Study On Motor Fault Diagnosis Based On Combined Method Of Cloud Computing

Nowadays,along with the continuous development of society economy,science and technology,all kinds of motors are playing an increasingly important role in industrial production and people's daily work.However,Once the motor fault occurs,the light will affect people's production and life,and will endanger people's lives as well as cause serious economic losses.Therefore,in order to meet the superior quality needs of industrial automation on the motor,in modern life,it is of great practical significance to research a method to diagnose the fault of motor.With the continuous development of science and technology,the motor diagnosis needs more accuracy and rapidity,there are some technical problems in the field of motor fault diagnosis.For example,the feature extraction of high dimensional fault data is not accurate,which leads to the problem of the low diagnostic accuracy,the limitation of a single diagnostic method,and the low computational efficiency,etc.Based on the above conditions.In this dissertation,we study the isospectral manifold learning algorithm and least squares support vector machine and wolf pack algorithm and combined diagnosis model cloud computing to solve the problem.Finally,the data of Case Western Reserve University are used as examples to simulate and verify.Specific contents are summarized as follows:Firstly,This dissertation aimed at the motor fault when high-dimensional data to extract the feature set is not precise diagnosis leads to the problem of low precision,we introduce a isospectral manifold learning algorithm to reduce the dimension of high-dimensional.This algorithm is used to reduce the dimensionality of the data processed by principal component analysis two times.At the same time,using the modified sparse reconstruction weight matrix to construct adjacency graph,so that more similar samples gathered after the dimension reduction,different samples are more evacuated,efficient realize the high dimension data redundancy reduction.Finally,It is compared with the principal component analysis,and the effect is remarkable.Then,the accuracy of the diagnostic is affected by the isospectral manifold learning algorithm and the parameters of the least squares support vector machine,the Wolf Pack Algorithm is used to optimize the parameters.The Fisher criterion function is used to optimize the combination parameters of the least squares support vector machine,and the fitness function used in optimizing the parameters of the isospectral manifold learning algorithm is the recognition rate of the nearest neighbor classification,the optimal diagnosis model is established by using the optimized parameters.Simulation results show that the model has good diagnosis results,Wolf Pack Algorithm and particle swarm optimization algorithm are compared and analyzed when the parameters are optimized.The diagnosis result is remarkable.Finally,according to the diversity of motor fault symptom and the limitations of a single diagnostic method,In this dissertation,using a combined diagnosis model is used to solve the problem,In this dissertation,support vector machine,fuzzy neural network and RBF neural network are combined and the optimal combination model is formed according to the minimization of the sum of diagnostic squared error,The experimental results show that the combined diagnostic model can make up for the defects of a single method,and compared with these three single diagnosis methods,it has higher fault recognition rate and better robustness.Finally,the cloud platform technology is used to solve the problem of long running time of the combination complex model,the efficiency of fault diagnosis is improved effectively.