A subway fault data analysis method based on a cloud platform

Abstract

The application relates to the field of subway fault analysis, in particular to a subway fault data analysis method based on a cloud platform, which comprises the following steps: acquiring train operation difference parameters and noise characteristic difference parameters of each key monitoring point to determine a monitoring difference state; determining a difference analysis strategy of each key monitoring point according to the monitoring difference state; when performing abnormal defect analysis, determining a defect analysis mode according to a noise cycle stability coefficient and an image gradient difference coefficient of the key monitoring point; under the condition that the abnormal defect analysis is completed, determining a monitoring compensation mode according to a defect interference radiation coefficient of the key monitoring point; when performing parameter optimization analysis, determining a setting mode of a point abnormality coefficient according to a noise abnormality frequency and a running interference proportion of the key monitoring point, and determining whether the point abnormality coefficient needs to be adjusted according to coincident abnormal parameters; and the application improves the accuracy of an operation parameter optimization result.

Description

Technical Field

[0001] This invention relates to the field of subway fault analysis, and in particular to a subway fault data analysis method based on a cloud platform. Background Technology

[0002] Noise and vibration during subway operation are an important part of subway operation fault analysis. During the continuous use of subway tracks, unknown track defects are easily generated, which can cause abnormal noise and affect the reliability of subway track noise analysis results. Therefore, how to compensate for track noise monitoring results based on the analysis results of actual noise signals and track surface defects to ensure the reliability of the adjustment results of subway operation parameters used to optimize noise conditions is a problem that urgently needs to be solved by those skilled in the art.

[0003] Chinese Patent Application Publication No. CN116011326A discloses a train operation optimization method, device, electronic equipment, and storage medium, relating to the field of rail transit technology. The method includes: dividing the entire train route into multiple sub-sections; solving a train operation optimization model based on the train's operation in each sub-section; and optimizing the train's operation curve based on the solution results. The train operation optimization model includes an objective function and constraints, with the objective function constructed to minimize the weighted sum of the train's total operating energy consumption and total operating noise. However, the above solution has the following problems: it fails to effectively compensate for the track noise acquisition results based on the analysis results of the actual noise signals and track surface defects, resulting in low reliability of the noise and vibration monitoring results, and consequently, an inability to timely and effectively improve track noise during subway operation. Summary of the Invention

[0004] To address this, the present invention provides a cloud-based subway fault data analysis method to overcome the problem in the prior art that fails to effectively compensate for the track noise acquisition results based on the analysis results of the actual noise signals and track surface defects, resulting in low reliability of the noise and vibration monitoring results and thus the inability to improve track noise during subway operation in a timely and effective manner.

[0005] To achieve the above objectives, this invention provides a subway fault data analysis method based on a cloud platform, comprising:

[0006] Obtain train operation difference parameters and noise characteristic difference parameters at each key monitoring point to determine the monitoring difference status;

[0007] Based on the monitoring difference status, a difference analysis strategy is determined for each key monitoring point. The difference analysis strategy is to conduct abnormal defect analysis or parameter optimization analysis for the key monitoring points.

[0008] In the abnormal defect analysis, the defect analysis method is determined based on the noise period stability coefficient and image gradient difference coefficient of the key monitoring points. The defect analysis method is to perform grayscale analysis of the track surface reference direction for the track surface area image, or to perform threshold extraction analysis based on the proportion of abnormal features.

[0009] Under the condition that the abnormal defect analysis is completed, the monitoring compensation method is determined according to the defect interference radiation coefficient of the key monitoring points. The monitoring compensation method is to adjust the abnormal coefficient of the point according to the regional distribution distance and regional density coefficient, or to adjust the abnormal coefficient of the point according to the defect interference radiation coefficient.

[0010] In the parameter optimization analysis, the setting method of the point anomaly coefficient is determined based on the frequency of noise anomalies and the proportion of operational interference at key monitoring points, and whether to adjust the point anomaly coefficient is determined based on the overlapping anomaly parameters.

[0011] Furthermore, when a key monitoring point is in a first preset difference state where the noise characteristic difference parameter is less than or equal to the preset noise characteristic difference parameter, the difference analysis strategy is to perform an abnormal defect analysis on that key monitoring point.

[0012] Furthermore, when a key monitoring point is in a second preset difference state where both the train operation difference parameter and the noise characteristic difference parameter are greater than the preset noise characteristic difference parameter, the difference analysis strategy is to perform parameter optimization analysis on that key monitoring point.

[0013] Furthermore, when performing abnormal defect analysis on a key monitoring point, the defect analysis method is determined based on the noise period stability coefficient and image gradient difference coefficient of the key monitoring point.

[0014] If the noise period stability coefficient of the key monitoring point is greater than the preset noise stability coefficient and the image gradient difference coefficient is less than or equal to the preset image gradient difference coefficient, then perform grayscale analysis of the track surface reference direction for the track surface area image.

[0015] If the noise period stability coefficient of the key monitoring point is less than or equal to the preset noise stability coefficient or the image gradient difference coefficient is greater than the preset image gradient difference coefficient, then the extraction threshold analysis is performed based on the proportion of abnormal features.

[0016] Furthermore, when performing grayscale analysis of the track surface reference direction for the track surface area image, the track surface area image corresponding to the key monitoring point is segmented to obtain several track surface analysis areas with the same area. The defect anomaly coefficient of each track surface analysis area is determined based on the lateral grayscale difference parameter and the extended grayscale difference parameter. The defect interference degree of the key monitoring point is determined based on the proportion of the key area and the reference anomaly coefficient.

[0017] The track surface analysis area where the defect anomaly coefficient is greater than the preset defect anomaly coefficient is designated as the key monitoring area.

[0018] Furthermore, when performing threshold extraction analysis based on the proportion of abnormal features, abnormal features are extracted for the track surface area image corresponding to the key monitoring points. Track surface reference features are determined based on the proportion of abnormal features in each category. Abnormal features in categories with a proportion of abnormal features less than the preset proportion of abnormal features are all recorded as track surface reference features.

[0019] The region extraction threshold is determined based on the proportion of abnormal features in the track surface reference features and the feature threshold parameter. The key monitoring areas of the key monitoring points are determined based on the region extraction threshold, and the defect interference degree of the key monitoring points is determined based on the proportion of the key areas.

[0020] The degree of defect interference is positively correlated with the proportion of key areas.

[0021] Furthermore, after the abnormal defect analysis is completed, the monitoring compensation method is determined based on the defect interference radiation coefficient of the key monitoring points.

[0022] If the defect interference radiation coefficient of a key monitoring point is greater than the preset defect interference radiation coefficient, the point anomaly coefficient will be reduced and adjusted according to the regional distribution distance and regional density coefficient, and a track operation warning will be issued.

[0023] If the defect interference radiation coefficient of a key monitoring point is less than or equal to the preset defect interference radiation coefficient, the point anomaly coefficient is reduced according to the defect interference radiation coefficient. The reduction value of the point anomaly coefficient is positively correlated with the defect interference radiation coefficient.

[0024] The defect interference radiation coefficient is determined based on the defect interference degree and the defect distribution degree. The condition for completing the abnormal defect analysis is that the defect interference degree of the key monitoring points is determined.

[0025] Furthermore, if the defect interference radiation coefficient of the key monitoring point is greater than the preset defect interference radiation coefficient, the key monitoring area of ​​the key monitoring point is divided into dense areas. The regional interference coefficient is determined according to the regional distribution distance and regional density coefficient of each dense area, and the point anomaly coefficient is reduced and adjusted according to the regional interference coefficient.

[0026] The decrease in the point anomaly coefficient is positively correlated with the regional interference coefficient;

[0027] The regional interference coefficient is negatively correlated with the regional distribution distance, and positively correlated with the regional density coefficient.

[0028] Furthermore, when performing parameter optimization analysis for a key monitoring point, the setting method of the point anomaly coefficient is determined based on the frequency of noise anomalies and the proportion of operational interference at the key monitoring point.

[0029] If the frequency of noise anomalies at key monitoring points is greater than the preset frequency of noise anomalies and the proportion of operational interference is greater than the preset proportion of operational interference, the point anomaly coefficient is determined based on the frequency of noise anomalies and the reference noise parameters. The point anomaly coefficient is positively correlated with the frequency of noise anomalies and the reference noise parameters, respectively.

[0030] If the frequency of noise anomalies at key monitoring points is less than or equal to the preset frequency of noise anomalies or the proportion of operational interference is less than or equal to the preset proportion of operational interference, the point anomaly coefficient is determined based on the reference noise parameters. The point anomaly coefficient is positively correlated with the reference noise parameters.

[0031] Furthermore, when determining the location anomaly coefficient based on the frequency of noise anomalies and reference noise parameters, the decision on whether to adjust the location anomaly coefficient is based on the coincident anomaly parameters.

[0032] If the coincidence anomaly parameter is greater than the preset coincidence anomaly parameter, the point anomaly coefficient is increased and adjusted according to the coincidence anomaly parameter and the proportion of operational interference.

[0033] The increase in the point anomaly coefficient is positively correlated with the coincidence anomaly parameter and the proportion of operational interference.

[0034] Compared with the prior art, the beneficial effects of the present invention are as follows: the technical solution of the present invention determines the monitoring difference status of each key monitoring point based on the train operation difference parameters and noise characteristic difference parameters, and analyzes the differences in the operation parameters and noise characteristics of the subway when it passes through points with abnormal noise by analyzing the differences in the monitoring difference status of each key monitoring point. This allows for a preliminary classification of the causes of abnormal noise, making the subsequent determined difference analysis strategy more consistent with the actual situation, avoiding the low analysis efficiency of the optimization parameter compensation process, and improving the accuracy of the optimization results of subway operation parameters.

[0035] Furthermore, in this invention, the monitoring difference status of each key monitoring point is determined based on the train operation difference parameters and noise characteristic difference parameters. By analyzing whether there is a regularity in the noise obtained from the key monitoring points and the differences in train operation, the degree of interference from subway operation parameters in the abnormal noise of the key analysis points is judged. For cases where the interference level of subway operation parameters is low, the track surface image is further analyzed to obtain the track defect status, and compensation is made for the optimization process of subway operation parameters accordingly. This invention improves the accuracy of the operation parameter optimization results.

[0036] Furthermore, in this invention, the defect analysis method is determined based on the noise period stability coefficient and image gradient difference coefficient of the key monitoring points. By analyzing the noise period stability coefficient and image gradient difference coefficient, the track surface condition and main defect status within the range of the key monitoring points are determined, thereby determining a targeted defect analysis method for the track surface image. This makes the defect analysis method more consistent with the actual track surface condition, thereby improving the accuracy and efficiency of defect identification results.

[0037] Furthermore, in this invention, when performing parameter optimization analysis on a key monitoring point, the setting method of the point anomaly coefficient is determined based on the frequency of abnormal noise and the proportion of operational interference at the key monitoring point. That is, by analyzing the frequency of abnormal noise occurrence and its overlap with abnormal operating parameters, the influence of subway operating parameters on the occurrence of abnormal noise is further determined, ensuring that the setting method of the point anomaly coefficient is more in line with the actual scenario, and the point anomaly coefficient is adjusted accordingly. This invention improves the accuracy of the optimization results for subway operating parameters. Attached Figure Description

[0038] Figure 1 This is a schematic diagram of the subway fault data analysis method based on a cloud platform according to the present invention;

[0039] Figure 2 This is a flowchart illustrating the difference analysis strategy for determining the differences of each key monitoring point based on the monitoring difference status of the present invention.

[0040] Figure 3 This is a flowchart illustrating the method for determining defect analysis based on the noise period stability coefficient and image gradient difference coefficient of key monitoring points in this invention.

[0041] Figure 4 This is a flowchart illustrating how the monitoring compensation method is determined based on the defect interference radiation coefficient of key monitoring points, according to the present invention. Detailed Implementation

[0042] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0043] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0044] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0045] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0046] Please see Figures 1 to 4 As shown, this embodiment of the invention provides a subway fault data analysis method based on a cloud platform, including:

[0047] Obtain train operation difference parameters and noise characteristic difference parameters at each key monitoring point to determine the monitoring difference status;

[0048] Based on the monitoring difference status, a difference analysis strategy is determined for each key monitoring point. The difference analysis strategy is to conduct abnormal defect analysis or parameter optimization analysis for the key monitoring points.

[0049] In the abnormal defect analysis, the defect analysis method is determined based on the noise period stability coefficient and image gradient difference coefficient of the key monitoring points. The defect analysis method is to perform grayscale analysis of the track surface reference direction for the track surface area image, or to perform threshold extraction analysis based on the proportion of abnormal features.

[0050] Under the condition that the abnormal defect analysis is completed, the monitoring compensation method is determined according to the defect interference radiation coefficient of the key monitoring points. The monitoring compensation method is to adjust the abnormal coefficient of the point according to the regional distribution distance and regional density coefficient, or to adjust the abnormal coefficient of the point according to the defect interference radiation coefficient.

[0051] In the parameter optimization analysis, the setting method of the point anomaly coefficient is determined based on the frequency of noise anomalies and the proportion of operational interference at key monitoring points, and whether to adjust the point anomaly coefficient is determined based on the overlapping anomaly parameters.

[0052] This invention is used to compensate for the adjustment results of subway train operating parameters during noise monitoring in subway operation, so as to ensure the effectiveness of the optimization results of operating parameters. The operating parameters of the subway train in this invention include, but are not limited to, speed, acceleration, and shifting strategy. In this invention, several noise monitoring points are set for the track area where noise monitoring is required. Each noise monitoring point is equipped with a noise sensor to detect the noise level of the corresponding noise monitoring point, and a noise analysis device is set to perform spectrum analysis on the collected noise signal. How to set the noise monitoring points and how to set the noise sensor and noise analysis device for each noise monitoring point are contents that are easy to understand for those skilled in the art, and will not be described in detail here. This invention does not specify the specific model of the noise sensor and noise analysis device used. Users can make adaptive settings according to the actual working scenario.

[0053] The noise reference values ​​at each noise monitoring point are periodically detected to determine key monitoring points. In this invention, a cyclic noise monitoring cycle is used, and the duration of the noise monitoring cycle can be determined by the user. One noise monitoring cycle duration is provided, which is 12 hours. At the end of each noise monitoring cycle, the noise reference values ​​at each noise monitoring point are detected. For a single noise monitoring point, the noise reference value is the average of the noise amplitude values ​​collected by each train passing through the noise point within a noise monitoring cycle.

[0054] This invention utilizes several subway operation monitoring records. Each subway operation monitoring record records at least one instance of noise monitoring results analysis and compensation for a noise monitoring point, including the point noise reference value, train operation difference parameters, noise characteristic difference parameters, gradient reference difference value, noise period stability coefficient, image gradient difference coefficient, defect anomaly coefficient, anomaly feature ratio, number of key monitoring areas within the distribution analysis range of dense key areas, defect interference radiation coefficient, noise anomaly frequency, operation interference ratio, and overlapping anomaly parameters. Each subway operation monitoring record also has a corresponding qualification mark, which records whether the effectiveness of the operation parameter optimization results meets the user's needs. It can be understood that the user can determine whether the effectiveness of the operation parameter optimization results meets the requirements based on self-defined indicators. For example, the self-defined indicator can be, but is not limited to, the degree of noise optimization. The degree of noise optimization is the decrease in the point noise reference value of the noise monitoring point between two noise monitoring cycles before and after the completion of operation parameter optimization.

[0055] Specifically, when a key monitoring point is in the first preset difference state where the noise characteristic difference parameter is less than or equal to the preset noise characteristic difference parameter, the difference analysis strategy is to perform an abnormal defect analysis on the key monitoring point.

[0056] The monitoring difference status includes a first preset difference status and a second preset difference status. If the train operation difference parameter of the key monitoring point is less than or equal to the preset train operation difference parameter and the noise characteristic difference parameter is greater than the preset noise characteristic difference parameter, then no further analysis will be performed for the key monitoring point.

[0057] Specifically, when a key monitoring point is in the second preset difference state where both the train operation difference parameter and the noise characteristic difference parameter are greater than the preset noise characteristic difference parameter, the difference analysis strategy is to perform parameter optimization analysis on that key monitoring point.

[0058] The key monitoring points are noise monitoring points whose noise reference values ​​are greater than preset noise reference values. The preset noise reference values ​​can be determined by the user based on the actual working scenario. For example, the user can set them based on the subway operation monitoring records. The higher the user's requirements for the effectiveness of the operation parameter optimization results, the smaller the preset noise reference values. A method for determining the preset noise reference values ​​is provided, which records the minimum noise reference value of the key monitoring points in the subway operation monitoring records that meets the user's requirements for the effectiveness of the operation parameter optimization results as the preset noise reference value.

[0059] The noise characteristic difference parameter is the sum of the products of the noise amplitude difference value and the noise frequency band difference value, respectively, and their corresponding difference influence coefficients. For a single key monitoring point, it represents the noise amplitude difference value of that key monitoring point in the current noise monitoring cycle. n represents the number of times a train passes through the key monitoring point within the current noise monitoring period; fi represents the noise amplitude value collected when the train passes through the key monitoring point for the i-th time within the current noise monitoring period; f0 represents the average noise amplitude value collected when the train passes through the key monitoring point each time within the current noise monitoring period; the noise amplitude value is the maximum noise amplitude value when the train passes through the key monitoring point; the noise frequency band difference value of the key monitoring point in the current noise monitoring period = period coverage frequency band range - overlapping frequency band range; the period coverage frequency band range is the absolute value of the difference between the maximum and minimum frequencies of the noise signal collected when the train passes through the key monitoring point each time within the current noise monitoring period; the overlapping frequency band range is the range of frequencies collected when the train passes through the key monitoring point each time within the current noise monitoring period. The sum of the absolute values ​​of the differences between the maximum and minimum frequencies of the overlapping frequency bands at key monitoring points. If the noise signals collected when the train passes through the key monitoring point each time during the current noise monitoring cycle have the same noise signal frequency band, then the frequency range of the noise signal is recorded as an overlapping frequency band. The noise signal frequency band is the frequency range of the noise signal. The values ​​of the noise amplitude difference value and the difference influence coefficient corresponding to the noise frequency band difference value can be determined by the user according to the actual working scenario. This embodiment provides a value of the noise amplitude difference value and the difference influence coefficient corresponding to the noise frequency band difference value. The difference influence coefficient corresponding to the noise amplitude difference value is 0.5, and the difference influence coefficient corresponding to the noise frequency band difference value is 0.5.

[0060] The train operation difference parameter is the sum of the equipment speed difference degree and the equipment acceleration difference degree, wherein the equipment speed difference degree... Vi is the train speed when it passes the key monitoring point for the i-th time in the current noise monitoring cycle, V0 is the average speed of the train when it passes the key monitoring point for each time in the current noise monitoring cycle, and the equipment acceleration difference is the absolute value of the difference between the maximum acceleration value and the minimum acceleration value of the train when it passes the key monitoring point for each time in the current noise monitoring cycle.

[0061] The values ​​of the preset train operation difference parameters and preset noise characteristic difference parameters can be determined by the user according to the actual working scenario. For example, the user can set them based on the subway operation monitoring records. A method for determining the values ​​of the preset train operation difference parameters is provided, in which the subway operation monitoring records for parameter optimization analysis of key monitoring points are recorded as state reference records, and the minimum value of the train operation difference parameters in the state reference records that meet the user's requirements for the effectiveness of the operation parameter optimization results is recorded as the preset train operation difference parameter. A method for determining the values ​​of the preset noise characteristic difference parameters is also provided, in which the minimum value of the noise characteristic difference parameters in the state reference records that meet the user's requirements for the effectiveness of the operation parameter optimization results is recorded as the preset noise characteristic difference parameter.

[0062] Specifically, when performing abnormal defect analysis on a key monitoring point, the defect analysis method is determined based on the noise period stability coefficient and image gradient difference coefficient of the key monitoring point.

[0063] If the noise period stability coefficient of the key monitoring point is greater than the preset noise stability coefficient and the image gradient difference coefficient is less than or equal to the preset image gradient difference coefficient, then perform grayscale analysis of the track surface reference direction for the track surface area image.

[0064] If the noise period stability coefficient of the key monitoring point is less than or equal to the preset noise stability coefficient or the image gradient difference coefficient is greater than the preset image gradient difference coefficient, then the extraction threshold analysis is performed based on the proportion of abnormal features.

[0065] In this invention, each noise monitoring point is simultaneously equipped with an image acquisition device to acquire track surface images within a preset monitoring range of each noise monitoring point, denoted as track surface area images. The preset monitoring range is determined based on the maximum sound source distance of each noise monitoring point. For a single noise monitoring point, the midpoint of its preset monitoring range is the location of the noise monitoring point, and the distance between the two ends of the preset monitoring range and the location of the noise monitoring point is the maximum sound source distance of the noise monitoring point. The maximum sound source distance is the maximum value of the distance between the sound source location of the noise signal that can be acquired by the noise sensor and noise analysis device installed at the noise monitoring point and the location of the noise monitoring point.

[0066] When performing anomaly defect analysis on a key monitoring point, an image of the track surface region within a preset detection range of the key monitoring point is acquired. The track surface region image corresponding to the key monitoring point is segmented to obtain several track surface analysis regions of equal area. An image gradient difference coefficient is detected for the track surface region image of the key monitoring point. The image gradient difference coefficient represents the number of gradient difference regions contained in the track surface region image of the key monitoring point. Gradient difference regions are track surface analysis regions whose gradient reference difference values ​​are greater than a preset gradient reference difference value. Gradient analysis is performed on each track surface analysis region to obtain the gradient reference value for each region. For a single track surface analysis region, the gradient reference difference value is the absolute value of the difference between the gradient reference value of that region and the average value of the gradient reference values ​​of all track surface analysis regions. The gradient reference value is the gradient amplitude of the image corresponding to the track surface analysis region. The preset gradient reference difference value can be determined by the user according to the actual working scenario. For example, users can set the preset gradient reference difference value based on subway operation monitoring records. The higher the user's requirement for the effectiveness of the operation parameter optimization results, the smaller the preset gradient reference difference value should be. A method for setting the preset gradient reference difference value is provided, in which the minimum value of the gradient reference difference value of each gradient difference region in the subway operation monitoring records that meets the user's requirement for the effectiveness of the operation parameter optimization results is recorded as the preset gradient reference difference value. The noise period stability coefficient is the absolute value of the difference between the maximum and minimum values ​​of the fundamental frequency of the noise signal obtained in the current noise monitoring period. How to perform gradient analysis to determine the gradient reference value for each track surface analysis area and how to obtain the fundamental frequency of the noise signal are contents that are easy to understand for those skilled in the art, and will not be elaborated here. Based on the noise period stability coefficient and the image gradient difference coefficient, the shadows and rust in the obtained track surface image and the possible defect states can be characterized, and the defect extraction method can be selected in a targeted manner to improve the reliability of defect extraction.

[0067] The values ​​of the preset noise stability coefficient and the preset image gradient difference coefficient can be determined by the user according to the actual working scenario. For example, the user can set them based on the subway operation monitoring records. A method for determining the value of the preset noise stability coefficient is provided, in which the subway operation monitoring records for performing grayscale analysis of the track surface area image in the track surface reference direction are recorded as the analysis reference records, and the minimum value of the noise period stability coefficient in the analysis reference records that meets the user's requirements for the effectiveness of the operation parameter optimization results is recorded as the preset noise stability coefficient. A method for determining the value of the preset image gradient difference coefficient is provided, in which the maximum value of the image gradient difference coefficient in the analysis reference records that meets the user's requirements for the effectiveness of the operation parameter optimization results is recorded as the preset image gradient difference coefficient.

[0068] Specifically, when performing grayscale analysis of the track surface reference direction for the track surface area image, the track surface area image corresponding to the key monitoring point is segmented to obtain several track surface analysis areas of the same area. The defect anomaly coefficient of each track surface analysis area is determined based on the lateral grayscale difference parameter and the extended grayscale difference parameter. The defect interference degree of the key monitoring point is determined based on the proportion of the key area and the reference anomaly coefficient.

[0069] The track surface analysis area where the defect anomaly coefficient is greater than the preset defect anomaly coefficient is designated as the key monitoring area.

[0070] Specifically, image grayscale value detection is performed for each track surface analysis area. For a single track surface analysis area, the lateral grayscale difference parameter is the maximum absolute value of the difference between the image grayscale values ​​of the track analysis areas included in the preset lateral analysis area of ​​the track surface analysis area. The track analysis areas included in the preset lateral analysis area are all laterally adjacent, and the common edge of two laterally adjacent track analysis areas is parallel to the extension direction of the track. The extension direction grayscale difference parameter is the maximum absolute value of the difference between the image grayscale values ​​of the track analysis areas included in the preset extension direction analysis area of ​​the track surface analysis area. The track analysis areas included in the preset extension direction analysis area are all extension direction adjacent, and the common edge of two extension direction adjacent track analysis areas is perpendicular to the extension direction of the track. The number of preset lateral analysis areas and the number of track analysis areas included in the preset extension direction analysis area can be determined by the user according to the actual working scenario. One example is that the number of track analysis areas included in the preset lateral analysis area is 5, and another example is that the number of track analysis areas included in the preset extension direction analysis area is 8.

[0071] The defect anomaly coefficient is the sum of the products of the horizontal grayscale difference parameter and the extended grayscale difference parameter and their corresponding anomaly reference coefficients, respectively. The values ​​of the anomaly reference coefficients corresponding to the horizontal grayscale difference parameter and the extended grayscale difference parameter can be set by the user according to the actual working scenario. One possible value for the anomaly reference coefficients corresponding to the horizontal grayscale difference parameter is 0.7, and the value for the anomaly reference coefficients corresponding to the extended grayscale difference parameter is 0.3.

[0072] The value of the preset defect anomaly coefficient can be determined by the user according to the actual working scenario. For example, the user can set it according to the subway operation monitoring record. The higher the user’s requirements for the effectiveness of the operation parameter optimization results, the larger the value of the preset defect anomaly coefficient. A method for determining the value of the preset defect anomaly coefficient is provided, which is the average value of the defect anomaly coefficient of each key monitoring area in the analysis reference record that meets the user’s requirements for the effectiveness of the operation parameter optimization results.

[0073] For a single key monitoring point, the defect interference degree is the product of the key area proportion and the reference anomaly coefficient. The key area proportion = the number of key monitoring areas of the key monitoring point / the number of track surface analysis areas of the key monitoring point. The reference anomaly coefficient is the average value of the defect anomaly coefficients of the key monitoring areas of the key monitoring point.

[0074] Specifically, when performing threshold analysis based on the proportion of abnormal features, abnormal features are extracted from the track surface area image corresponding to the key monitoring points. Track surface reference features are determined based on the proportion of abnormal features in each category. Abnormal features in categories with a proportion of abnormal features less than the preset proportion of abnormal features are all recorded as track surface reference features.

[0075] The region extraction threshold is determined based on the proportion of abnormal features in the track surface reference features and the feature threshold parameter. The key monitoring areas of the key monitoring points are determined based on the region extraction threshold, and the defect interference degree of the key monitoring points is determined based on the proportion of the key areas.

[0076] The degree of defect interference is positively correlated with the proportion of key areas.

[0077] This process involves extracting abnormal features from track surface images at key monitoring points and classifying abnormal features into categories based on their feature similarity. The feature similarity between abnormal features of the same category is less than a preset feature similarity. The value of this preset feature similarity can be determined by the user based on the actual work scenario. For example, the user can set it based on subway operation monitoring records. The higher the user's requirement for the effectiveness of the operation parameter optimization results, the larger the preset feature similarity value. A method for determining the preset feature similarity value is provided, where the average feature similarity between different categories of abnormal features in the subway operation monitoring records that meet the user's requirements for the effectiveness of the operation parameter optimization results is recorded as the preset feature similarity. How to extract abnormal features and how to determine the feature similarity between abnormal features are easily understood by those skilled in the art and will not be elaborated here.

[0078] For anomalies of a single category, the percentage of anomalies in that category is calculated as: (Number of anomalies in that category) / (Number of anomalies extracted from the track surface region image). The user can determine the value of this preset percentage based on the actual work scenario; for example, it can be set based on subway operation monitoring records. A method for determining the preset percentage is provided where the maximum percentage of anomalies for each track surface reference feature in the subway operation monitoring records that meets the user's requirements for the effectiveness of the operation parameter optimization results is recorded as the preset percentage. Feature threshold parameters are detected for each track surface reference feature. For a single track surface reference feature, the feature threshold parameter is the defect image obtained after image segmentation. The upper limit of the grayscale range used to determine when the sum of the entropies of the image and background parts is maximized is defined as follows: the region extraction threshold is the sum of the products of the feature threshold parameters of each track surface reference feature and their corresponding threshold reference coefficients. The threshold reference coefficients corresponding to the feature threshold parameters of each track surface reference feature are negatively correlated with the proportion of abnormal features in their respective categories. The region extraction threshold is used as the lower limit of the grayscale range used for defect image segmentation of the track surface region image. The track surface analysis area containing defective images is designated as the key monitoring area. How to detect the feature threshold parameters and how to perform defect image segmentation of the track surface region image based on the region extraction threshold are content that is easily understood by those skilled in the art and will not be elaborated here.

[0079] Specifically, once the abnormal defect analysis is completed, the monitoring compensation method is determined based on the defect interference radiation coefficient of the key monitoring points.

[0080] If the defect interference radiation coefficient of a key monitoring point is greater than the preset defect interference radiation coefficient, the point anomaly coefficient will be reduced and adjusted according to the regional distribution distance and regional density coefficient, and a track operation warning will be issued.

[0081] If the defect interference radiation coefficient of a key monitoring point is less than or equal to the preset defect interference radiation coefficient, the point anomaly coefficient is reduced according to the defect interference radiation coefficient. The reduction value of the point anomaly coefficient is positively correlated with the defect interference radiation coefficient.

[0082] The defect interference radiation coefficient is determined based on the defect interference degree and the defect distribution degree. The condition for completing the abnormal defect analysis is that the defect interference degree of the key monitoring points is determined.

[0083] For a single key monitoring point, the defect interference radiation coefficient is calculated as ln(defect interference degree × defect distribution degree), where the defect distribution degree is the number of dense key areas. For a single key monitoring area, if the number of key monitoring areas within its distribution analysis range is greater than the preset number of areas, then the key monitoring area is designated as a dense key area. The midpoint of the distribution analysis range is the location of the key monitoring area. The distribution analysis range is the set of locations on the track whose distance from the location of the key monitoring area is less than the preset analysis distance. The value of the preset analysis distance can be determined by the user based on the actual working scenario. For example, the user can determine the value based on the subway operation monitoring... The monitoring records are set, and a preset analysis distance value is provided. The preset analysis distance value is 5% of the length of the track surface covered by the preset monitoring range corresponding to the key monitoring point. The value of the preset area number can be determined by the user according to the actual working scenario. For example, the user can set it according to the subway operation monitoring records. The higher the user's requirements for the effectiveness of the operation parameter optimization results, the smaller the value of the preset area number. A method for determining the preset area number is provided, which is the minimum value of the number of key monitoring areas within the distribution analysis range of dense key areas in the subway operation monitoring records that meets the user's requirements for the effectiveness of the operation parameter optimization results.

[0084] The unadjusted point anomaly coefficient and the reference noise parameter of the key monitoring point corresponding to the point anomaly coefficient are positively correlated. The value of the preset defect interference radiation coefficient can be determined by the user according to the actual working scenario. For example, the user can set it according to the subway operation monitoring record. A method for determining the value of the preset defect interference radiation coefficient is provided. The subway operation monitoring record that reduces the point anomaly coefficient according to the defect interference radiation coefficient is recorded as the compensation reference record. The maximum value of the defect interference radiation coefficient of each key monitoring point in the compensation reference record that meets the user's requirements for the effectiveness of the operation parameter optimization results is recorded as the preset defect interference radiation coefficient.

[0085] Specifically, if the defect interference radiation coefficient of a key monitoring point is greater than the preset defect interference radiation coefficient, the key monitoring area of ​​the key monitoring point is divided into dense areas. The regional interference coefficient is determined according to the regional distribution distance and regional density coefficient of each dense area, and the point anomaly coefficient is reduced and adjusted according to the regional interference coefficient.

[0086] The decrease in the point anomaly coefficient is positively correlated with the regional interference coefficient;

[0087] The regional interference coefficient is negatively correlated with the regional distribution distance, and positively correlated with the regional density coefficient.

[0088] Specifically, for any dense area, each of the key monitoring areas contained therein exists within the distribution analysis range of at least one dense key area within that dense area. For a single key monitoring point, the regional interference coefficient... m represents the number of densely populated areas at this key monitoring point, and s represents the number of areas at this key monitoring point. k Let g be the regional density coefficient of the k-th dense area of ​​this key monitoring point. k The interference determination coefficient is the regional density coefficient corresponding to the kth dense area of ​​the key monitoring point. The regional density coefficient is the number of key monitoring areas contained in the dense area. The interference determination coefficient corresponding to the regional density coefficient of each dense area is negatively correlated with the regional distribution distance of the corresponding dense area. For a single dense area, the regional distribution distance is the minimum distance between the location of each key monitoring area and the location of the key monitoring point in the dense area.

[0089] Specifically, when performing parameter optimization analysis for a key monitoring point, the setting method of the point anomaly coefficient is determined based on the frequency of noise anomalies and the proportion of operational interference at the key monitoring point.

[0090] If the frequency of noise anomalies at key monitoring points is greater than the preset frequency of noise anomalies and the proportion of operational interference is greater than the preset proportion of operational interference, the point anomaly coefficient is determined based on the frequency of noise anomalies and the reference noise parameters. The point anomaly coefficient is positively correlated with the frequency of noise anomalies and the reference noise parameters, respectively.

[0091] If the frequency of noise anomalies at key monitoring points is less than or equal to the preset frequency of noise anomalies or the proportion of operational interference is less than or equal to the preset proportion of operational interference, the point anomaly coefficient is determined based on the reference noise parameters. The point anomaly coefficient is positively correlated with the reference noise parameters.

[0092] For a single key monitoring point, the noise anomaly frequency is the number of times abnormal traffic noise occurs at that key monitoring point within the current noise monitoring period. If the noise amplitude value detected when a train passes through the key monitoring point is greater than the preset noise reference value, then an abnormal traffic noise event is recorded. The operational interference ratio is calculated as: operational overlap frequency / noise anomaly frequency. The operational overlap frequency is the number of times abnormal operational interference noise occurs at that key monitoring point within the current noise monitoring period. If the noise amplitude value detected when a train passes through the key monitoring point is greater than the preset noise reference value, then the train is considered to have experienced abnormal traffic noise. If the speed of travel when passing through the key monitoring point is greater than the reference speed, an abnormal operation interference noise is recorded. The reference speed is the minimum noise amplitude value corresponding to the abnormal passage noise recorded in each record of the subway operation monitoring record. The reference noise parameter is the average noise amplitude value corresponding to the abnormal passage noise determined in each record within the current noise monitoring cycle. If the noise abnormality frequency of the key monitoring point is greater than the preset noise abnormality frequency and the operation interference ratio is greater than the preset operation interference ratio, the point abnormality coefficient and the abnormal noise index are positively correlated. The abnormal noise index is the noise abnormality frequency and the reference noise parameter.

[0093] The values ​​of the preset noise anomaly frequency and the preset operational interference ratio can be determined by the user based on the actual working scenario. For example, the user can set them based on subway operation monitoring records. A method for determining the preset noise anomaly frequency is provided, in which the subway operation monitoring records that determine the point anomaly coefficient based on the noise anomaly frequency and reference noise parameters are recorded as the setting reference records. The minimum value of the noise anomaly frequency of key monitoring points in the setting reference records that meet the user's requirements for the effectiveness of the operation parameter optimization results is recorded as the preset noise anomaly frequency. A method for determining the preset operational interference ratio is provided, in which the minimum value of the operational interference ratio of key monitoring points in the setting reference records that meet the user's requirements for the effectiveness of the operation parameter optimization results is recorded as the preset operational interference ratio.

[0094] Specifically, when determining the location anomaly coefficient based on the frequency of noise anomalies and reference noise parameters, the decision on whether to adjust the location anomaly coefficient is based on the coincident anomaly parameters.

[0095] If the coincidence anomaly parameter is greater than the preset coincidence anomaly parameter, the point anomaly coefficient is reduced and adjusted according to the coincidence anomaly parameter and the proportion of operational interference.

[0096] The decrease in the point anomaly coefficient is positively correlated with the coincidence anomaly parameter and the proportion of operational interference.

[0097] When determining the point anomaly coefficient based on the noise anomaly frequency and reference noise parameters, the point anomaly coefficient = lg(noise anomaly frequency × reference noise parameter). The overlapping anomaly parameter is the average noise amplitude value corresponding to each determined abnormal operation interference noise within the current noise monitoring cycle. The value of the preset overlapping anomaly parameter can be determined by the user according to the actual working scenario. For example, the user can set it based on the subway operation monitoring records. The higher the user's requirement for the effectiveness of the operation parameter optimization results, the smaller the value of the preset overlapping anomaly parameter. A method for determining the value of the preset overlapping anomaly parameter is provided, which records the minimum value of the overlapping anomaly parameter corresponding to each adjustment of the point anomaly coefficient in the setting reference record that meets the user's requirement for the effectiveness of the operation parameter optimization results as the preset overlapping anomaly parameter. If the overlapping anomaly parameter is greater than the preset overlapping anomaly parameter, the point anomaly coefficient is reduced according to the operation interference degree. The reduction value of the point anomaly coefficient is positively correlated with the operation interference degree, which is the product of the overlapping anomaly parameter and the operation interference ratio.

[0098] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0099] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for analyzing subway fault data based on a cloud platform, characterized in that, include: Obtain train operation difference parameters and noise characteristic difference parameters at each key monitoring point to determine the monitoring difference status; Based on the monitoring difference status, a difference analysis strategy is determined for each key monitoring point. The difference analysis strategy is to conduct abnormal defect analysis or parameter optimization analysis for the key monitoring points. In the abnormal defect analysis, the defect analysis method is determined based on the noise period stability coefficient and image gradient difference coefficient of the key monitoring points. The defect analysis method is to perform grayscale analysis of the track surface reference direction for the track surface area image, or to perform threshold extraction analysis based on the proportion of abnormal features. Under the condition that the abnormal defect analysis is completed, the monitoring compensation method is determined according to the defect interference radiation coefficient of the key monitoring points. The monitoring compensation method is to adjust the abnormal coefficient of the point according to the regional distribution distance and regional density coefficient, or to adjust the abnormal coefficient of the point according to the defect interference radiation coefficient. In the parameter optimization analysis, the setting method of the point anomaly coefficient is determined based on the frequency of noise anomalies and the proportion of operational interference at key monitoring points, and whether to adjust the point anomaly coefficient is determined based on the overlapping anomaly parameters. The train operation difference parameter is the sum of the equipment speed difference degree and the equipment acceleration difference degree, wherein the equipment speed difference degree... , Let be the train's speed when it passes this key monitoring point for the i-th time within the current noise monitoring cycle. The noise characteristic difference parameter is the average train speed during each passing of the key monitoring point within the current noise monitoring cycle. It is the sum of the products of the noise amplitude difference value and the noise frequency band difference value, respectively, and their corresponding difference influence coefficients. For a single key monitoring point, it represents the noise amplitude difference value at that key monitoring point during the current noise monitoring cycle. , where n is the number of times a train passes through this key monitoring point during the current noise monitoring period. This represents the noise amplitude value collected when the train passes through this key monitoring point for the i-th time within the current noise monitoring cycle. This is the average noise amplitude value collected when the train passes through the key monitoring point each time within the current noise monitoring cycle. The noise amplitude value is the maximum noise amplitude value when the train passes through the key monitoring point. The noise frequency band difference value of the key monitoring point in the current noise monitoring cycle = cycle coverage frequency band range - overlapping frequency band range. The cycle coverage frequency band range is the absolute value of the difference between the maximum and minimum frequencies of the noise signals collected when the train passes through the key monitoring point each time within the current noise monitoring cycle. The overlapping frequency band range is the sum of the absolute values ​​of the differences between the maximum and minimum frequencies of the overlapping frequency bands of the key monitoring point within the current noise monitoring cycle. For a single key monitoring point, the noise period stability coefficient is the absolute value of the difference between the maximum and minimum values ​​of the fundamental frequency of the noise signal acquired within the current noise monitoring period. The image gradient difference coefficient is the number of gradient difference regions contained in the track surface area image of the key monitoring point. The gradient difference region is the track surface analysis region where the gradient reference difference value is greater than the preset gradient reference difference value. For a single track surface analysis region, the gradient reference difference value is the absolute value of the difference between the gradient reference value of the track surface analysis region and the average value of the gradient reference values ​​of all track surface analysis regions. The defect interference radiation coefficient = ln(defect interference degree × defect distribution degree), where the defect distribution degree is the number of dense key areas. The noise anomaly frequency is the number of times abnormal traffic noise exists at the key monitoring point within the current noise monitoring period. If the noise amplitude value detected when a train passes through the key monitoring point is greater than the preset noise reference value, then an abnormal traffic noise is recorded. The running interference ratio = running overlap frequency / noise anomaly frequency, where the running overlap frequency is the number of times abnormal running interference noise exists at the key monitoring point within the current noise monitoring period. For a single dense area, the regional distribution distance is the minimum distance between the location of each key monitoring area and the location of the key monitoring point within the dense area. The regional density coefficient is the number of key monitoring areas contained within the dense area. The overlap anomaly parameter is the average noise amplitude value corresponding to each determined abnormal operation interference noise within the current noise monitoring cycle. The unadjusted location anomaly coefficient is positively correlated with the reference noise parameter of the key monitoring point corresponding to the location anomaly coefficient.

2. The subway fault data analysis method based on a cloud platform according to claim 1, characterized in that, When any key monitoring point is in the first preset difference state where the noise characteristic difference parameter is less than or equal to the preset noise characteristic difference parameter, the difference analysis strategy is to perform an abnormal defect analysis on that key monitoring point.

3. The subway fault data analysis method based on a cloud platform according to claim 2, characterized in that, When any key monitoring point is in the second preset difference state where the train operation difference parameter is greater than the preset train operation difference parameter and the noise characteristic difference parameter is greater than the preset noise characteristic difference parameter, the difference analysis strategy is to perform parameter optimization analysis for that key monitoring point.

4. The subway fault data analysis method based on a cloud platform according to claim 3, characterized in that, When performing abnormal defect analysis on a key monitoring point, the defect analysis method is determined based on the noise period stability coefficient and image gradient difference coefficient of the key monitoring point. If the noise period stability coefficient of the key monitoring point is greater than the preset noise stability coefficient and the image gradient difference coefficient is less than or equal to the preset image gradient difference coefficient, then perform grayscale analysis of the track surface reference direction for the track surface area image. If the noise period stability coefficient of the key monitoring point is less than or equal to the preset noise stability coefficient or the image gradient difference coefficient is greater than the preset image gradient difference coefficient, then the extraction threshold analysis is performed based on the proportion of abnormal features.

5. The subway fault data analysis method based on a cloud platform according to claim 4, characterized in that, When performing grayscale analysis of the track surface reference direction for the track surface area image, the track surface area image corresponding to the key monitoring point is segmented to obtain several track surface analysis areas with the same area. The defect anomaly coefficient of each track surface analysis area is determined based on the lateral grayscale difference parameter and the extended grayscale difference parameter. The defect interference degree of the key monitoring point is determined based on the proportion of the key area and the reference anomaly coefficient. The track surface analysis area where the defect anomaly coefficient is greater than the preset defect anomaly coefficient is designated as the key monitoring area.

6. The subway fault data analysis method based on a cloud platform according to claim 5, characterized in that, When performing threshold extraction analysis based on the proportion of abnormal features, abnormal features are extracted for the track surface area image corresponding to the key monitoring points. Track surface reference features are determined based on the proportion of abnormal features in each category. Abnormal features in categories with a proportion of abnormal features less than the preset proportion of abnormal features are all recorded as track surface reference features. The region extraction threshold is determined based on the proportion of abnormal features in the track surface reference features and the feature threshold parameter. The key monitoring areas of the key monitoring points are determined based on the region extraction threshold, and the defect interference degree of the key monitoring points is determined based on the proportion of the key areas. The degree of defect interference is positively correlated with the proportion of key areas.

7. The subway fault data analysis method based on a cloud platform according to claim 6, characterized in that, Once the abnormal defect analysis is completed, the monitoring compensation method is determined based on the defect interference radiation coefficient of the key monitoring points. If the defect interference radiation coefficient of a key monitoring point is greater than the preset defect interference radiation coefficient, the point anomaly coefficient will be reduced and adjusted according to the regional distribution distance and regional density coefficient, and a track operation warning will be issued. If the defect interference radiation coefficient of a key monitoring point is less than or equal to the preset defect interference radiation coefficient, the point anomaly coefficient is reduced according to the defect interference radiation coefficient. The reduction value of the point anomaly coefficient is positively correlated with the defect interference radiation coefficient. The defect interference radiation coefficient is determined based on the defect interference degree and the defect distribution degree. The condition for completing the abnormal defect analysis is that the defect interference degree of the key monitoring points is determined.

8. The subway fault data analysis method based on a cloud platform according to claim 7, characterized in that, If the defect interference radiation coefficient of the key monitoring point is greater than the preset defect interference radiation coefficient, the key monitoring area of ​​the key monitoring point is divided into dense areas. The regional interference coefficient is determined according to the regional distribution distance and regional density coefficient of each dense area. The point anomaly coefficient is reduced and adjusted according to the regional interference coefficient. The decrease in the location anomaly coefficient is positively correlated with the regional interference coefficient, the regional interference coefficient is negatively correlated with the regional distribution distance, and the regional interference coefficient is positively correlated with the regional density coefficient.

9. The subway fault data analysis method based on a cloud platform according to claim 8, characterized in that, When performing parameter optimization analysis for a key monitoring point, the setting method of the point anomaly coefficient is determined based on the frequency of noise anomalies and the proportion of operational interference at the key monitoring point. If the frequency of noise anomalies at key monitoring points is greater than the preset frequency of noise anomalies and the proportion of operational interference is greater than the preset proportion of operational interference, the point anomaly coefficient is determined based on the frequency of noise anomalies and the reference noise parameters. The point anomaly coefficient is positively correlated with the frequency of noise anomalies and the reference noise parameters, respectively. If the frequency of noise anomalies at key monitoring points is less than or equal to the preset frequency of noise anomalies or the proportion of operational interference is less than or equal to the preset proportion of operational interference, the point anomaly coefficient is determined based on the reference noise parameters. The point anomaly coefficient is positively correlated with the reference noise parameters.

10. The subway fault data analysis method based on a cloud platform according to claim 9, characterized in that, When determining the location anomaly coefficient based on the frequency of noise anomalies and reference noise parameters, the need for adjustment of the location anomaly coefficient is determined based on the coincident anomaly parameters. If the coincidence anomaly parameter is greater than the preset coincidence anomaly parameter, the point anomaly coefficient is increased and adjusted according to the coincidence anomaly parameter and the proportion of operational interference. The increase in the point anomaly coefficient is positively correlated with the coincidence anomaly parameter and the proportion of operational interference.

Images