Research On Fault Diagnosis Of High Voltage Circuit Breaker Based On Spark Platform

With the rapid development of smart grids,power big data has entered a new era.The circuit breaker is an important device to ensure the stable operation of the power grid,and the high voltage circuit breaker(HVCB)fault diagnosis method can ensure the safe and stable operation of the power system.In the power system,through the initial failure planned maintenance to the current state maintenance of the fault found,the efficient operation of the circuit breaker can be ensured,and the economic loss caused by the planned maintenance can be reduced.Moreover,Spark is a memory computing platform for data processing.It has the characteristics of fast processing speed,diversified mode operation,universal functions,and support for multiple programming languages to process data.It provides for the processing and analysis of circuit breaker monitoring signal data.A new research idea was created.This article first introduces the basic structure and function of the circuit breaker and the working principle of the spring operating mechanism,analyzes the common fault types of the circuit breaker and the advantages and disadvantages of some common fault diagnosis methods.Through the interaction between the circuit breaker operating status signal and the fault type Corresponding relationship,HVCB fault diagnosis analysis method based on the Spark platform is proposed.Select the support vector machine(SVM)and naive bayes(NB)in the Spark machine learning library as the classification method of the circuit breaker fault diagnosis,preconditioning the travel signal data of the moving contact of the HVCB and select the characteristic value,and use the selected characteristic signal data as the experiment Input to simulate the fault state of the circuit breaker spring mechanism,and use it as the experimental output according to the fault label.Secondly,the SVM algorithm model constructed in Spark machine learning is used for data analysis and processing,and related performance indicators are obtained through model evaluation,and the classification and diagnosis results are output with the optimal model.On this basis,the parameter tuning of the algorithm and the comparison of several different kernel functions are carried out.And compared with traditional linear regression(LR)and NB classification model.The results show that through algorithm evaluation and model parameter tuning,it is found that the SVM model based on the radial basis kernel function has better accuracy.Compared with other kernel function models,the performance is better,and a better diagnosis effect is achieved.And the improved SVM model in Spark machine learning has improved the speed and accuracy of the LR and NB classification model in the fault diagnosis of the HVCB spring mechanism.The classification accuracy rate has reached 0.93,and the classification The diagnosis effect is more obvious,the running time is shorter than the NB model,which can meet the needs of fast data processing and high accuracy.Therefore,it is feasible to apply the improved SVM model in Spark machine learning to the fault diagnosis of HVCB spring mechanism.Combined with the model method of circuit breaker fault diagnosis,Lab VIEW is used to design and implement the fault diagnosis system.According to the related functions realized by the system,the diagnosis system is divided into four modules: serial port module,data reading,diagnosis algorithm and model curve display.Perform programming and interface design for each module,and run related programs.The results show that the interface can analyze and display the diagnosis results clearly and efficiently.