Title :
Inverse Fault Detection and Diagnosis Problem in Discrete Dynamic Systems
Author :
Li, Wei ; Shen, Hao
Author_Institution :
Hangzhou Dianzi Univ., Hangzhou
Abstract :
This paper investigates an inverse fault detection and diagnosis problem in discrete dynamic systems. The problem is how to adjust the system parameters according to observation value of inputs and outputs so that the system is concordant. First we formulate the problem as a least square problem with interval coefficients. Then two algorithms for this problem are presented. The first algorithm based on the expected value of observation value of inputs and outputs. We are only required to solve a classical least square problem in this algorithm and the algorithm is robust. The second algorithm by using linear programming approach can deal with large scale systems and suit for on line adjustment.
Keywords :
discrete systems; fault diagnosis; large-scale systems; least squares approximations; linear programming; diagnosis problem; discrete dynamic systems; interval coefficients; inverse fault detection; large scale systems; least square problem; linear programming; Cybernetics; Electrical fault detection; Fault detection; Fault diagnosis; Least squares methods; Machine learning; Optimization methods; Power system dynamics; Power system protection; Power system relaying; Discrete dynamic system; Fault detection and diagnosis; Inverse problem;
Conference_Titel :
Machine Learning and Cybernetics, 2007 International Conference on
Print_ISBN :
978-1-4244-0973-0
Electronic_ISBN :
978-1-4244-0973-0
DOI :
10.1109/ICMLC.2007.4370307
