Research On Fault Diagnosis Of Causality Diagram Based On Petri Nets

Fault diagnosis technology is a widely used engineering science.It has an inseparable relationship with the actual industry and is the crystallization of contemporary industry growth.Industry practice proves that exploring fault diagnosis technology has unparalleled value to the production and life of contemporary society.Causality diagram theory is based on belief network,which is a new theory to overcome its shortcomings.It can be used for troubleshooting.This paper mainly focuses on causality diagram theory,combines Petri net theory and BAM neural network theory to study new algorithms for minimizing cut sets and methods for reducing the number of inspections during fault diagnosis.The main contents are as follows:(1)Design a new algorithm to find the minimum cut set of the causality diagram.Combining causality diagram theory with Petri net theory,the causality diagram is converted into Petri nets,and the algorithm of the basic library is defined.A new algorithm is designed to solve the minimal cut set of causality graphs.The algorithm omits the steps of cutting a cut and a final cut,saving time for fault diagnosis.At the same time,two examples are used to demonstrate the specific implementation process of the algorithm and verify its application and effectiveness.(2)Reduce the number of inspections during fault diagnosis.After obtaining the minimum cut set,combined with the relevant knowledge and theory of BAM neural network,the neural network training samples are obtained through the obtained minimum cut sets;an algorithm is designed according to the training samples to obtain the associative memory matrix;The state of the monitoring point is used as input,and the output vector is obtained using the formula;finally,the output vector is observed to determine the cause of the failure(an event with a median value of 1 indicates occurrence).