paper a comparison between each classical attractive force is presented. Then the new attractive force based on fractional potential is introduced. Finally, this attractive force is applied on an example and compare with the Ge and Cui method which is the classical potential fields for motion planning of mobile robots in a dynamic environment.Copyright

Luenberger Observer based Sensor and Actuator Fault Detection for Chemical Reactor

Abstract Abstract In this paper, we propose the application of a principal observer based fault diagnosis approaches for chemical process. The Continuous Stirred Tank Reactor (CSTR) model with parametric uncertainties is linearized around a chosen steady state. A Lunberger observer is employed to generate conversion estimates for the CSTR using only temperature measurements. Results of numerical simulations of a no isothermal CSTR with coolant jacket dynamics are presented. An average robustness with respect to interpolation errors, disturbances and model parametric uncertainties is achieved by using the obtained Lunberger observer only for the sensor faults. But the detection of actuator fault is not achieved by this approach.

A MIMO Time-Varying System Control Via a Stable Dynamic Inversion Methodology: Case of an Induction Machine

In This paper, we study a MIMO nonlinear systems control via the linear time-varying form and stable dynamic inversion methodology. We apply a poles placement state feedback control of MIMO nonlinear systems of an induction machine using trajectory linearization control. The nonlinear system is linearized along a nominal trajectory to yield a MIMO linear time-varying one which is transformed into the controller canonical form through a Lyapunov coordinate transformation, which serves to compute the control input design. The nominal trajectory is not easy to calculate because systems are governed by many physical constraints and the requested movement that the system needs to do must be, first of all, feasible. So a stable inversion design has been used to generate the nominal trajectory from the desired output. The effectiveness of the developed approach is checked on a real speed trajectory tracking of an asynchronous motor.