Research On Fault Warning Method Of Mine Dry Transformer Based On Grey Model

As an important part of the underground power supply system,the running state of mine dry type transformer affects the safety and reliability of power supply system.At present,the maintenance workers mainly adopt the way of power outages and preventive maintenance of the preventive test to evaluate the state of transformer,not only with over repair and under repair conditions,but process a lower accuracy when it response the operation condition of transformer.The working environment of mine dry transformer is mean,there are many complex factors in the operating environment,and there is lack of effective operation state and fault prediction methods.Therefore,the research of mine dry transformer early warning system is of great practical significance.In this paper,by reading relevant references and taking grey system theory as the basis of research,the structural characteristics of the prediction algorithm,the operating process and the calculating methods are discussed.Based on the key points of optimization methods,the improvement thoughts and the inherent errors in the structure of grey model are discussed.The common faults of mine explosion-proof dry transformer and its developing process are described in detail.In order to achieve efficient and accurate transformer fault warning,based on GM(1,1)this paper firstly introduced Markov factor V and GNNM(1,1)to reconstruct the accumulation sequenceX⁽⁸⁾⁽¹⁾,with the introduction of Markov factor V,the GM M(1,1)model is obtained,which combined the no aftereffect of Markov chain and the new information priority principle,and improved the sensitivity of grey model to abnormal data.Then based on particle swarm optimization,the GM M(1,1)model is optimized and improved.Finally the G M M(r,1)transformer fault prediction algorithm is designed,which with highly prediction accuracy.In order to demonstrate the validity of GM M(r,1)transformer fault prediction algorithm,this paper simulated the winding fault of transformer.The simulation results of GM M(r,1)transformer fault prediction algorithm shows that:The GM M(r,1)fault prediction algorithm,which based on the Particle Swarm Optimization algorithm runs stably and with a better accuracy.The potential transformer faults that may occur in the future can be predicted by using GM M(r,1)prediction algorithm,and the overall data fluctuation trend can be reflected in the prediction curve.In addition,the abnormal data can be reflected in the prediction curve in advance when the original data fluctuates,which can meet the accuracy requirement of the dry-type transformer fault pre-warning,achieved the purpose of extending the service life of the mine flameproof dry-type transformer and improving the actual operation efficiency of the equipment.