Research on Grey Prediction of Electromagnetic Relay Contact Resistance

Abstract

When the relay contacts are switched on the load circuit, contact pairs on-state directly reflects the relays reliability, and it also indirectly reflects machine performance of the relay. Therefore, the static contact resistance is one of the key features of the relays state estimation and failure warning. However, the contact resistance of the relay contact is uncertainty characteristics, small-scale sequence trend model is difficult to obtain an ideal prediction performance, and therefore grey model which based on small-scale sequence for prediction of electrical contact reliability of the relay is drawn on. Using the static contact resistance measurement value of the relay contact resistance as training samples to set up GM (1, 1) model and obtaining development grey numberaand endogenous control grey numberbby least square method, so as to get differential equation of original datas predicted value and then forecast the original data. In order to improve the prediction accuracy and weaken the influence of randomness, sliding smoothing formula to the original data is used to get improved prediction model. By comparing the difference between the accuracy of improved model and not improved model, it concludes that the improved model has possibility and accuracy higher.