Research On Fault Classification Of ZPW-2000A Track Circuit Based On Fuzzy Cognitive Map

ZPW-2000 A track circuit is an important part of the railway transportation equipment,ensuring the safe operation of the train.When the track circuit failure,it will lead to a series of problems,such as train collision,train rear-end,economic losses,and even personal safety.Therefore,it is very necessary to timely carry out the maintenance of equipment of track circuit faults to ensure the safe operation of the train and improve the efficiency of railway transportation.Currently,the fault detection and discrimination of track circuit is still manual test and analysis in present,which could have several disadvantages such as a lot of time is spent in determining the cause of the fault,low fault detection efficiency,large amount of labor in present and error-prone while analyzing.In view of above drawbacks,base on the monitoring data of ZPW-2000 A track circuit,rough set theory and Fuzzy Cognitive Map(FCM)algorithm are combined,which are applied to fault classification for track circuit to classify the faults in ZPW-2000 A track circuit in this dissertation.Moreover,according to the various faults,some maintenance advices are given.The main contents of this paper are as follows.Firstly,the basic theory of ZPW-2000 A is deeply studied.Various equipment voltage of track circuit is analyzed base on monitoring data of ZPW-2000 A track circuit.The main track input voltage,the minor track input voltage,the main track adjust output voltage,the minor track adjust output voltage,the track relay voltage,the minor track relay voltage,check condition of the minor track,the merit leaves voltage of transmitter and the power of transmitter are selected as condition attributes,which are extracted to complete feature extraction.And according to fault data determine fault types,the track circuit fault information system is constructed.Secondly,in allusion to continuous attributes in the characteristic attribute data,an information entropy discretization algorithm is applied to get discrete data.Meanwhile,in order to solve these problem such as more attribute nodes,computational complexity and long time of the algorithm on computer during Fuzzy Cognitive Map is built,an improved principal component heuristic reduction algorithm is proposed to attribute reduction of discretization information system,achieving data dimensionality and the characteristic of the secondary extraction.The simulation results show that the classification indexes of track circuit failure data are improved by reduction.Thirdly,the several different methods for weight matrix solving have been compared to select Least Squares(LS)algorithm to solving weight matrix of FCM.Meanwhile,the LS-FCM classification model is constructed.Finally,a fault classification model for track circuit is proposed,which is base on rough set theory and LS-FCM.According to track circuit test data of a signaling depot history data,and compared with the classification model for track circuit based on LS-FCM,the experimental results show that rough set theory and FCM algorithm are applied to fault classification for ZPW-2000 A track circuit has a higher classification accuracy.