Research On Fault Diagnosis Of Ultrahigh Pressure Abrasive Water Jet Cutting Machine Tool

Abrasive water jet cutting is a special processing technology which does not generate heat in the cutting process,so it has a wide range of applications.The cost of an ultra-high pressure abrasive water jet cutting machine is expensive,and its structure is complex.The coupling between components and components is closely related.If the failure occurs,and the operator only relies on his own experience to determine the cause of the failure,it will often make the ultra-high pressure abrasive water jet cutting machine not be diagnosed in time,resulting in serious damage to the equipment and bring serious damage to enterprises and personnel.Therefore,the technology of fault early warning and diagnosis for ultra-high pressure abrasive water jet cutting machine tool is more urgent.Aiming at the requirement of intelligent fault diagnosis for ultra-high pressure abrasive water jet cutting machine tool,a fault diagnosis method suitable for ultra-high pressure abrasive water jet cutting machine tool is developed by using sensor technology,artificial neural network model and optimized neural network model,and establishing an intelligent module system for machine tool fault prediction.Firstly,analysing the working principle of abrasive water jet and the characteristic parameters of components and common faults to select the characteristic parameters and fault types studied in this paper.Secondly,comparing the advantages and disadvantages of the existing intelligent fault diagnosis methods and selecteing the method of neural network is to analyze the fault diagnosis process.The original data of machine tools are collected by sensor technology.For the training data requirement,the sample data are sorted out and establishing the characteristic parameter matrix and fault type matrix.Then,by setting the number of nodes in each layer of BP neural network and coding the fault type,the sample data and corresponding fault type codes are used as the training data of BP neural network respectively,and then establishing the BP neural network model.Verifying the applicability and feasibility of fault diagnosis for abrasive water jet cutting machine tool by the test data.Then the RBF neural network and genetic algorithm are used to optimize BP neural network to improve the accuracy of fault diagnosis.By comparison,the optimized model of genetic algorithm is more suitable for the fault diagnosis of ultra-high pressure abrasive water jet cutting machine tool,and the diagnosis output of the optimized neural network is determined by corresponding decision-making method.Decision analysis is used to determine what type of fault is,so as to guide the maintenance personnel to repair the equipment.Finally,utilizing GUIDE toolkit of matlab,neural network model and the corresponding decision-making method is to design an intelligent module of fault diagnosis.