Research On Historical Data Comparative Analysis Technology Of Track Geometric State Detection

The historical data of the geometric size of the railway track reflects the quality of the track,and has a profound effect on the railway engineering department.The track detection technology can be divided into dynamic detection and static detection according to the principle of detection.Static detection is no load,using the track inspection instrument to detect the track;but dynamic detection is the existence of certain loads in the detection process,using high-speed track inspection car to detect the track;The two types of historical data are of guiding significance for all railway works.In order to make the full value of historical data and reliable to be used,the thesis studies three aspects: One is the mileage alignment of historical data at the level of data expression;Two is the prediction of track irregularities;Three is the self-diagnosis guiding ideas of historical data for rail inspection equipment.The research of this paper will provide a theoretical basis and reliable method for track maintenance and provide a train of thought for the full use of historical data value;the purpose is to ensure the high quality of the track,thereby prolonging its service life.In order to achieve the above goal,a specific repair operation is usually carried out on the track.In general repair work,it needs to be determined: operation time(repair time node),location(mileage)and content(correction effect).The core of getting the above information is the change rule of track irregularity.Prediction of track irregularities based on data of railway track geometric size is important for railway departments to grasp the development of track quality.However,the reference of historical data will be affected by mileage bias,such as the loss of mileage and cumulative mileage error,the data may be inconsistent with the results of each inspection,which is the data waveform cannot be aligned.the paper predicts the track quality index(TQI)as the observation value,it is proposed that multiple sets of original data be sequentially verified by one step,the data waveform will be evaluated with each other by inter-correlation coefficient,the effective and reference values can be acquired after the mileage of the original data of each group is aligned.In the content of track irregularities prediction,two methods are provided: One is based on the dynamic test data,the prediction is carried out through the Auto-regressive Integrated Moving Average Model(abbreviated as ARIMA model);The other is based on static data,through its data performance,we find the function relation with the highest matching degree,and predict the development of track irregularity.At the same time,the static data not only can guide the track adjustment,but also can reflect the abnormal working condition of the corresponding sensors on the track inspection instrument when the track is not adjusted.Finally,through the historical data provided by Huizhou railway section of the Guangzhou railway group,after aligning its mileage,the TQI data with reference can be obtained,through the ARIMA model of model driven,the mean of relative error between the predicted results and the measured values is 1.75%;According to the idea of data driven,the confidence level of the prediction result of the distribution function can reach 92%.;Through the historical data provided by Jiangxi EVERBRIGHT Measurement and Control Technology Limited Company,referring to industry standards,it can be reversed to determine whether the track is adjust between the two checks;At the same time,when the corresponding sensor is abnormal,the historical data shows that there will be a discrete interval around the agreed true value,in this paper,a case of fiber-optic gyroscope is used to study the accumulated static detection historical data of the left track alignment orientation;The absolute value of the data residuals is compared with the industry standard of the track inspection instrument corresponding to the project,and the qualification of the sensor corresponding to the data can be qualified;Based on it,In order to capture the characteristics of historical data and to provide the train of thought for the self-diagnosis of rail inspection equipment.