Research On Fault Diagnosis System Of JH39 Mechanical Press Based On Fault Tree Inversion

Because the JH39-800 mechanical press has many fault type and difficult to diagnose,this paper study and discuss a method of fault diagnosis and then create a fault diagnosis system,which make quick judgments on the abnormality of the lubrication system,and find the fault timely.This paper start from the principle of JH39-800 mechanical press parts,then get familiar with the whole system,understand the function and principle of the lubrication system in detail,collect the basic information and data,analysis the common faults of each component and explore the fault reason,build the fault tree,do the qualitative and quantitative analysis.BP neural network algorithm is constructed,train the BP neural network and simulated it in MATLAB.The fault diagnosis is carried out according to the detected fault signal,and the failure probability will be calculated by using the fault tree inversion method.Establish the database system of JH39-800 mechanical press fault diagnosis,then take the BP neural network as the core,make the fault diagnosis procedures based on the databse,combine the TWINCAT2 automation software to construct the fault diagnosis expert control system with BECKHOFF system.Deal with the fault signal through some module of BECKHOFF,and take the signal into the diagnosis system to get the diagnosis results.Compare the experimental results with the simulation results,analysis the ability of diagnosis system to identify all kinds of fault signal,determine the effect of the fault diagnosis expert system.The experimental results show that the fault diagnosis system has a certain ability to identify the fault signal,it can effectively judge the fault which is helpful to judge the internal faults of JH39-800 mechanical press.At the same time,this method can also be applied to other mechanical fault diagnosis.