Adaptive evaluation method and system for catenary operation state

Abstract

The application discloses a catenary operation state self-adaptive evaluation method and system; the method comprises the following steps: constructing a catenary operation state evaluation index system, and acquiring time sequence detection data of each evaluation index; dynamically calculating the adaptive window length at the current evaluation moment, extracting a data subset matched with the adaptive window length from the time sequence detection data; calculating a first objective weight vector by using an improved CRITIC method; updating a gain coefficient vector, and calculating a second objective weight vector; combining the second objective weight vector with a subjective weight vector determined by using an analytic hierarchy process by using a combined weight method, generating a final combined weight vector, and evaluating the catenary operation state at the current evaluation moment until the evaluation task is completed; and the method and system effectively improve the accuracy and timeliness of the catenary state evaluation.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of state evaluation, and particularly relates to an adaptive evaluation method and system for the running state of a catenary. BACKGROUND

[0002] As the power source of the traction power supply system of a high-speed railway, the running state of the catenary directly relates to the current collection quality and operation safety of the train. Precise evaluation of the running state of the catenary is a key technical basis for realizing the transition from "scheduled repair" to "state repair". At present, the core of constructing an evaluation model lies in how to scientifically determine the weight system of each evaluation index.

[0003] Existing weight determination methods mainly include subjective weighting methods, objective weighting methods, and combination weighting methods combining the two. Subjective weighting methods such as the analytic hierarchy process (AHP) rely on expert experience, but have the problem of strong subjectivity; objective weighting methods such as the CRITIC method are based on data-driven, but generally adopt a static weight mode, i.e., once the weight is determined, it remains unchanged. However, the catenary is a dynamic system, and the degradation rate and importance of each index will change over time. Static weights are difficult to adapt to the dynamic evolution characteristics of the running state of the catenary, resulting in lagging evaluation results.

[0004] It is worth noting that in the weight system calculation process, the basic range of data is crucial to the accuracy of objective weights. Existing objective weighting methods (such as the standard CRITIC method) do not consider the time dimension characteristics of data. Essentially, this method regards the system as static, and the calculated weight reflects the "average" importance of the index in the entire historical period. However, for a dynamic system such as the catenary, the mutual relationship between its state and index will evolve over time. Using global static data to calculate the weight cannot capture the fluctuations in the importance of the index caused by recent state changes, resulting in an evaluation model that cannot sensitively reflect the current latest running situation and is difficult to identify new risks. Therefore, how to adaptively determine the length of the data time window and establish a dynamic weight optimization mechanism with self-learning ability has become a key challenge to improve the accuracy and timeliness of the evaluation of the catenary state.

[0005] In summary, due to its static nature, neglect of the time dimension of data, and lack of adaptive optimization ability, the existing weight determination method is difficult to meet the demand for precise and real-time evaluation of the dynamically changing catenary state. Therefore, there is an urgent need to develop a new adaptive evaluation method that can dynamically adjust the weight and has self-learning ability. SUMMARY

[0006] To solve the problems in the background art, the purpose of the present application is to provide an adaptive evaluation method for the running state of a catenary, which comprises the following steps:

[0007] constructing an overhead line operation state evaluation index system, and obtaining time sequence detection data of each evaluation index under the index system;

[0008] dynamically calculating an adaptive window length at a current evaluation time extracting a data subset matched with the adaptive window length from the time sequence detection data;

[0009] based on the data subset, a first objective weight vector is calculated by using an improved CRITIC method;

[0010] according to a preset update period, a gain coefficient vector is updated according to a historical evaluation result and an actual operation state, and a second objective weight vector is obtained according to the first objective weight vector and the updated gain coefficient vector;

[0011] a combination weight method is used to combine the second objective weight vector and a subjective weight vector determined by an analytic hierarchy process, so as to generate a final combination weight vector at the current evaluation time;

[0012] based on the final combination weight vector at the current evaluation time, an overhead line evaluation value is obtained, and an overhead line operation state at the current evaluation time is evaluated;

[0013] wherein the method for dynamically calculating the adaptive window length at the current evaluation time is as follows:

[0014] an update law considering a warning accuracy deviation change rate is designed, a window length is dynamically adjusted, and a preliminary adaptive window length is obtained:

[0015]

[0016] wherein, is a second step factor, is a weight factor, is a warning accuracy deviation rate factor; is a pseudo-gradient at a time point of updated online by using a pseudo-gradient estimation algorithm; is a warning accuracy deviation of an evaluation model, is a window length change amount, is a pseudo-gradient at a time point of -1;

[0017] a minimum value and a maximum value of the adaptive window length are set, and a final adaptive window length at the current evaluation time is obtained.

[0018] ​Furthermore, the accuracy deviation of the warning The calculation method is as follows:

[0019] Within a preset update period, evaluations are performed based on the adaptive window length of the previous evaluation time. The resulting dataset consists of all the comprehensive evaluation values ​​corresponding to the adaptive window length of the previous evaluation time. ;

[0020] Define the target value corresponding to the actual state. If a comprehensive evaluation value is obtained If a fault or repair occurs in the overhead contact system within a preset warning period, then ;otherwise, ;

[0021] The accuracy deviation of the early warning For comprehensive evaluation value Its target value The root mean square error, , For the dataset The total number of samples.

[0022] Furthermore, based on the aforementioned data subset, the first objective weight vector is calculated using the improved CRITIC method, including the following steps;

[0023] Get The evaluation time is to extract a window of length from the time-series detection data. Recent data subset This constitutes an evaluation matrix. ,in This represents the sample number within the window. The serial number represents the evaluation indicator. The number of evaluation indicators; Indicates the evaluation time The The first sample Each indicator value corresponds to the evaluation matrix. Normalization is performed to obtain the standardized matrix. , For the normalized version The first sample Individual indicator values;

[0024] Based on the standardized matrix Calculate the first Standard deviation coefficient of each evaluation indicator , ,in For the first Each evaluation metric in adaptive window length within the standard deviation, For the first Each evaluation metric in adaptive window length The mean within;

[0025] Calculate the first Conflicts among evaluation indicators , ,in For adaptive window length Indicators are calculated from the data within. With indicators The Pearson correlation coefficient;

[0026] Calculate the first The first objective weight of each evaluation indicator The first objective weights of all evaluation indicators constitute the first objective weight vector. .

[0027] Furthermore, according to a preset update cycle, the gain coefficient vector is updated based on historical evaluation results and actual operating conditions, including the following steps:

[0028] Define update cycle The gain coefficient vector inside is ,in For the first The gain coefficient of each evaluation index, initially with a value of 1. The serial number represents the evaluation indicator. To evaluate the number of indicators; to count the number of indicators. Each evaluation indicator was updated in the past pre-set update cycle. The number of times it was assessed as a severe condition. and the number of times the failure actually occurred afterward. ;

[0029] Calculate the first Early warning accuracy of each evaluation indicator , It is a smoothing factor;

[0030] Calculate the update cycle Inner The gain coefficient of each evaluation indicator ,in For learning rate, To pre-set the reliability threshold;

[0031] The gain coefficient vector K(T) is obtained based on the gain coefficient of the obtained evaluation index.

[0032] Furthermore, the step of combining the second objective weight vector with the subjective weight vector determined by the analytic hierarchy process using the combined weighting method to generate the final combined weight vector at the current evaluation moment includes the following steps:

[0033] The subjective weights of each evaluation indicator are calculated using the analytic hierarchy process (AHP), and the relative importance of the indicators is quantified through expert scoring to construct a judgment matrix. ,in Indicates the first The first evaluation indicator and the first The relative importance of each evaluation indicator , , The number of evaluation indicators;

[0034] After the judgment matrix passes the consistency test, the eigenvectors of the judgment matrix are calculated and standardized to obtain the analytic hierarchy process (AHP) weight vectors of the evaluation indicators. ,in, Indicates the first The weights of each evaluation indicator using the analytic hierarchy process;

[0035] The second objective weight vector is obtained by using the combined weight method. Compared with the subjective weight vector determined by the analytic hierarchy process By combining the weights, we obtain the final combined weight vector at the current evaluation time. ,in, Indicates the first The second objective weight of each evaluation indicator , .

[0036] Furthermore, based on the first objective weight vector and the updated gain coefficient vector, the second objective weight vector is obtained, including the following steps:

[0037] Multiply the first objective weight vector by the updated gain coefficient vector to obtain the second objective weight vector. .

[0038] Furthermore, the overhead contact line operation status evaluation index system includes pull-out value evaluation index and conductor height evaluation index reflecting safety performance; hard point evaluation index and height difference evaluation index reflecting smoothness performance; and pantograph-catenary pressure evaluation index and conductor slope evaluation index reflecting current collection performance.

[0039] This invention also provides an adaptive evaluation system for the operating status of overhead contact lines, comprising:

[0040] The status assessment index system module is used to construct an index system for assessing the operating status of the overhead contact system and to obtain time-series detection data of each assessment index under the index system.

[0041] The data subset construction module is used to dynamically calculate the adaptive window length at the current evaluation time. Extract the nearest match to the adaptive window length from the time-series detection data. The matched subset of data;

[0042] The first objective weight vector calculation module is used to calculate the first objective weight vector based on a data subset using the improved CRITIC method.

[0043] The gain coefficient vector update module is used to update the gain coefficient vector according to a preset update cycle, based on historical evaluation results and actual operating conditions; and to obtain a second objective weight vector based on the first objective weight vector and the updated gain coefficient vector.

[0044] The final combined weight calculation module is used to combine the second objective weight vector with the subjective weight vector determined by the analytic hierarchy process using the combined weight method to generate the final combined weight vector at the current evaluation time.

[0045] The operation status assessment module is used to obtain the catenary assessment value based on the final combined weight vector at the current assessment time, and to assess the catenary operation status at the current assessment time.

[0046] Furthermore, the present invention adopts the following technical solution:

[0047] A non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the adaptive evaluation method for the operating status of the overhead contact line as described above.

[0048] Furthermore, the present invention adopts the following technical solution:

[0049] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the adaptive evaluation method for the operating status of the overhead contact line as described above.

[0050] The beneficial technical effects of this invention are as follows:

[0051] (1) Dynamic adaptive optimization of weights is realized: by dynamically adjusting the length of the data window, the weight calculation can focus on recent running data, effectively overcoming the problem of the traditional static weights responding to state changes with lag, and significantly improving the timeliness of the evaluation;

[0052] (2) Solved the adaptability problem of fixed time window: The sensitivity and stability of the evaluation are automatically balanced by the data-driven approach, which avoids the slow response caused by long window and prevents the result fluctuation caused by short window, thus enhancing the applicability of the method under different operating conditions.

[0053] (3) A weight self-learning mechanism was established: Through the adaptive update of the gain coefficient, the model can continuously optimize the weight allocation based on the historical evaluation results, thereby achieving a continuous improvement in evaluation accuracy and enhancing the reliability of the model.

[0054] (4) A complete intelligent evaluation framework has been constructed: dynamic window selection, weight calculation and feedback correction are organically combined, providing a more scientific and adaptive weight determination method for the evaluation of the contact network status, and providing a reliable basis for accurate operation and maintenance decisions. Attached Figure Description

[0055] Figure 1 This invention provides an overhead contact line operation status evaluation system for embodiments of the invention.

[0056] Figure 2 A flowchart illustrating the adaptive evaluation method for overhead contact line operation status provided in an embodiment of the present invention;

[0057] Figure 3 This is a schematic diagram of the connection of the adaptive evaluation system module for the operation status of the overhead contact line provided in an embodiment of the present invention. Detailed Implementation

[0058] This invention discloses an adaptive evaluation method and system for the operating status of overhead contact lines. The method includes constructing an evaluation system for the operating status of the overhead contact lines and acquiring time-series detection data for each evaluation index; dynamically calculating the adaptive window length at the current evaluation time and extracting a recent data subset from the time-series detection data; calculating a first objective weight vector using an improved CRITIC method; updating the gain coefficient vector and calculating a second objective weight vector; and combining the second objective weight vector with a subjective weight vector determined by the analytic hierarchy process using a combined weight method to generate a final combined weight vector, which is then used to evaluate the operating status of the overhead contact lines at the current evaluation time until the evaluation task is completed. The method and system of this invention effectively improve the accuracy and timeliness of overhead contact line status evaluation.

[0059] The adaptive evaluation method and system for overhead contact line operation status provided by the present invention will be further described clearly and completely below with reference to the accompanying drawings:

[0060] Example 1

[0061] Figure 1 This embodiment provides an evaluation system for the operating status of overhead contact lines. Figure 2This is a flowchart illustrating the adaptive evaluation method for the operating status of an overhead contact line provided in an embodiment of the present invention. This embodiment provides an adaptive evaluation method for the operating status of an overhead contact line, the method comprising the following steps:

[0062] Step (1): Construct a catenary operation status evaluation system and obtain time-series detection data of each evaluation index under the evaluation system; the evaluation index includes pull-out value evaluation index and conductor height evaluation index reflecting safety performance, hard point evaluation index and height difference evaluation index reflecting smoothness performance, and pantograph-catenary pressure evaluation index and conductor slope evaluation index reflecting current collection performance; wherein, the time-series detection data are the time-series measurement values ​​of these index parameters, which are obtained through dynamic or static measurement by catenary detection equipment. Specifically: pull-out value and conductor height are directly measured by catenary geometric parameter detection device (such as laser measuring instrument or image recognition system); hard point and height difference are calculated based on pantograph acceleration sensor or continuous scanning data of contact wire height; pantograph-catenary pressure is collected in real time by force sensor and arc detection device mounted on pantograph; conductor slope is obtained by dynamic detection data of pull-out value, and then post-processing to calculate its rate of change over a certain baseline length;

[0063] Step (2): Dynamically calculate the adaptive window length at the current evaluation time. According to the adaptive window length Extract the nearest match to the adaptive window length from the time-series detection data. The matched subset of data;

[0064] Step (3): Based on the data subset, calculate the first objective weight vector using the improved CRITIC method;

[0065] Step (4): Update the gain coefficient vector according to the preset update cycle, based on the historical evaluation results and the actual operating conditions; multiply the first objective weight vector with the updated gain coefficient vector to obtain the second objective weight vector;

[0066] Step (5): Use the combined weighting method to combine the second objective weight vector with the subjective weight vector determined by the analytic hierarchy process to generate the final combined weight vector at the current evaluation time;

[0067] Step (6): Evaluate the operating status of the overhead contact system at the current evaluation time based on the final combined weight vector at the current evaluation time;

[0068] Repeat steps (2) to (6) to achieve continuous evaluation of the operating status of the touch screen.

[0069] According to the adaptive evaluation method for overhead contact line operation status provided by the present invention, step (2) involves dynamically calculating the adaptive window length at the current evaluation time. This includes the following steps:

[0070] Step (2.1): Define the early warning accuracy bias of the evaluation model. , For the evaluation time;

[0071] The accuracy deviation of the early warning The calculation method is as follows:

[0072] Within a preset update period, evaluations are performed based on the adaptive window length of the previous evaluation time. The resulting dataset consists of all the comprehensive evaluation values ​​corresponding to the adaptive window length of the previous evaluation time. ;

[0073] It should be noted that the comprehensive evaluation value refers to the overall evaluation value of the entire overhead contact system. In the calculation, the comprehensive evaluation value is a single numerical evaluation result obtained by weighting and combining all indicator data. First, the original data of each indicator is normalized. Then, the normalized data is multiplied element by element by the final combined weight vector. Finally, the weighted results of all indicators are summed to obtain the comprehensive evaluation value.

[0074] Define the target value corresponding to the comprehensive evaluation value. The status level classification criteria in Table 1 are well known to those skilled in the art. If the comprehensive evaluation value... If the status falls under the "Severe" level in Table 1 and a fault or repair occurs in the overhead contact system within a preset warning time (it should be noted that the warning time refers to the reasonable time interval between the discovery of an anomaly in the overhead contact system during assessment and the actual implementation of maintenance actions, usually set at 7-30 days, depending on the characteristics of the line and maintenance capabilities), then... ;otherwise, ;

[0075] The accuracy deviation of the early warning For comprehensive evaluation value Its target value The root mean square error, , For the dataset The total number of samples in the sample;

[0076] Table 1. Criteria for Classifying Status Levels

[0077] Rank Excellent Good Attention Severe Comprehensive evaluation value (0.7653,1] (0.479,0.7653] (0.1870,0.4790] (0,0.1870]

[0078] Step (2.2): Update online using the pseudo gradient estimation algorithm :

[0079]

[0080] in, For the first step of the long factor, As a penalty factor;

[0081] Step (2.3): Design an update law that considers the rate of change of the early warning accuracy deviation, dynamically adjust the window length, and obtain a preliminary adaptive window length. :

[0082]

[0083] in, This is the second step length factor. As a weighting factor, This is the rate factor for deviation in early warning accuracy. This represents the change in window length. for The pseudo gradient at time -1;

[0084] In step (2.3), based on the results obtained in steps (2.1) and (2.2) , Perform preliminary adaptive window length The calculation of the adaptive window length at the current evaluation time requires, when finally determining the adaptive window length, to use the preliminary adaptive window length calculated in step (2.3). Constrained at a preset minimum value and maximum value Between, that is, the adaptive window length at the final current evaluation time. .

[0085] According to the adaptive evaluation method for catenary operation status provided by the present invention, step (3) of calculating the first objective weight vector based on the data subset using the improved CRITIC method includes the following steps;

[0086] Step (3.1): Obtain The evaluation time is to extract a window of length from the time-series detection data. Recent data subset This constitutes an evaluation matrix. ,in This represents the sample number within the window. The serial number represents the evaluation indicator. The number of evaluation indicators; Indicates the evaluation time The The first sample Regarding the measurement values ​​of these indicators, it should be noted that these parameters are obtained through dynamic or static measurements using contact wire inspection equipment. Specifically: pull-out value and conductor height are directly measured using contact wire geometric parameter detection devices (such as laser measuring instruments or image recognition systems); hard points and height differences are calculated based on pantograph acceleration sensors or continuous scanning data of the contact wire height; pantograph-catenary pressure is collected in real time by force sensors and arc detection devices mounted on the pantograph; conductor slope is obtained by dynamically detecting pull-out value data and then calculating its rate of change over a certain baseline length through post-processing; and the evaluation matrix... Normalization is performed to obtain the standardized matrix. , The normalized version The first sample The measured values ​​of each indicator;

[0087] Step (3.2): Based on the standardized matrix Calculate the first Standard deviation coefficient of each evaluation indicator , ,in For the first Each evaluation metric in adaptive window length within the standard deviation, For the first Each evaluation metric in adaptive window length The mean within;

[0088] Step (3.3): Calculate the first... Conflicts among evaluation indicators , ,in For adaptive window length Indicators are calculated from the data within. With indicators The Pearson correlation coefficient is calculated using the following formula:

[0089]

[0090] in, For the first Each evaluation metric in adaptive window length The mean within, For the first Each evaluation metric in adaptive window length The mean within, , ;

[0091] Step (3.4): Calculate the first... The first objective weight of each evaluation indicator The first objective weights of all evaluation indicators constitute the first objective weight vector. .

[0092] According to the adaptive evaluation method for overhead contact line operation status provided by the present invention, step (4) involves updating the gain coefficient vector according to the historical evaluation results and actual operation status at a preset update cycle; and multiplying the first objective weight vector with the updated gain coefficient vector for correction to obtain the second objective weight vector, which includes the following steps:

[0093] Step (4.1): Define the update cycle The gain coefficient vector inside is ,in For the first The gain coefficient of each evaluation index, initially with a value of 1. The serial number represents the evaluation indicator. To evaluate the number of indicators; to count the number of indicators. Each evaluation indicator was updated in the past pre-set update cycle. The number of times it was assessed as a severe condition. and the number of times the failure actually occurred afterward. ;

[0094] Step (4.2): Calculate the first... Early warning accuracy of each evaluation indicator , It is a smoothing factor;

[0095] Step (4.3): Calculate the update cycle Inner The gain coefficient of each evaluation indicator ,in For learning rate, To pre-set the reliability threshold;

[0096] The gain coefficient vector K(T) is obtained based on the gain coefficient of the obtained evaluation index;

[0097] Step (4.4): Multiply the first objective weight vector with the updated gain coefficient vector to obtain the second objective weight vector. .

[0098] According to the adaptive evaluation method for the operating status of the overhead contact line provided by the present invention, step (5) involves using a combined weighting method to combine the second objective weight vector with the subjective weight vector determined by the analytic hierarchy process to generate the final combined weight vector at the current evaluation moment. This includes the following steps:

[0099] Step (5.1): Calculate the subjective weights of each evaluation indicator using the analytic hierarchy process (AHP), quantify the relative importance of the evaluation indicators through expert scoring, and construct a judgment matrix. ,in Indicates the first The first evaluation indicator and the first The relative importance of each evaluation indicator , , The number of evaluation indicators;

[0100] After the judgment matrix passes the consistency test, the eigenvectors of the judgment matrix are calculated and standardized to obtain the analytic hierarchy process (AHP) weight vectors of the evaluation indicators. ,in, Indicates the first The weights of each evaluation indicator using the analytic hierarchy process;

[0101] Specifically, performing a consistency check on the judgment matrix to determine whether there are contradictions within the matrix is ​​well known to those skilled in the art. The order of the judgment matrix is ​​less than or equal to 7, and the formula for calculating the consistency check is as follows:

[0102] ;

[0103] ;

[0104] Here, the order n of the judgment matrix represents the number of evaluation indicators. Let CI be the largest eigenvalue of the judgment matrix, CR be the random consistency ratio, and RI be the average random consistency index. The average random consistency index of each order judgment matrix is:

[0105] n 1 2 3 4 5 6 7 RI 0 0 0.52 0.89 1.12 1.26 1.36

[0106] If CR < 0.1, the judgment matrix passes the consistency test; if CR ≥ 0.1, the judgment matrix needs to be adjusted. , , Optimize the consistency of the judgment matrix until CR < 0.1;

[0107] For the judgment matrix that passes the consistency test, solve its characteristic equation. To obtain the largest eigenvalue Corresponding feature vector ; for the feature vector Standardization processing is required. The weights of the analytic hierarchy process are obtained. ,in express The One element, ;

[0108] Step (5.2): Use the combined weighting method to combine the second objective weight vector Compared with the subjective weight vector determined by the analytic hierarchy process By combining the weights, we obtain the final combined weight vector at the current evaluation time. ,in, Indicates the first The second objective weight of each evaluation indicator , .

[0109] Based on the final combined weight vector at the current assessment time, the original data of each indicator are normalized and then multiplied element-wise with the final combined weight vector. Finally, the weighted results of all indicators are summed to obtain the comprehensive assessment value. According to the comprehensive assessment value, the catenary operation status at the current assessment time is assessed according to the contents of Table 1. According to the status level classification standard given in Table 1, different assessment values ​​fall into different levels (e.g., excellent, good, attention, serious). The operators arrange different maintenance tasks according to different levels, and finally realize the status assessment and maintenance of the catenary.

[0110] The method of this invention dynamically adjusts the data window length, enabling weight calculation to focus on recent operational data. This effectively overcomes the problem of delayed response to state changes in traditional static weights, significantly improving the timeliness of the evaluation. Furthermore, by dynamically adjusting the data window length, the adaptability problem of fixed time windows is solved. Through a data-driven approach, the sensitivity and stability of the evaluation are automatically balanced, avoiding both the sluggish response caused by long windows and the result fluctuations caused by short windows, thus enhancing the applicability of the method under different operating conditions. In addition, by updating the gain coefficient within a preset update cycle, this invention enables the model to continuously optimize the weight allocation based on historical evaluation results, establishing a weight self-learning mechanism. This achieves continuous improvement in evaluation accuracy and enhances the reliability of the model.

[0111] In summary, this invention organically combines dynamic window selection, weight calculation, and feedback correction, providing a more scientific and adaptive weight determination method for overhead contact line status assessment and a reliable foundation for accurate operation and maintenance decisions.

[0112] Example 2

[0113] Figure 3 This is a schematic diagram of the module connection for the adaptive evaluation system for catenary operation status provided in this embodiment; this embodiment provides an adaptive evaluation system for catenary operation status, including:

[0114] The status assessment index system module is used to construct an index system for assessing the operating status of the overhead contact system and to obtain time-series detection data of each assessment index under the index system.

[0115] The data subset construction module is used to dynamically calculate the adaptive window length at the current evaluation time. According to the adaptive window length Extract the nearest match to the adaptive window length from the time-series detection data. The matched subset of data;

[0116] The first objective weight vector calculation module is used to calculate the first objective weight vector based on the data subset using the improved CRITIC method.

[0117] The gain coefficient vector update module is used to update the gain coefficient vector according to the historical evaluation results and actual operating conditions at a preset update cycle; and to correct by multiplying the first objective weight vector with the updated gain coefficient vector to obtain the second objective weight vector.

[0118] The final combined weight calculation module is used to combine the second objective weight vector with the subjective weight vector determined by the analytic hierarchy process using the combined weight method to generate the final combined weight vector at the current evaluation time.

[0119] The operation status assessment module is used to assess the operation status of the overhead contact system at the current assessment time based on the final combined weight vector at the current assessment time.

[0120] Furthermore, the present invention adopts the following technical solution:

[0121] A non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the adaptive evaluation method for the operating status of the overhead contact line as described above.

[0122] Furthermore, the present invention adopts the following technical solution:

[0123] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the adaptive evaluation method for the operating status of the overhead contact line as described above.

[0124] From the above description of the embodiments, those skilled in the art will clearly understand that the facilities of the present invention can be implemented using software plus necessary general-purpose hardware platforms. Embodiments of the present invention can be implemented using existing processors, or by dedicated processors used for this or other purposes for suitable systems, or by hardwired systems. Embodiments of the present invention also include non-transitory computer-readable storage media, comprising machine-readable media for carrying or having machine-executable instructions or data structures stored thereon; such machine-readable media can be any available medium accessible by a general-purpose or special-purpose computer or other machine with a processor. For example, such machine-readable media can include RAM, ROM, EPROM, EEPROM, CD-ROM or other optical disc storage, disk storage or other magnetic storage devices, or any other medium that can be used to carry or store the required program code in the form of machine-executable instructions or data structures and is accessible by a general-purpose or special-purpose computer or other machine with a processor. When information is transmitted or provided to a machine via a network or other communication connection (hardwired, wireless, or a combination of hardwired and wireless), that connection is also considered a machine-readable medium.

[0125] 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 such changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. An adaptive evaluation method for the operating status of overhead contact lines, characterized in that, The method includes the following steps: Construct an evaluation index system for the operation status of the overhead contact system, and obtain time-series detection data for each evaluation index under the index system; Dynamically calculate the adaptive window length at the current evaluation time. Extract the nearest match to the adaptive window length from the time-series detection data. The matched subset of data; The first objective weight vector is calculated based on a subset of data using the improved CRITIC method. According to the preset update cycle, the gain coefficient vector is updated based on historical evaluation results and actual operating conditions; the second objective weight vector is obtained based on the first objective weight vector and the updated gain coefficient vector. The second objective weight vector is combined with the subjective weight vector determined by the analytic hierarchy process using the combined weight method to generate the final combined weight vector at the current evaluation time. Based on the final combined weight vector at the current assessment time, the catenary assessment value is obtained, and the catenary operation status at the current assessment time is assessed. Among them, the adaptive window length at the current evaluation time is dynamically calculated. The method is as follows: The design considers the update law of the rate of change of early warning accuracy deviation, dynamically adjusts the window length, and obtains an initial adaptive window length. : ； in, This is the second step length factor. As a weighting factor, This is the rate factor for deviation in early warning accuracy. To utilize the pseudo gradient estimation algorithm to update the post-gradient online The pseudo gradient at time step; To assess the accuracy bias of the model's early warning system, This represents the change in window length. for The pseudo gradient at time -1; ; For the first step of the long factor, As a penalty factor; ; Set the minimum value for adaptive window length. and maximum value The final adaptive window length at the current evaluation time. .

2. The adaptive evaluation method for the operating status of overhead contact lines according to claim 1, characterized in that, The accuracy deviation of the early warning The calculation method is as follows: Within a preset update period, evaluations are performed based on the adaptive window length of the previous evaluation time. The resulting dataset consists of all the comprehensive evaluation values ​​corresponding to the adaptive window length of the previous evaluation time. ; Define the target value corresponding to the actual state. If a comprehensive evaluation value is obtained If a fault or repair occurs in the overhead contact system within a preset warning period, then ;otherwise, ; The accuracy deviation of the early warning For comprehensive evaluation value Its target value The root mean square error, , For the dataset The total number of samples.

3. The adaptive evaluation method for the operating status of the overhead contact line according to claim 1, characterized in that, Based on the data subset, the first objective weight vector is calculated using the improved CRITIC method, including the following steps; Get The evaluation time is to extract a window of length from the time-series detection data. Recent data subset This constitutes an evaluation matrix. ,in This represents the sample number within the window. The serial number represents the evaluation indicator. The number of evaluation indicators; Indicates the evaluation time The The first sample Each indicator value corresponds to the evaluation matrix. Normalization is performed to obtain the standardized matrix. , For the normalized version The first sample Individual indicator values; Based on the standardized matrix Calculate the first Standard deviation coefficient of each evaluation indicator , ,in For the first Each evaluation metric in adaptive window length within the standard deviation, For the first Each evaluation metric in adaptive window length The mean within; Calculate the first Conflicts among evaluation indicators , ,in For adaptive window length Indicators are calculated from the data within. With indicators The Pearson correlation coefficient; Calculate the first The first objective weight of each evaluation indicator The first objective weights of all evaluation indicators constitute the first objective weight vector. .

4. The adaptive evaluation method for the operating status of overhead contact lines according to claim 1, characterized in that, The gain coefficient vector is updated according to a preset update cycle, based on historical evaluation results and actual operating conditions, including the following steps: Define update cycle The gain coefficient vector inside is ,in For the first The gain coefficient of each evaluation index, initially with a value of 1. The serial number represents the evaluation indicator. To evaluate the number of indicators; to count the number of indicators. Each evaluation indicator was updated in the past pre-set update cycle. The number of times it was assessed as a severe condition. and the number of times the failure actually occurred afterward. ; Calculate the first Early warning accuracy of each evaluation indicator , It is a smoothing factor; Calculate the update cycle Inner The gain coefficient of each evaluation indicator ,in For learning rate, To pre-set the reliability threshold; The gain coefficient vector K(T) is obtained based on the gain coefficient of the obtained evaluation index.

5. The adaptive evaluation method for the operating status of overhead contact lines according to claim 1, characterized in that, The step of combining the second objective weight vector with the subjective weight vector determined by the analytic hierarchy process to generate the final combined weight vector at the current evaluation moment using the combined weight method includes the following steps: The subjective weights of each evaluation indicator are calculated using the analytic hierarchy process (AHP), and the relative importance of the indicators is quantified through expert scoring to construct a judgment matrix. ,in Indicates the first The first evaluation indicator and the first The relative importance of each evaluation indicator , , The number of evaluation indicators; After the judgment matrix passes the consistency test, the eigenvectors of the judgment matrix are calculated and standardized to obtain the analytic hierarchy process (AHP) weight vectors of the evaluation indicators. ,in, Indicates the first The weights of each evaluation indicator using the analytic hierarchy process; The second objective weight vector is obtained by using the combined weight method. Compared with the subjective weight vector determined by the analytic hierarchy process By combining the weights, we obtain the final combined weight vector at the current evaluation time. ,in, Indicates the first The second objective weight of each evaluation indicator , .

6. The adaptive evaluation method for the operating status of the overhead contact line according to claim 4, characterized in that, The second objective weight vector is obtained by combining the first objective weight vector with the updated gain coefficient vector, including the following steps: Multiply the first objective weight vector by the updated gain coefficient vector to obtain the second objective weight vector. ,in, This is the first objective weight vector.

7. The adaptive evaluation method for the operating status of overhead contact lines according to claim 1, characterized in that, The overhead contact system operation status evaluation index system includes pull-out value evaluation index and conductor height evaluation index, which reflect safety performance; hard point evaluation index and height difference evaluation index, which reflect smoothness performance; and pantograph-catenary pressure evaluation index and conductor slope evaluation index, which reflect current collection performance.

8. An adaptive evaluation system for catenary operation status, used to implement the adaptive evaluation method for catenary operation status as described in any one of claims 1-7, characterized in that, include: The status assessment index system module is used to construct an index system for assessing the operating status of the overhead contact system and to obtain time-series detection data of each assessment index under the index system. The data subset construction module is used to dynamically calculate the adaptive window length at the current evaluation time. Extract the nearest match to the adaptive window length from the time-series detection data. The matched subset of data; The first objective weight vector calculation module is used to calculate the first objective weight vector based on a data subset using the improved CRITIC method. The gain coefficient vector update module is used to update the gain coefficient vector according to a preset update cycle, based on historical evaluation results and actual operating conditions; and to obtain a second objective weight vector based on the first objective weight vector and the updated gain coefficient vector. The final combined weight calculation module is used to combine the second objective weight vector with the subjective weight vector determined by the analytic hierarchy process using the combined weight method to generate the final combined weight vector at the current evaluation time. The operation status assessment module is used to obtain the catenary assessment value based on the final combined weight vector at the current assessment time, and to assess the catenary operation status at the current assessment time.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the adaptive evaluation method for the operating status of the overhead contact line as described in any one of claims 1 to 7.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the adaptive evaluation method for the operating status of the overhead contact line as described in any one of claims 1 to 7.

Images