Research On Abrasive Belt Wear Monitoring System Based On Rail Grinding Car

The rapid development of my country’s high-speed railway construction,and the continuous increase in vehicle weight and traffic density have made the problem of rail wear increasingly prominent.At present,rail grinding has been regarded as the most important means of line maintenance.The commonly used rail repair methods are mainly grinding wheel grinding and milling cutter milling.In recent years,abrasive belt manufacturing technology has become more mature,and abrasive belt grinding has shown high-quality characteristics such as high efficiency,environmental protection and economy,and has been gradually promoted and applied to small rail grinding equipment.The service life of a single belt is short,and the grinding performance of the belt changes significantly with different wear conditions.In order to better exert the performance of the abrasive belt and avoid problems such as belt fracture and rail damage caused by excessive wear,this article is based on the actual working conditions of rail abrasive belt grinding,and explores the relationship between the current signal change of the abrasive belt drive motor and the abrasive belt wear state.Through the establishment of a monitoring system,the current signal characteristics are used as input to determine the current wear state of the abrasive belt.The specific research content is as follows:The characteristics of rail abrasive belt grinding are analyzed and compared with ordinary grinding methods.The differences are mainly reflected in the processing methods,process parameters and working conditions,and quality acceptance standards.On this basis,the different wear stages of the abrasive belt in the whole life cycle are introduced,and the main wear forms of the abrasive belt in each stage are analyzed.Starting from the working principle of the grinding unit and the motor,the relationship between the change of the phase current of the belt drive motor and the grinding process parameters is explored by establishing the mechanical equation.Comparing the different characterization methods of the abrasive belt wear degree,it is decided to use the abrasive belt wear mass percentage to quantify the abrasive belt wear degree.The pre-experiment for rail belt grinding is designed.Based on the results of the experiment,a preliminary analysis of the status information of the belt grinding is carried out.According to the analysis results of the preliminary experiment,the experimental process parameters were determined,and a formal experiment was carried out to obtain the current signal sample set.By comparing the current information of the motor idling process and the grinding process,the main frequency of the signal and the noise component below 1000 Hz are analyzed.After the original current signal is filtered and processed by the RMS value method,the time domain,frequency domain,and time-frequency domain features are extracted.From the perspective of abrasive wear,the reasons for the related changes in signal characteristics are analyzed.Correlation analysis is made between the extracted signal features and the abrasive belt wear state,and the signal features that have a moderate or above correlation with the abrasive belt wear are used as the classification basis for the abrasive belt wear state,and are randomly divided into training samples and test samples.The machine learning algorithm based on support vector machine and the machine learning algorithm based on BP neural network are used for model training respectively to distinguish the wear state of the abrasive belt.The accuracy of the two recognition methods reached 92.59%and 97.53%,respectively.The abrasive belt wear monitoring system was built based on Lab VIEW,which completed the functions of signal acquisition and preprocessing,time domain analysis,frequency domain analysis,time frequency domain analysis and abrasive belt wear monitoring.Introduced the usage of different modules of the system and the functions used to realize different functions.Finally,the online monitoring function of abrasive belt wear based on BP neural network is realized by calling MATLAB function in Lab VIEW,and the system performance test is completed,and the result proves that the system is reliable.