Research On Electrical Performance Simulation Experiment And Life Prediction Method Of Silver-based Contact Materials

Low-voltage electrical appliances are important components in various electrical systems.As the precision and complexity of various electrical systems increase,the reliability requirements for low-voltage electrical appliances are gradually increasing.As a part of the low-voltage electrical appliance that performs the breaking function,contacts have an important influence on the normal operation of low-voltage electrical appliances.Therefore,the reliability research on contact can improve the reliability and life of the low-voltage electrical appliance.In this paper,a simulation experiment system for the electrical properties of contact materials was designed and developed,and the electrical performance degradation data of the three contact materials of silver chromium oxide,silver tin oxide and silver nickel during the experiment was collected,and the degradation mechanism was analyzed.The life prediction of the silver nickel oxide contacts was completed based on the collected data.In this thesis,the push rod electromagnetic relay was used as the simulation object,and the mechanical action structure,measurement control circuit and corresponding human-computer interaction software of the contact material electrical performance simulation experiment system was designed.The mechanical structure part can realize the function of adjusting the mechanical parameters such as contact distance and overtravel,the circuit part can realize the functions of motion control,signal conditioning and data communication.The software can collect the original voltage and current signals,and analyze and calculate the parameters of contact static pressure,contact resistance,arcing time,arcing energy,rebound time,rebound energy and other parameters.The static pressure,contact resistance,arcing data,rebound data,etc.of the three contacts was compared and analyzed during the DC load,and the degradation mechanism of the contacts was analyzed and explained.Then a method for predicting the life of contact materials based on PSO-BP neural network was proposed.Firstly,contact resistance,static pressure,arcing time,rebound time and other parameters of the contact materials were pre-processed.Then the last bounce time,bounce energy,and bounce time were selected for prediction according to monotonicity,correlation,discreteness indicators.The model was built and trained through the PSO-BP neural network.The training results showed that the three selected performance parameters and neural network model have better prediction accuracy.Finally,in view of the problems that BP neural network was prone to overfitting and poor performance of the test set,a deep confidence network(DBN)was constructed to map the characteristics of performance data to high-dimensional space and use the high-dimensional feature data to make predictions.At the same time,the particle swarm optimization algorithm was used to determine the optimal number of DBN network nodes.The results show that the prediction accuracy of the five-layer DBN network was higher and the training time was moderate,which can predict the life of the contact material greatly.This study can provide a method for the simulation and experiment of the electrical properties of the contact materials used in electromagnetic relays.It also has practical value for the analysis of contact material failure and life prediction.