Power transformation equipment evaluation method and system based on multi-dimensional parameters

By using a multi-dimensional parameter evaluation method to collect, filter, model, and manage substation equipment data, the problem of insufficient evaluation accuracy of substation equipment is solved, enabling accurate evaluation of equipment status and full-cycle management, thereby improving the reliability of equipment operation and management efficiency.

CN121327684APending Publication Date: 2026-01-13国网湖北省电力有限公司江陵县供电公司 +1

Patent Information

Application Number
CN202511426829.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-09-30
Publication Date
2026-01-13

AI Technical Summary

Technical Problem

The accuracy of the operational status assessment of power equipment in the existing technology is limited, and the monitoring data has not been effectively utilized for optimization, resulting in inaccurate equipment status assessment.

Method used

A multi-dimensional parameter-based evaluation method for power equipment is adopted. The system collects data through a sensing module, filters key parameters through a screening module, establishes a dynamic correlation model through an analysis module, calculates a health index through an evaluation module, generates visual charts and health records through an output module, and automatically updates the data through a management module, thereby achieving full-cycle management.

Benefits of technology

It improves the accuracy and efficiency of health assessment of power equipment, enabling timely detection of potential problems, reducing maintenance costs and safety risks, and ensuring reliable equipment operation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121327684A_ABST
    Figure CN121327684A_ABST
Patent Text Reader

Abstract

The invention discloses a power transformation equipment evaluation method and system based on multi-dimensional parameters, and relates to the field of power transformation equipment safety management, and the method comprises a sensing module which is used for collecting the operation information and initial health characterization data of power transformation equipment, and carrying out the classification of the collected information and data; the screening module is used for receiving the classified information and data in the sensing module, screening key parameters strongly related to the health state of the equipment from the classified information and data, and labeling health influence dimensions corresponding to the parameters; according to the method, by collecting and classifying various types of operation and health data of the power transformation equipment, key parameters highly related to equipment health are accurately screened, meanwhile, health influence dimensions are marked, dimension weights are determined, core evaluation elements can be focused, and irrelevant data interference is reduced.
Need to check novelty before this filing date? Find Prior Art

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power transformation equipment safety management, in particular to a power transformation equipment evaluation method and system based on multi-dimensional parameters. BACKGROUND

[0002] Power transformation equipment is a core link of the power system, connecting the power generation end and the power consumption end, and its core functions are voltage transformation, power distribution and safety control. By monitoring key parameters such as temperature, insulation performance and mechanical properties online, and combining algorithm evaluation health level, the traditional periodic maintenance is replaced, and predictive maintenance is realized, which can avoid fault risks in advance, prolong equipment life, reduce unplanned power outages, and ensure the continuous and stable operation of the power transformation system.

[0003] The patent application with the application number 201811366930.3 discloses a power transmission and transformation equipment health evaluation system based on big data. The application aims to solve the problem that "China's power system has become the world's largest power network related to national economy and people's livelihood. The reliability, efficient operation and effective management of power equipment have become increasingly important for the safety and stability of the power system. How to quickly mine and find the health status and defect information of equipment from massive power equipment monitoring data has become an important concern for researchers and power companies. The analysis and processing of new stream data generated by numerous sensors in the smart grid have brought great challenges to the health evaluation of equipment."

[0004] However, the existing technology for the operation safety evaluation of power transformation equipment is mainly based on the monitoring of power transformation equipment operation data. Although the improvement direction has made great progress in the accuracy of data monitoring, the monitoring data used in the evaluation of the safety of equipment operation state based on monitoring data is not optimized, so that the accuracy of equipment operation state evaluation is limited.

[0005] Therefore, we propose a power transformation equipment evaluation method and system based on multi-dimensional parameters. SUMMARY

[0006] In view of the above shortcomings of the prior art, the present application provides a power transformation equipment evaluation method and system based on multi-dimensional parameters, which can effectively solve the problems of the prior art.

[0007] To achieve the above purpose, the present application realizes the following technical solutions;

[0008] The present application discloses a power transformation equipment evaluation system based on multi-dimensional parameters, which comprises:

[0009] The perception module is used for collecting operation information and initial health characterization data of the power transformation equipment and classifying the collected information and data; the screening module is used for receiving the classified information and data in the perception module, screening key parameters strongly related to the health state of the equipment from the classified information and data, and marking health influence dimensions corresponding to each parameter; the analysis module is used for receiving the screened key parameters with the marking, establishing a dynamic correlation model between each parameter, identifying parameter abnormal correlation characteristics, and synchronously generating a correlation analysis report; the evaluation module is used for obtaining the correlation analysis report, setting a health level evaluation logic, and calculating an equipment health index based on the parameter abnormal correlation characteristics in the analysis report, and synchronously performing a health level division operation; the output module is used for receiving the health level, the health index and the correlation analysis report in the evaluation module, generating a visual health state chart, synchronously predicting a health trend of the power transformation equipment, and packaging the generated visual health state chart and the prediction result to transmit to a preset receiving end; and the management module is used for receiving the visual health state chart and the prediction result to create a time-stamped unique health file of the power transformation equipment.

[0010] The perception module is connected with the screening module through a wireless network, the screening module is connected with the analysis module and the evaluation module through a wireless network, and the analysis module is connected with the output module and the management module through a wireless network.

[0011] Further, the perception module comprises a multi-source perception unit, a data preprocessing unit and a classification unit.

[0012] The multi-source perception unit is configured with special sensors corresponding to different core components of the power transformation equipment, and is used for collecting operation electric parameters of the power transformation equipment, including voltage, current, power, physical state parameters including temperature, vibration, humidity, and initial health characterization parameters including insulation resistance and oil dissolved gas concentration.

[0013] The data preprocessing unit is used for performing outlier rejection, data smoothing and denoising and data normalization processing on the collected raw data, wherein the normalization processing adopts a min-max normalization method to map the data to the interval [0, 1];

[0014] The classification unit classifies the preprocessed data according to a two-level classification standard of equipment components-data types, wherein the equipment components include transformer body, switch equipment and mutual inductor, and the data types include real-time collection data and periodic collection data.

[0015] Further, the screening module performs a process of key parameter screening and health influence dimension marking, which comprises the following steps:

[0016] Calculate the mutual information entropy of each parameter after classification and the historical failure data of the equipment, and screen out candidate parameters with mutual information entropy values greater than a preset threshold.

[0017] The candidate parameters are subjected to grey correlation analysis to calculate the correlation degrees of each parameter with the health benchmark state of the equipment, and the top N parameters in the correlation degrees are further selected as the key parameters;

[0018] The key parameters are labeled to the preset health impact dimensions, the health impact dimensions include a safety dimension associated with the risk of safety accidents caused by equipment failure, a reliability dimension associated with the length of time of normal operation of the equipment, and an economic dimension associated with the maintenance cost and life consumption of the equipment, and based on the number and influence degree of the key parameters under each dimension, the initial weight of each health impact dimension is determined by analytic hierarchy process.

[0019] Further, the operation of establishing a dynamic correlation model and identifying abnormal correlation characteristics of the parameters in the analysis module is:

[0020] The time series data of the selected key parameters are input to construct a dynamic correlation model based on a long short-term memory network, the input layer dimension of the model is consistent with the number of key parameters, the hidden layer is set to 2-3 layers, the output layer is the predicted value of each key parameter, the adaptive moment estimation optimizer is used in the model training process, and the mean square error is used as the loss function;

[0021] The deviation degree of the actual value of the parameter and the predicted value of the model is calculated , represents the actual collection value of the i-th key parameter at time t, represents the model predicted value of the parameter at time t, represents the standard deviation of the historical normal operation data of the parameter;

[0022] When the deviation degree of any key parameter is greater than the preset deviation threshold, or the deviation degrees of two or more key parameters are simultaneously greater than 80% of the respective preset deviation thresholds, it is determined that the identified abnormal correlation characteristics of the parameters, and the abnormal parameter identification, the time of abnormal occurrence and the deviation degree value are recorded in the correlation analysis report.

[0023] Further, the evaluation module is subjected to:

[0024] Based on the abnormal correlation characteristics of each key parameter, the health impact coefficient of each key parameter is determined:

[0025] When the key parameter is normal, the health impact coefficient is 1; when the key parameter has a slight abnormality, the health impact coefficient is 0.8; when the key parameter has a moderate abnormality, the health impact coefficient is 0.5; when the key parameter has a severe abnormality, the health impact coefficient is 0.2, wherein the degree of abnormality is determined by comparing the deviation degree with the respective preset threshold values;

[0026] Recalculating the device health index:

[0027] ;

[0028] wherein: is the initial weight of the jth health impact dimension in the screening module obtained by AHP; is the number of key parameters under the jth health impact dimension; is the set of key parameters under the jth health impact dimension; is the health impact coefficient; is the normalized value of the ith key parameter, which is output by the data preprocessing unit of the perception module and is subject to .

[0029] Further, the health level evaluation logic set in the evaluation module is:

[0030] According to the value range of the device health index and the severity of the parameter abnormality correlation characteristics, the device health level is divided into: excellent, good, general, warning, and failure;

[0031] The grade division logic is:

[0032] When ≥ 0.9 and there is no abnormal key parameter, the health level is excellent; when 0.8≤ < 0.9 and only a single slightly abnormal parameter exists, the health level is good; when 0.6≤ < 0.8 and there are no more than two slightly abnormal parameters or a single moderately abnormal parameter, the health level is general; when 0.4≤ < 0.6 and there is a single severely abnormal parameter or two or more moderately abnormal parameters, the health level is warning; when < 0.4 or there are two or more severely abnormal parameters, the health level is failure.

[0033] Further, the operation content of the output module to generate a visual health status chart and predict the health trend stage of the power transformation device is:

[0034] The generated visual health status chart includes:

[0035] A key parameter time series change curve graph for showing the change trend and abnormal point marking of each key parameter over time;

[0036] A health impact dimension radar chart with the health sub-index of each dimension as the radius for showing the health status of each dimension, wherein the health sub-index is the comprehensive index under each health impact dimension, i.e., the health sub-index of the jth health impact dimension ;

[0037] health level change histogram, for health level and health index change in the last N cycles;

[0038] quantifying the health trend prediction value of the future pth preset period:

[0039]

[0040] wherein: is the preset number of historical periods participating in the prediction; is the time decay coefficient, ∈(0,1); is the health index of the t-kth period; is the trend adjustment factor, >0; is the average change rate of the health index of the last m periods, representing the historical change trend of the health state;

[0041] Finally, the visual health state chart, the health trend prediction value of the future p periods and the corresponding health level prediction interval are packaged according to a preset format, transmitted to a preset receiving end through a preset encryption transmission protocol.

[0042] Further, the management module creates a unique health record of the power transformation equipment process:

[0043] The health record is associated with the unique identifier of the equipment, which includes the equipment number, and the record content includes: equipment basic information, key parameter original data and preprocessing results of each period, associated analysis report, health index and health level record, and health trend prediction result, wherein the equipment basic information includes model, factory date, and installation location.

[0044] Configure the automatic update logic of the record: when the sensing module completes the new period data acquisition and classification, trigger the screening module, the analysis module, the evaluation module and the output module to execute the processing procedure in turn, generate new health state data, and automatically update to the corresponding health record, and record the update timestamp.

[0045] On the other hand, a power transformation equipment evaluation method based on multi-dimensional parameters, comprising:

[0046] ​The operation electric parameters, physical state parameters and initial health characteristic parameters of the power transformation equipment are collected, after the abnormal value elimination, the self-adaptive weighted moving average denoising and the min-max normalization preprocessing, the classification is completed according to the equipment component-data type two-level standard; the candidate parameters strongly related to the equipment fault state are screened through the mutual information entropy, the key parameters are determined combined with the grey correlation analysis, the health influence dimension is labeled, and the initial weight of each dimension is calculated by using the analytic hierarchy process; the long short-term memory network dynamic correlation model is constructed by using the time sequence data of the key parameters, the deviation degree of the actual value and the model prediction value of the parameters is calculated, the abnormal correlation characteristics are identified, and the correlation analysis report is generated; the health grade evaluation logic is set, the health influence coefficient is determined according to the abnormal degree of the key parameters, the equipment health index is calculated by substituting the initial weight, and the health grade is divided according to the index range and the abnormal characteristics; the time sequence curve graph of the key parameters, the health dimension radar graph and the health grade column graph are generated, the future period health trend is predicted based on the time attenuation coefficient and the trend adjustment factor, and after being packaged, the encrypted protocol is transmitted to the preset receiving end; the health archives containing the basic information, the period data and the evaluation results are created according to the unique identification of the associated equipment, the archives are automatically updated after the new period data is collected, and the update time stamp is recorded.

[0047] Compared with the known prior art, the technical scheme provided by the present application has the following beneficial effects:

[0048] The present application provides a power transformation equipment evaluation method and system based on multi-dimensional parameters, which in the execution process, through collecting and classifying multiple types of operation and health data of power transformation equipment, accurately screening key parameters strongly related to equipment health, and labeling health influence dimensions and determining dimension weights, can focus on core evaluation elements and reduce irrelevant data interference.

[0049] In addition, by establishing a dynamic correlation model to identify parameter abnormal correlation characteristics, potential health problems of the equipment can be captured in time, health indexes are calculated based on abnormal characteristics and preset logic, health grades are divided, combined with the abnormal priority degradation principle of the safety dimension, the evaluation accuracy and safety priority are ensured, visual charts and health trend prediction can intuitively master the equipment state and development trend, and finally a unique health archive with a time stamp is created and automatically updated, realizing the whole cycle tracking and management of the equipment health state, improving the efficiency and accuracy of the power transformation equipment health evaluation as a whole, and helping the reliable operation of the equipment, reducing maintenance cost and safety risk. BRIEF DESCRIPTION OF DRAWINGS

[0050] In order to more clearly illustrate the technical scheme in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creating any inventive labor.

[0051] Figure 1 A structure schematic diagram of a power transformation equipment evaluation system based on multi-dimensional parameters;

[0052] Figure 2 A flow schematic diagram of a power transformation equipment evaluation method based on multi-dimensional parameters. DETAILED DESCRIPTION

[0053] To make the objects, technical solutions, and advantages of the embodiments of the present application clearer, the following will be combined with the accompanying drawings for the embodiments of the present application to make a clear and complete description of the technical solutions in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0054] The present application will be further described in combination with the embodiments.

[0055] Embodiment 1

[0056] A power transformation equipment evaluation system based on multi-dimensional parameters in the embodiment comprises: Figure 1 as shown in the figure, comprising:

[0057] a perception module, configured to collect operation information and initial health characterization data of the power transformation equipment, and to classify the collected information and data;

[0058] The perception module comprises a multi-source perception unit, a data preprocessing unit, and a classification unit.

[0059] The multi-source perception unit is configured with special sensors corresponding to different core components of the power transformation equipment, and is configured to collect operation electric parameters of the power transformation equipment, including voltage, current, power, physical state parameters including temperature, vibration, humidity, and initial health characterization parameters including insulation resistance and oil dissolved gas concentration.

[0060] The data preprocessing unit is configured to perform outlier rejection, data smoothing and denoising, and data normalization processing on the collected raw data, wherein the normalization processing adopts a min-max normalization method to map the data to the interval [0, 1].

[0061] The classification unit classifies the preprocessed data according to a two-level classification standard of equipment components-data types, wherein the equipment components include transformer body, switch device, and mutual inductor, and the data types include real-time collection data and periodic collection data.

[0062] Supplementary explanation:

[0063] The outlier rejection adopts a dynamic threshold method:

[0064] First, the mean and standard deviation of a certain type of parameter in the same collection cycle are calculated, and then the dynamic threshold is determined in combination with the current operating load rate of the device , represent the mean, standard deviation, represent the preset multiple, initially 1, when the data value is outside the range, it is determined to be an abnormal value and is removed, and after removal, linear interpolation of the valid values at adjacent times is used for completion;

[0065] Data smoothing and denoising uses adaptive weighted moving average: for the time series data after removing abnormal values, set the base window size, and according to the data change rate , represent the parameter value at time t, the window weight is dynamically adjusted, the greater the change rate, the higher the weight of recent data, and the expression is: wherein, , and subject to ;

[0066] The screening module is configured to receive information and data classified in the perception module, screen key parameters strongly related to the health status of the device from the classified information and data, and label the health influence dimensions corresponding to each parameter.

[0067] The process of the screening module performing key parameter screening and health influence dimension labeling is as follows:

[0068] Calculate the mutual information entropy of each parameter after classification and the historical fault data of the device, and screen out candidate parameters with mutual information entropy values greater than a preset threshold, wherein the mutual information entropy is used to represent the close degree of association between the parameter and the fault state of the device.

[0069] wherein the mutual information entropy value calculation logic is:

[0070] , represent the value set of the parameter to be evaluated, the set of device fault states, including normal, minor fault, and serious fault, represent the joint probability distribution of the parameter value x and the fault state y, represent the edge probability distribution of the parameter value x, and the edge probability distribution of the fault state y, The greater the value, the closer the association between the parameter and the fault state of the device.

[0071] The above formula quantifies the degree of association between the joint probability distribution and the edge probability distribution, and introduces the mutual information entropy in information theory into the screening of the parameters of the power transformation device, which can objectively capture the nonlinear association between the parameters and the faults. Compared with traditional linear correlation analysis, it can more comprehensively cover the fault association characteristics, provide a scientific and quantitative basis for preliminary screening of key parameters, and reduce subjective screening bias.

[0072] The correlation degrees of the candidate parameters with the health benchmark state of the equipment are calculated by grey correlation analysis, and the parameters with the top N correlation degrees are further selected as the key parameters;

[0073] The detailed process is as follows:

[0074] Let the health benchmark state sequence be (the sequence is composed of the equipment factory standard or historical optimal operation data), and the sequence of the i th candidate parameter be , n represents the number of data collection points; then the correlation coefficient is calculated as , , wherein the resolution coefficient is denoted as ∈(0, 1) is used to adjust the sensitivity of the correlation coefficient; and the correlation degree is quantitatively expressed as , and the correlation degree is closer to 1, indicating that the consistency of the candidate parameter with the health benchmark state is higher;

[0075] The above formula constructs the health benchmark state sequence based on the equipment factory standard or historical optimal operation data, introduces the resolution coefficient to calculate the correlation coefficient for the sequence of the i th candidate parameter, and quantifies the consistency of the parameter with the health benchmark through the correlation degree. This formula can effectively handle the scene of partial information uncertainty in the health assessment of the power transformation equipment, the flexible setting of the resolution coefficient can adapt to the analysis requirements of different parameters, avoid the screening deviation caused by a single standard, and accurately screen the candidate parameters with high consistency with the health benchmark of the equipment, thereby providing support for the determination of the key parameters;

[0076] The key parameters are labeled to the preset health influence dimensions, the health influence dimensions include the safety dimension, the reliability dimension, and the economic dimension, the safety dimension is related to the risk of safety accidents caused by equipment failure, the reliability dimension is related to the normal operation time of the equipment, and the economic dimension is related to the maintenance cost and life loss of the equipment. Based on the number and influence degree of the key parameters in each dimension, the initial weight of each health influence dimension is determined through the analytic hierarchy process;

[0077] Specifically,

[0078] a judgment matrix is constructed as , , wherein the importance ratio of the i th dimension relative to the j th dimension is denoted as > 0, = 1 / ; the maximum eigenvalue of the judgment matrix and the corresponding eigenvector are calculated as , the initial weight of each dimension is obtained after normalization of the eigenvector , j = 1, 2, 3; finally, consistency check is performed, the consistency index and the consistency ratio are calculated, and when <0.1, it is determined that the weight distribution meets the consistency requirement;

[0079] In the formula: is the maximum eigenvalue of the judgment matrix; is a preset average random consistency index;

[0080] The analysis module is configured to receive the screened key parameters with labels, establish a dynamic correlation model between the parameters, identify abnormal correlation characteristics of the parameters, and synchronously generate a correlation analysis report.

[0081] The operation of establishing the dynamic correlation model and identifying the abnormal correlation characteristics of the parameters in the analysis module is as follows:

[0082] The time series data of the screened key parameters are taken as input to construct a dynamic correlation model based on a long short-term memory network. The input layer dimension of the model is consistent with the number of key parameters, the hidden layer is set to 2-3 layers, the output layer is the predicted value of each key parameter, the adaptive moment estimation optimizer is used in the model training process, and the mean square error is used as the loss function.

[0083] The deviation degree of the actual value of the parameter and the predicted value of the model is calculated , represents the actual collection value of the i-th key parameter at time t, represents the model predicted value of the parameter at time t, represents the standard deviation of the historical normal operation data of the parameter.

[0084] When the deviation degree of any key parameter is greater than a preset deviation threshold, or the deviation degrees of two or more key parameters are simultaneously greater than 80% of the respective preset deviation thresholds, it is determined that the identified abnormal correlation characteristics of the parameters, and the abnormal parameter identifier, the abnormal occurrence time, and the deviation degree value are recorded in the correlation analysis report.

[0085] The evaluation module is configured to obtain the correlation analysis report, set a health level evaluation logic, and calculate a device health index based on the abnormal correlation characteristics of the parameters in the analysis report, and synchronously perform a health level division operation.

[0086] When the evaluation module calculates the device health index based on the abnormal correlation characteristics of the parameters, it is subject to:

[0087] Based on the abnormal correlation characteristics of each key parameter, a health influence coefficient of each key parameter is determined:

[0088] When the key parameter is normal, the health influence coefficient is 1; when the key parameter has a slight abnormality, the health influence coefficient is 0.8; when the key parameter has a moderate abnormality, the health influence coefficient is 0.5; and when the key parameter has a severe abnormality, the health influence coefficient is 0.2, wherein the abnormality degree is determined by the deviation degree determined by comparison with a preset threshold value respectively;

[0089] Recalculating the device health index:

[0090]

[0091] wherein: is the initial weight of the jth health impact dimension in the screening module obtained by AHP; is the number of key parameters under the jth health impact dimension; is the set of key parameters under the jth health impact dimension; is the health impact coefficient; is the normalized value of the ith key parameter, which is output by the data preprocessing unit of the perception module and is subject to ;

[0092] The above formula integrates the three-dimensional factors of dimension weight, parameter abnormality degree and parameter normalized value, which not only reflects the importance difference of different health dimensions, but also accurately quantifies the influence of single parameter abnormality on the overall health, so that the health index can comprehensively and selectively reflect the device health status and provide intuitive and quantitative basis for health grade division;

[0093] The health grade evaluation logic set in the evaluation module is:

[0094] According to the value range of the device health index and the severity of the parameter abnormality correlation characteristics, the device health grade is divided into: excellent, good, general, warning and failure;

[0095] The grade division logic is:

[0096] When ≥ 0.9 and there is no key parameter abnormality, the health grade is excellent; when 0.8 ≤ < 0.9 and only a single light abnormality parameter exists, the health grade is good; when 0.6 ≤ < 0.8 and there are no more than two light abnormality parameters or a single moderate abnormality parameter, the health grade is general; when 0.4 ≤ < 0.6 and there is a single severe abnormality parameter or two or more moderate abnormality parameters, the health grade is warning; when < 0.4 or there are two or more severe abnormality parameters, the health grade is failure;

[0097] When performing the health grade division operation, if there is a key parameter abnormality under the safety dimension, perform a grade reduction on the health grade, except when the health grade is failure, to prioritize the safety of device operation;

[0098] ​The output module receives the health level, the health index and the correlation analysis report in the evaluation module, generates a visual health state chart, synchronously predicts the health trend of the power transformation equipment, and transmits the generated visual health state chart and the prediction result to a preset receiving end in a preset format through a preset encryption transmission protocol.

[0099] The operation content of the output module in the stage of generating the visual health state chart and the prediction of the health trend of the power transformation equipment is as follows:

[0100] The generated visual health state chart includes:

[0101] A key parameter time sequence change curve diagram for showing the change trend of each key parameter with time and an abnormal point mark;

[0102] A health influence dimension radar chart taking the health sub-index of each dimension as a radius for showing the health state of each dimension, wherein the health sub-index is a comprehensive index under each health influence dimension, that is, the health sub-index of the jth health influence dimension , which constitutes a hierarchical correlation with ;

[0103] A health level change bar chart for the health level and the health index change in the last N periods;

[0104] The health trend prediction value of the pth preset period in the future is quantified as:

[0105] ;

[0106] In the formula: is the preset historical period number participating in the prediction; is a time attenuation coefficient, ∈(0, 1); is the health index of the t-kth period; is a trend adjustment factor, > 0; is the average change rate of the health index in the last m periods, representing the historical change trend of the health state;

[0107] The above formula dynamically adapts the device health fluctuation characteristics through the time attenuation coefficient, captures the steepness of the historical change trend through the trend adjustment factor, breaks the limitations of fixed time weight and ignoring change trend in traditional prediction, makes the prediction value highlight the influence of the recent health state and fit the overall trend of the device health change, and significantly improves the adaptability and accuracy of the prediction of the health trend of the power transformation equipment;

[0108] Finally, the visual health state chart, the health trend prediction value of the pth period in the future and the corresponding health level prediction interval are packaged in a preset format and transmitted to a preset receiving end through a preset encryption transmission protocol;

[0109] wherein, The greater the value of the device health state fluctuation is, The greater the value is, on the contrary The smaller the value is; The greater the value of the device health index historical change trend is, The greater the value is, on the contrary The smaller the value is;

[0110] The management module is used for receiving the visual health state chart and the prediction result to create a unique health record of the power transformation equipment with a time stamp;

[0111] The management module creates a unique health record of the power transformation equipment, and the process is as follows:

[0112] The health record is associated with a unique device identifier, which represents the device number, and the record content includes: device basic information, key parameter original data and preprocessing result of each cycle, associated analysis report, health index and health grade record, and health trend prediction result, wherein the device basic information includes model, factory date, and installation location.

[0113] The configuration record automatically updates the logic: when the sensing module completes the new cycle data acquisition and classification, the screening module, the analysis module, the evaluation module and the output module are triggered to execute the processing flow in turn to generate new health state data, and automatically update to the corresponding health record, and record the update time stamp.

[0114] The sensing module is connected to the screening module through wireless network interaction, the screening module is connected to the analysis module and the evaluation module through wireless network interaction, and the analysis module is connected to the output module and the management module through wireless network interaction.

[0115] In the embodiment, the perception module runs to collect operation information and initial health characterization data of the power transformation equipment, classifies the collected information and data, and the screening module runs to receive the classified information and data in the perception module, screens key parameters strongly related to the health state of the equipment from the classified information and data, labels the health influence dimensions corresponding to each parameter, and then the analysis module receives the screened key parameters with labels, establishes a dynamic correlation model between the parameters to identify abnormal correlation characteristics of the parameters, synchronously generates a correlation analysis report, the evaluation module further acquires the correlation analysis report, sets health level evaluation logic, and calculates the equipment health index based on the abnormal correlation characteristics of the parameters in the analysis report, synchronously performs health level division operation, and the output module receives the health level, health index and correlation analysis report in the evaluation module, generates a visual health state chart, synchronously predicts the health trend of the power transformation equipment, and then packages the generated visual health state chart and prediction result to transmit to a preset receiving end, and finally the management module receives the visual health state chart and prediction result to create a unique health file of the power transformation equipment with a time stamp.

[0116] The system in the above embodiment can accurately collect and optimize the operation and health data of the power transformation equipment, accurately identify key influencing factors and abnormal correlations, scientifically evaluate the health state of the equipment, intuitively present the state changes and predict the trends, and can also establish an equipment-specific health file, avoid failure risks in advance, ensure the safe operation of the equipment, prolong the service life of the equipment, reduce maintenance costs, provide clear data support for operation and maintenance, and effectively improve the operation and maintenance efficiency and management level of the power transformation equipment.

[0117] Embodiment 2

[0118] In the specific implementation level, based on embodiment 1, the embodiment refers to Figure 2 The embodiment 1 is further specifically described as a power transformation equipment evaluation system based on multi-dimensional parameters.

[0119] A power transformation equipment evaluation method based on multi-dimensional parameters, comprising:

[0120] Collecting operation electric parameters, physical state parameters and initial health characterization parameters of the power transformation equipment, performing abnormal value elimination, adaptive weighted moving average denoising and min-max normalization preprocessing, and then classifying according to the equipment component-data type two-level standard;

[0121] Screening candidate parameters strongly related to the equipment failure state by mutual information entropy, determining key parameters by combining gray correlation analysis, labeling the health influence dimensions, and calculating the initial weights of each dimension by using analytic hierarchy process;

[0122] A long short-term memory network dynamic correlation model is constructed by using key parameter time sequence data, a deviation degree of the actual value and the model predicted value of the parameter is calculated, abnormal correlation characteristics are identified, and a correlation analysis report is generated;

[0123] A health level evaluation logic is set, a health influence coefficient is determined according to the abnormal degree of the key parameter, a device health index is calculated by substituting the initial weight, and the health level is divided according to the index range and abnormal characteristics;

[0124] A key parameter time sequence curve chart, a health dimension radar chart and a health level column chart are generated, a future period health trend is predicted based on a time attenuation coefficient and a trend adjustment factor, and after being packaged, the health trend is transmitted to a preset receiving end through an encryption protocol;

[0125] A health record containing basic information, periodic data and evaluation results is generated based on the unique identification of the associated device, and the record is automatically updated by the process triggered after new periodic data is collected, and the update timestamp is recorded.

[0126] In summary, in the above-mentioned embodiments, the method and system can accurately screen the key parameters that are strongly related to the health of the device by collecting and classifying multiple types of operation and health data of the power transformation device, and can focus on the core evaluation elements by labeling the health influence dimensions and determining the dimension weights, thereby reducing irrelevant data interference. In addition, by establishing a dynamic correlation model to identify abnormal correlation characteristics of the parameters, potential health problems of the device can be captured in a timely manner. Based on abnormal characteristics and preset logic, the health index is calculated and the health level is divided, and the safety dimension abnormality priority degradation principle is combined to ensure that the evaluation is accurate and safety is prioritized. Visualization charts and health trend prediction can help to intuitively understand the state and development trend of the device. Finally, a unique health record with a timestamp is created and automatically updated to realize whole-cycle tracking and management of the health state of the device, thereby improving the efficiency and accuracy of the health evaluation of the power transformation device and helping to ensure reliable operation of the device, reduce maintenance costs and safety risks.

[0127] The above embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can modify the technical solutions described in the foregoing embodiments, or make equivalent substitutions for part of the technical features; and these modifications or substitutions will not change the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A multi-dimensional parameter based electrical equipment evaluation system, comprising: The method comprises the following steps: The perception module is used to collect the operation information and initial health characterization data of the power transformation equipment, and to classify the collected information and data; The screening module is used to receive the classified information and data in the perception module, screen the key parameters strongly related to the health state of the equipment from the classified information and data, and label the health influence dimensions corresponding to each parameter; The analysis module is used to receive the key parameters screened with labels, establish a dynamic correlation model between the parameters, identify the abnormal correlation characteristics of the parameters, and synchronously generate a correlation analysis report; The evaluation module is used to obtain the correlation analysis report, set up a health level evaluation logic, and calculate the equipment health index based on the abnormal correlation characteristics of the parameters in the analysis report, and synchronously perform a health level division operation; The output module is used to receive the health level, health index and correlation analysis report in the evaluation module, generate a visual health state chart, synchronously predict the health trend of the power transformation equipment, and package the generated visual health state chart and prediction results for transmission to a preset receiving end; The management module is used to receive the visual health state chart and prediction results to create a unique health file of the power transformation equipment with a time stamp.

2. The multi-dimensional parameter-based substation evaluation system of claim 1, wherein, The perception module comprises a multi-source perception unit, a data preprocessing unit and a classification unit; The multi-source perception unit is configured with special sensors corresponding to different core components of the power transformation equipment, and is used to collect the operation electric parameters of the power transformation equipment, including voltage, current and power; physical state parameters, including temperature, vibration and humidity; and initial health characterization parameters, including insulation resistance and oil dissolved gas concentration; The data preprocessing unit is used to perform outlier rejection, data smoothing and denoising and data normalization processing on the collected raw data, wherein the normalization processing adopts a min-max normalization method to map the data to the [0, 1] interval; The classification unit classifies the preprocessed data according to a two-level classification standard of equipment components and data types, wherein the equipment components include transformer body, switch equipment and mutual inductor, and the data types include real-time collection data and periodic collection data.

3. The multi-dimensional parameter-based substation equipment evaluation system of claim 1, wherein, The screening module performs the following process of key parameter screening and health influence dimension labeling: Calculate the mutual information entropy of each parameter after classification and the historical fault data of the equipment, and screen out candidate parameters with mutual information entropy values greater than a preset threshold; Calculate the correlation degree of each parameter and the health benchmark state of the equipment through gray correlation analysis on the candidate parameters, and further screen out the top N parameters as key parameters according to the correlation degree; Label each key parameter to a preset health influence dimension, wherein the health influence dimension includes a safety dimension related to the risk of safety accidents caused by equipment failure, a reliability dimension related to the length of time of normal operation of the equipment, and an economic dimension related to the maintenance cost and life loss of the equipment, and determine the initial weight of each health influence dimension based on the number and influence degree of the key parameters in each dimension through analytic hierarchy process.

4. The multi-dimensional parameter based substation equipment evaluation system of claim 1, wherein, The operation of establishing a dynamic correlation model and identifying abnormal correlation characteristics of parameters in the analysis module is as follows: The dynamic correlation model based on the long short-term memory network is constructed by taking the time sequence data of the screened key parameters as input, the dimension of the input layer of the model is consistent with the number of key parameters, the hidden layer is set to 2-3 layers, the output layer is the predicted value of each key parameter, the adaptive moment estimation optimizer is used in the model training process, and the mean square error is used as the loss function; deviation degree of the actual value of the parameter and the predicted value of the model , denotes the actual collected value of the i-th key parameter at time t, denotes the model predicted value of the parameter at time t, denotes the standard deviation of the historical normal operation data of the parameter; When the deviation degree of any key parameter is greater than the preset deviation threshold, or the deviation degrees of two or more key parameters are greater than 80% of the respective preset deviation thresholds at the same time, it is determined that the identified parameter abnormal correlation feature is abnormal, and the abnormal parameter identification, the abnormal occurrence time and the deviation degree value are recorded in the correlation analysis report.

5. The multi-dimensional parameter based substation equipment evaluation system of claim 1, wherein, When the evaluation module calculates the equipment health index based on the parameter abnormal correlation feature, it is subject to: Based on the abnormal correlation feature of each key parameter, the health influence coefficient of each key parameter is determined: When the key parameter is normal, the health influence coefficient is 1; when the key parameter has slight abnormality, the health influence coefficient is 0.8; when the key parameter has moderate abnormality, the health influence coefficient is 0.5; and when the key parameter has severe abnormality, the health influence coefficient is 0.2, wherein the abnormality degree is determined by the deviation degree is determined by comparing with the respectively preset threshold value. The equipment health index is recalculated: ; In the formula: is the initial weight of the jth health impact dimension obtained by the analytic hierarchy in the screening module; is the number of key parameters under the jth health impact dimension; is the set of key parameters under the jth health impact dimension; is the health impact coefficient; is the normalized value of the ith key parameter, which is output by the data preprocessing unit of the perception module and is subject to .

6. The multi-dimensional parameter based substation equipment evaluation system of claim 5, wherein, The health level evaluation logic set in the evaluation module is: According to the value range of the equipment health index and the severity of the parameter abnormal correlation feature, the equipment health level is divided into: excellent, good, general, warning and failure; The grade division logic is: When a health rating of excellent when ≥ 0.9 with no critical parameter abnormalities; a health rating of good when 0.8 ≤ < 0.9 with only a single minor abnormal parameter; a health rating of fair when 0.6 ≤ < 0.8 with no more than two minor abnormal parameters or a single moderate abnormal parameter; When 0.4≤ When <0.6 and there is a single severe abnormality parameter or two or more moderate abnormality parameters, the health level is warning; when When <0.4 or there are two or more severe abnormality parameters, the health level is failure.

7. The multi-dimensional parameter based substation equipment evaluation system of claim 1, wherein, The operation content of the output module for generating a visual health state chart and predicting the health trend stage of the power transformation equipment is: The generated visual health state chart includes: A key parameter time sequence change curve diagram for showing the change trend of each key parameter with time and abnormal point marking; a health impact dimension radar chart, with a health sub-index of each dimension as a radius, used to show the health status of each dimension, wherein the health sub-index is a comprehensive index under each health impact dimension, i.e., the health sub-index of the jth health impact dimension ; A health level change bar chart for the health level and health index change in the last N cycles; Quantitative health trend prediction value of the pth preset period in the future: ; In the formula: is the preset number of historical periods participating in prediction; is the time attenuation coefficient, ∈(0, 1); is the health index of the t-kth period; is the trend adjustment factor, > 0; represents: the average change rate of the health index of the last m periods, representing the historical change trend of the health state; Finally, the visualization will include health status charts and health trend predictions for the next p periods. The corresponding health level prediction ranges are packaged according to a preset format and transmitted to a preset receiving end through a preset encrypted transmission protocol.

8. The multi-dimensional parameter based substation equipment evaluation system of claim 1, wherein, The management module creates a unique health file of the power transformation equipment, and the process is: The health file is associated with the unique identification of the equipment, which includes the equipment number, and the file content includes: equipment basic information, key parameter original data and preprocessing results of each cycle, correlation analysis report, health index and health level record, and health trend prediction result, wherein the equipment basic information includes model, factory date and installation location; The automatic updating logic of the file is configured: after the sensing module completes the new cycle data acquisition and classification, the screening module, the analysis module, the evaluation module and the output module are triggered to execute the processing process in turn, a new health state data is generated, and the health state data is automatically updated to the corresponding health file, and the update time stamp is recorded.

9. The multi-dimensional parameter based substation equipment evaluation system of claim 1, wherein, The sensing module is connected with the screening module through wireless network interaction, the screening module is connected with the analysis module and the evaluation module through wireless network interaction, and the analysis module is connected with the output module and the management module through wireless network interaction.

10. A method for evaluating power transformation equipment based on multi-dimensional parameters, which is a method for implementing the system for evaluating power transformation equipment based on multi-dimensional parameters according to any one of claims 1-9, characterized in that, It includes: The running electrical parameters, physical state parameters and initial health characteristic parameters of the power transformation equipment are collected, preprocessed through abnormal value elimination, adaptive weighted moving average denoising and min-max normalization, and classified according to the equipment component-data type two-level standard; The candidate parameters strongly related to the equipment failure state are screened through mutual information entropy, the key parameters are determined combined with gray correlation analysis, the health influence dimensions are marked, and the initial weights of each dimension are calculated by using analytic hierarchy process; A long short-term memory network dynamic correlation model is constructed with key parameter time series data, a deviation degree of the actual value of the parameter and the predicted value of the model is calculated, abnormal correlation characteristics are identified, and a correlation analysis report is generated; A health level evaluation logic is set, a health influence coefficient is determined according to the abnormal degree of the key parameter, a device health index is calculated by substituting the initial weight, and the health level is divided according to the index range and abnormal characteristics; A key parameter time series curve graph, a health dimension radar graph and a health level column graph are generated, a future period health trend is predicted based on a time attenuation coefficient and a trend adjustment factor, and after packaging, the trend is transmitted to a preset receiving end through an encryption protocol; A health record containing basic information, periodic data and evaluation results is created with the unique identification of the associated device, and the record is automatically updated after new periodic data is collected, and the update timestamp is recorded.

Citation Information

Patent Citations

  • A Health Assessment System for Power Transmission and Transformation Equipment Based on Big Data

    CN109711664B

Cited By

  • Steel-concrete nondestructive testing method and device based on time-varying analysis

    CN121955193A