Research On State Evaluation And Early Warning Methods Of Railway Signal Turnout Equipment

In recent years,with the rapid development of China’s high-speed railway,the number of railway signal equipment has increased year by year,and some defects are exposed in the traditional experience-based and periodic maintenance method gradually.The problem of "over-maintenance" and "under-maintenance" have become important factors affecting the safe operation of railway.How to realize the transition from "periodic repair","fault repair" to "state repair" and how to timely and accurately evaluate the state of signal equipment have become important measures to ensure the efficient and safe operation of railway.As basic signal equipment,its importance is self-evident.S700 K switch machine is the outdoor execution device of the switch,so the working environment is complex and the faults account for about70% of all the faults of the switch.If it has potential safety hazards or failures are not repaired in time,it will lead to very serious consequences.Aiming at the above problems,this thesis divides the state of the S700 K switch machine into fault state and degradation state in order to accurately evaluate the state of the switch,and carries out targeted research according to the characteristics of the two states.The main work of this thesis is as follows:(1)The common failure modes and mechanisms of S700 K switch machine are analyzed and summarized,and the power curve corresponding to each state of the switch is determined as the research object.(2)For fault data,the traditional feature extraction method cannot reproduce the features in the data well for small sample data due to the small number of fault samples in the field.Aiming at this feature,the Deep Belief Network(DBN)is adopted to extract the deep abstract features of the fault.At the same time,it is considered that the number of neurons in the hidden layer has a great influence on the effect of feature extraction and the number of neurons artificially set is difficult to give full play to the best performance of the network,so the Particle Swarm Optimization(PSO)is used to optimize the number of neurons in each layer of the network within the specified range to obtain the best stacking structure of the network.Finally,the Extreme Learning Machine(ELM)is used to classify and identify faults,which makes up for the time-consuming process of network feature extraction,and completes the fault diagnosis of turnout.(3)During the normal operation of the equipment,most of the faults are not sudden faults,but gradually degenerate from the normal state to the fault state as the equipment ages.For this kind of fault,there must be a certain connection between its intermediate state and the fault state.In order to find out the connection between the large amount of degraded data generated during normal operation and the normal and fault states,a clustering algorithm is used in this thesis to cluster the degraded data.The time domain features of the data are extracted in order to ensure the clustering effect,then Pearson Correlation Coefficient is used to establish the similar relationship between the degraded state and the normal and faulty states,thereby defining the degree of turnout degradation and calibrating the training data.The early warning mechanism and maintenance suggestions are established to prepare for the prediction of the next degradation state.(4)According to the state association constructed in the previous step,the Convolutional Neural Network(CNN)is used to automatically extract the characteristics of the data.Considering that the original data has certain timing characteristics,this article selects a gate that has certain advantages in processing time series.The Gated Recurrent Unit(GRU)replaces the fully connected layer in the network to predict each state.With its unique gate structure,a relatively good effect is finally achieved.To sum up,this thesis analyzes the characteristics of the two states and uses different models to complete the diagnosis and prediction of faults and degradation states,and then realize the assessment of the state of the turnouts.It provides a reference for further realizing the transformation from "fault repair" mode to "state repair" mode and improving the efficiency of railway operation.