Research On Equipment Fault Prediction Technology For Automatic Gate Machine Of Rail Transit

The development of urban rail transit has led more and more passengers to choose rail transit as a way to travel,automatic fare collection system directly serve the passengers.The state of the equipment has directly impact on the service level.High-intensity operation increases the equipment failure rate.At present,the equipments of automatic fare collection adopt the mode of failure maintenance and regular maintenance,which leads to high safety and high cost.In view of this,the concept of stata maintenance is proposed in this thesis,and the Automatic Gate Machine(AGM)is chosen as the research object.The research mainly study the technology of failure prediction,which can realize the active maintenance of the equipment and to ensure the perfect equipment.It is of great significance to study the topic.Failure prediction needs to understand the mechanism of the failure of the equipment.This paper first analyzes the classification,operation mode,equipment composition,function,working principle,etc.of the AGM,and has a comprehensive understanding of it.On this basis,the degradation of equipment state,the failure rate curves of mechanical components and electronic components are qualitatively analyzed,the reliability indices and life distribution rules of the equipment are quantitatively analyzed.The influencing factors of failure are analyzed from four aspects of human-machine-ring-pipe,and combines the actual failure records of automatic ticket machines in Nanjing Rail Transit Station,summarizes the equipment failures into reader failure,sensor failure recovery module,passenger passage display failure,gatedoor failure,communication failure and then analyze the various failure mechanisms.The failure condition information of the equipment can be acquired with reliable instruments,which can help to achieve state maintenance.This paper puts forward the real-time status detection in the automatic gate machine equipment to achieve the process for the follow-up equipment maintenance management to provide guidance.Provide improving guidance for the management of the maintenance of the follow-up equipment.In the study of the failure prediction model construction of the equiment,the long and short memory recurrent neural network is selected as the prediction method,and elaborate the process of building the fault prediction model.Passenger flow,trouble free operation time,failure times,temperature and season are selected as the input of failure prediction of circul neural network,the fault sequence of equipment is taken as output,and then establish parameters of per layer,such as the number of neurons,activation function,learning rate and activation times,then the BP-LSTM and PSO-LSTM are established.Finally the correlation coefficient,recall rate and accuracy are selected as the indexes to evaluate the performance of the model.The validity and accuracy of the two models are verified by the example of cross-validation algorithm.It is concluded that the two models have certain accuracy.The failure prediction model optimized by PSO has better prediction performance.The failure prediction model optimized by PSO has better prediction performance.The failure prediction technology can guarantee the countermeasures before the failure,and greatly improve the maintenance and management level of the equipment.