A GMDH neural network-based approach to robust fault diagnosis : Application to the DAMADICS benchmark problem

This paper focuses on the problem of designing a robust fault detection scheme with group method of data handling (GMDH) neural networks. One of the main objectives is to show how to determine the structure and parameters of the neural network as well as how to estimate the modelling uncertainty of the resulting neural model. It is shown that the algorithms proposed to tackle the above tasks make it possible to obtain a suitably accurate mathematical description of the system. This description allows developing a technique of generating an adaptive threshold that permits robust fault detection. Another objective is related with an application-oriented study regarding the proposed approaches. In particular, the paper presents the results concerning the application of the techniques and algorithms for the modelling and fault detection of a valve actuator (DAMADICS benchmark).

[1]  R. Fletcher.  Practical Methods of Optimization , 1988 .

[2]  Michel Kinnaert,et al.  Diagnosis and Fault-Tolerant Control , 2004, IEEE Transactions on Automatic Control.

[3]  J. P. Norton,et al.  Fast and robust algorithm to compute exact polytope parameter bounds , 1990 .

[4]  Nariman Sepehri,et al.  Diagnosis of process valve actuator faults using a multilayer neural network , 2003 .

[5]  Duc Truong Pham,et al.  Neural Networks for Identification, Prediction and Control , 1995 .

[6]  J. Norton,et al.  Bounding Approaches to System Identification , 1996 .

[7]  Petre Stoica,et al.  Decentralized Control , 2018, The Control Systems Handbook.

[8]  Paul M. Frank,et al.  Issues of Fault Diagnosis for Dynamic Systems , 2010, Springer London.

[9]  Thomas Parisini,et al.  Keynote paper: Fault diagnosis and neural networks: A power plant application , 1995 .

[10]  Paul M. Frank,et al.  Robust Observer-Based Fault Diagnosis in Non-Linear Uncertain Systems , 2000 .

[11]  W. Cholewa,et al.  Fault Diagnosis: Models, Artificial Intelligence, Applications , 2004 .

[12]  Marcin Witczak,et al.  Genetic programming based approaches to identification and fault diagnosis of non-linear dynamic systems , 2002 .

[13]  Jie Chen,et al.  Robust Model-Based Fault Diagnosis for Dynamic Systems , 1998, The International Series on Asian Studies in Computer and Information Science.

[14]  Eric Walter,et al.  Identification of Parametric Models: from Experimental Data , 1997 .

[15]  Paul M. Frank,et al.  Nonlinear observers for fault detection and isolation , 1999 .

[16]  N. T. Russell,et al.  Modular neural network modelling for long-range prediction of an evaporator , 2000 .

[17]  Joseba Quevedo,et al.  Development and Application of Methods for Actuator Diagnosis in Industrial Control Systems (DAMADICS): A Benchmark Study , 2003 .

[18]  Sylviane Gentil,et al.  Reformulation of parameter identification with unknown-but-bounded errors , 1988 .

[19]  Józef Korbicz,et al.  Observers and Genetic Programming in the Identification and Fault Diagnosis of Non-Linear Dynamic Systems , 2004 .

[20]  Józef Korbicz,et al.  A novel genetic programming approach to nonlinear system modelling: application to the DAMADICS benchmark problem , 2004, Eng. Appl. Artif. Intell..

[21]  P. Laycock,et al.  Optimum Experimental Designs , 1995 .

[22]  Zbigniew Michalewicz,et al.  Handbook of Evolutionary Computation , 1997 .

[23]  Alan F. Murray,et al.  Confidence estimation methods for neural networks : a practical comparison , 2001, ESANN.

[24]  Sylviane Gentil,et al.  Recursive membership estimation for output-error models , 1990 .

[25]  Javier Segovia,et al.  Parameter estimation of dynamic GMDH neural networks with the bounded-error technique , 2002 .

[26]  H. Nijmeijer,et al.  New directions in nonlinear observer design , 1999 .

[27]  Joseba Quevedo,et al.  Introduction to the DAMADICS actuator FDI benchmark study , 2006 .