Search

CN-121980308-A - Subway steel rail insulation degradation positioning method, device, computing equipment and storage medium

CN121980308ACN 121980308 ACN121980308 ACN 121980308ACN-121980308-A

Abstract

The application provides a subway rail insulation degradation positioning method, a device, a computing device and a storage system, wherein the subway rail insulation degradation positioning method comprises the steps of establishing a simulation model based on a topological structure of a target subway line, generating a first rail potential distribution model based on a constructed training data set, extracting a first rail potential gradient characteristic quantity reflecting an insulation state, and determining an optimal rail insulation defect detection threshold value through an optimization algorithm; acquiring site potential data, constructing a second steel rail potential distribution model, extracting second steel rail potential gradient feature quantity through multi-resolution spatial derivative analysis, comparing the second steel rail potential gradient feature quantity with an optimal detection threshold value, identifying potential steel rail insulation defect segments, carrying out time sequence frequency statistics on position information of the potential steel rail insulation defect segments, and determining the final high-confidence steel rail insulation degradation position. And the automatic positioning and intelligent operation and maintenance of the insulation degradation area of the subway steel rail reflux type direct current traction power supply system are realized.

Inventors

  • WANG AIMIN
  • YANG XIN

Assignees

  • 西华大学

Dates

Publication Date
20260505
Application Date
20260116

Claims (10)

  1. 1. The method for positioning the insulation degradation of the subway steel rail is characterized by comprising the following steps of: Establishing a simulation model based on the topological structure of the target subway line, and setting various simulated insulation defect working conditions to construct a training data set; constructing a first steel rail potential distribution model based on the training data set, extracting first steel rail potential gradient feature quantity reflecting an insulation state, and determining an optimal steel rail insulation defect detection threshold value through an optimization algorithm; Acquiring field potential data of the target subway line, constructing a second steel rail potential distribution model according to the field potential data, extracting second steel rail potential gradient feature quantity through multi-resolution spatial derivative analysis, comparing the second steel rail potential gradient feature quantity with the optimal detection threshold value, and identifying potential steel rail insulation defect sections; And carrying out time sequence frequency statistics on the position information of the potential steel rail insulation defect section, and outputting a high-confidence steel rail insulation degradation position under the condition that the statistical result meets the preset judgment condition.
  2. 2. The method of claim 1, wherein the setting a plurality of simulated insulation defect conditions to construct a training data set comprises: Determining the positions of a transformer substation and a station based on the topological structure of the target subway line; Arranging monitoring points based on the positions of the transformer substation and the station, and generating an initial monitoring point distribution scheme; And setting a plurality of insulation defect working conditions according to the initial monitoring point distribution scheme in the simulation model, and collecting potential simulation data and corresponding real defect position information of each monitoring point to form the training data set.
  3. 3. The method of claim 1, wherein constructing a first rail potential distribution model based on the training dataset and extracting a first rail potential gradient feature quantity reflecting an insulation state, determining an optimal rail insulation defect detection threshold value by an optimization algorithm, comprises: Constructing a first rail potential distribution function with continuous whole lines through a cubic spline interpolation algorithm based on the training data set; adopting a multi-resolution analysis method, and respectively calculating a first-order spatial derivative and a second-order spatial derivative of the first rail potential distribution function under at least two different scales; Carrying out weighted fusion on the first-order spatial derivative and the second-order spatial derivative under multiple scales to generate comprehensive first rail potential gradient characteristic quantity; obtaining a degradation diagnosis position by comparing the first rail potential gradient feature quantity with a detection threshold value, wherein the detection threshold value has the expression: , And The global mean and standard deviation of the feature quantity in the training data set are respectively obtained by a particle swarm optimization algorithm with the minimum position deviation of diagnosis and real degradation as a target Thereby obtaining the steel rail insulation defect detection threshold value 。
  4. 4. The method of claim 1, wherein the acquiring the field potential data of the target subway line, and constructing a second rail potential distribution model from the field potential data, comprises: Acquiring the site potential data based on monitoring points of an actual line deployed on the target subway line; And generating a full-line continuous rail potential distribution function by using a spline interpolation algorithm which is the same as the construction process of the first rail potential distribution model based on the field rail potential data, so as to obtain the second rail potential distribution model.
  5. 5. The method of claim 4, wherein extracting a second rail potential gradient feature by multi-resolution spatial derivative analysis, comparing the second rail potential gradient feature to the optimal detection threshold, identifying potential rail insulation defect segments, comprises: Adopting a multi-resolution analysis method which is the same as the generation process of the potential gradient characteristic quantity of the first steel rail, Calculating the spatial gradient change rate of the rail potential distribution function, and generating a second rail potential gradient characteristic quantity; And comparing the second steel rail potential gradient characteristic quantity with the optimal detection threshold value, and identifying all continuous sections of which the second steel rail potential gradient characteristic quantity exceeds the optimal detection threshold value as potential steel rail insulation defect sections.
  6. 6. The method according to claim 1, wherein the performing time sequence frequency statistics on the position information of the potential steel rail insulation defect segment, and outputting the high confidence steel rail insulation degradation position when the statistics result meets the preset judgment condition, includes: in a preset continuous time window, carrying out time sequence frequency statistics on the position information of the potential steel rail insulation defect section; Screening out the space sections which are repeatedly identified as the potential steel rail insulation defect sections in the current period and are more than or equal to the preset period times when the output period of each continuous time window is finished, marking the space sections as the high-confidence steel rail insulation degradation positions and outputting the space sections; And for the space section smaller than the preset period times, outputting the space section as interference or sporadic signal filtering processing.
  7. 7. The method according to claim 6, wherein the step of screening out, at the end of the output period of each of the continuous time windows, the spatial section repeatedly identified as the potential rail insulation defect section number of times equal to or greater than a preset period number of times in the current period, marked as the high-confidence rail insulation degradation position and outputting includes: Setting a diagnosis period and an aggregation output period, wherein the diagnosis period is a diagnosis interval of each time, and the aggregation output period is an output diagnosis report interval corresponding to the output period of the continuous time window; Performing one-time insulation state diagnosis at the corresponding end time point of each diagnosis period, and generating a potential defect area list containing all potential steel rail insulation defect sections in the diagnosis period, wherein each record contains a space position, diagnosis time and confidence; Counting the frequency of each space position interval aiming at the results of all insulation state diagnosis in the current aggregation output period at the corresponding ending time point of each aggregation output period, and marking the interval as the high-confidence steel rail insulation degradation position if the frequency of the interval identified as the potential defect in the current aggregation output period is greater than or equal to the preset period frequency; and outputting all the high-confidence steel rail insulation degradation positions in the current aggregation output period to an operation and maintenance platform, emptying the current aggregation output period statistical buffer memory, and entering the next aggregation output period.
  8. 8. The utility model provides a subway rail insulation degradation positioner which characterized in that includes: the building module is configured to build a simulation model based on the topological structure of the target subway line, and set various simulated insulation defect working conditions to build a training data set; The determining module is configured to construct a first steel rail potential distribution model based on the training data set, extract first steel rail potential gradient feature quantity reflecting the insulation state and determine an optimal steel rail insulation defect detection threshold value through an optimization algorithm; the identification module is configured to acquire field potential data of the target subway line, construct a second steel rail potential distribution model according to the field potential data, extract second steel rail potential gradient feature quantity through multi-resolution spatial derivative analysis, compare the second steel rail potential gradient feature quantity with the optimal detection threshold value and identify potential steel rail insulation defect sections; And the statistics module is configured to perform time sequence frequency statistics on the position information of the potential steel rail insulation defect section, and output a high-confidence steel rail insulation degradation position under the condition that the statistics result meets a preset judgment condition.
  9. 9. A computing device, comprising: A memory and a processor; The memory is used for storing computer executable instructions, and the processor is used for executing the computer executable instructions to realize the steps of the subway rail insulation degradation positioning method according to any one of claims 1 to 7.
  10. 10. A computer-readable storage medium storing computer instructions which, when executed by a processor, implement the steps of the subway rail insulation degradation positioning method of any one of claims 1 to 7.

Description

Subway steel rail insulation degradation positioning method, device, computing equipment and storage medium Technical Field The application relates to the technical field of rail transit power supply safety, in particular to a subway rail insulation degradation positioning method. The application also relates to a subway steel rail insulation degradation positioning device, a computing device and a computer readable storage medium. Background In a direct current traction power supply system, a steel rail is used as a traction current return path, and the degradation of the ground insulation performance is a main source for causing serious consequences such as stray current leakage, electrochemical corrosion of a buried metal structure, equipment interference and the like. Insulation degradation is typically caused by environmental factors such as moisture, dust, tunnel penetration, and material aging, joint corrosion, etc., the evolution of which manifests itself as a local or regional drop in ground resistance, thereby creating an abnormal potential difference between the rail and its grounding device, such abnormal potential distribution having spatial non-uniformity and dimensional sensitivity. The current detection means for the problems mainly focus on manual inspection, single-point potential measurement and equivalent resistance test, and the problems of limited detection range, incapability of supporting continuous monitoring, low positioning accuracy, lack of quantitative evaluation mechanism and the like generally exist. Although there have been studies attempting to invert the insulation defect location by building a resistive network model or an electromagnetic field simulation model, the model is often based on uniform parameters, ignoring line topology changes and the dynamics of current carrying distribution, resulting in poor matching of simulation results to field data. In addition, the existing method relies on experience threshold values for defect judgment, lacks a self-adaptive optimization mechanism, is difficult to adapt to different line structures and operation conditions, and is easy to report by mistake or miss. More importantly, the current technology does not form a complete closed loop positioning system from data acquisition to feature modeling and threshold optimization to regional output, and particularly, the identification capability is limited because local potential mutation caused by insulation degradation and global potential fluctuation caused by load change cannot be effectively distinguished on the feature level. Therefore, development of a novel positioning method integrating line topology modeling, multi-scale spatial feature analysis, self-adaptive threshold optimization and physical post-processing mechanism is needed to realize high-precision and automatic identification of an insulation degradation region and engineering landing. Disclosure of Invention In view of the above, the embodiment of the application provides a method for positioning insulation degradation of a subway steel rail, which aims to solve the technical defects in the prior art. The embodiment of the application also provides a subway steel rail insulation degradation positioning device, a computing device and a computer readable storage medium. According to a first aspect of an embodiment of the present application, there is provided a method for positioning insulation degradation of a subway rail, including: Establishing a simulation model based on the topological structure of the target subway line, and setting various simulated insulation defect working conditions to construct a training data set; constructing a first steel rail potential distribution model based on the training data set, extracting first steel rail potential gradient feature quantity reflecting an insulation state, and determining an optimal steel rail insulation defect detection threshold value through an optimization algorithm; Acquiring field potential data of the target subway line, constructing a second steel rail potential distribution model according to the field potential data, extracting second steel rail potential gradient feature quantity through multi-resolution spatial derivative analysis, comparing the second steel rail potential gradient feature quantity with the optimal detection threshold value, and identifying potential steel rail insulation defect sections; And carrying out time sequence frequency statistics on the position information of the potential steel rail insulation defect section, and outputting a high-confidence steel rail insulation degradation position under the condition that the statistical result meets the preset judgment condition. Optionally, the setting a plurality of simulated insulation defect conditions to construct a training data set includes: Determining the positions of a transformer substation and a station based on the topological structure of the target subway line; Arrangin