Relay protection state maintenance studying and judging system based on data analysis

Abstract

The invention discloses a relay protection state maintenance research and judgment system based on data analysis, and relates to the technical field of electric power system protection equipment. The system collects operation parameters in real time through a sensor group and monitoring equipment, preprocesses the operation parameters, and determines the relay protection state maintenance state based on a preprocessed data set; calculating a current harmonic distortion rate THD, a power stability index VSI and a vibration acceleration coefficient ZDJ, performing associated calculation to obtain an equipment health state comprehensive index HCI, and performing primary evaluation with a preset first equipment health threshold M and a preset second equipment health threshold N; and constructing a time sequence model by using an autoregressive integral moving average model (ARIMA), predicting a future data sequence, calculating a fault risk prediction index FRP in combination with a time sequence model prediction formula YCWL and HCI, and performing secondary comparison evaluation to generate maintenance research and judgment information. Through the system, the state of the relay protection equipment can be monitored in real time, early warning can be performed in time, and safe operation of the equipment is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of power system protection equipment, and in particular to a relay protection status inspection and analysis system based on data analysis. Background Art

[0002] With the continuous development of modern power systems and the continued growth of power loads, the safe and stable operation of power systems faces unprecedented challenges. Relay protection equipment, as an important component of the power system, shoulders the heavy responsibility of protecting the safe operation of the power system. However, traditional relay protection equipment fault detection methods often rely on manual experience and regular maintenance, which is not only time-consuming and labor-intensive, but also difficult to detect potential equipment faults in real time and accurately. Therefore, developing a data analysis-based relay protection status inspection and judgment system to monitor the operating status of relay protection equipment in real time and predict its possible faults has important practical significance and broad application prospects.

[0003] Existing fault detection methods for relay protection equipment suffer from several key shortcomings. First, traditional periodic maintenance methods make it difficult to promptly identify potential issues in equipment operation, which can lead to increased suddenness and severity of faults. Second, fault diagnosis methods that rely on manual experience are limited by personnel's expertise and accumulated experience, making it difficult to ensure the accuracy and consistency of diagnostic results. Furthermore, existing monitoring methods often lack in-depth mining and analysis of equipment operating data, failing to comprehensively and systematically reflect the health status and fault trends of the equipment, resulting in low accuracy and reliability in fault prediction. Summary of the Invention

[0004] In view of the deficiencies in the prior art, the present invention provides a relay protection status inspection and judgment system based on data analysis, which solves the problems in the above-mentioned background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a relay protection status inspection and judgment system based on data analysis, including a data acquisition module, a relay equipment analysis module, a comprehensive algorithm module, a time series module and an inspection and judgment module;

[0006] The data acquisition module is used to set detection points around the relay protection equipment and install sensor groups and monitoring equipment to collect operating parameters of the relay protection equipment in real time, and pre-process the collected operating parameters to generate an operating data set;

[0007] The relay equipment analysis module calculates and obtains the current harmonic distortion rate THD, the power stability index VSI and the vibration acceleration coefficient ZDJ based on the operation data set obtained after preprocessing;

[0008] The comprehensive algorithm module is used to perform correlation calculations on the obtained current harmonic distortion rate THD, power stability index VSI and vibration acceleration coefficient ZDJ to obtain the equipment health status comprehensive index HCI, and perform an initial comparative evaluation with the preset first equipment health threshold M and second equipment health threshold N, and initiate a secondary comparative evaluation based on the evaluation results;

[0009] The time series module is used to construct a time series model through the autoregressive integrated moving average model ARIMA, and split the training set and the test set through the cross-validation method, determine the autoregressive term p, the number of differences d and the moving average term q of the time series model by drawing the partial autocorrelation function PACF line graph and the autocorrelation function ACF line graph, use the maximum likelihood estimation method MLE to estimate the parameters of the time series model, and construct the time series model prediction formula YCWL;

[0010] The maintenance analysis module is used to predict the future data sequence based on the time series model and associate it with the equipment health status comprehensive index HCI, obtain the fault risk prediction index FRP through statistical calculation, perform a secondary comparative evaluation of the fault risk prediction index FRP and the preset fault risk threshold Z, and generate maintenance analysis information based on the evaluation results.

[0011] Preferably, the data acquisition module includes a data acquisition unit and a data preprocessing unit;

[0012] The data acquisition unit is used to collect operating parameters of the relay protection device in real time through a sensor group and a monitoring device, wherein the sensor group includes a current sensor, a voltage sensor and a vibration sensor; the monitoring device includes a harmonic analyzer, a frequency meter, a load monitor, a reactive power meter, a phase meter and a resonance frequency meter;

[0013] The current amplitude df, harmonic amplitude xf and fundamental frequency amplitude jf are obtained by a current sensor;

[0014] The harmonic frequency xp and harmonic amplitude xf are obtained by a harmonic analyzer;

[0015] The fundamental frequency jp and harmonic frequency xp are obtained by a frequency meter;

[0016] The load change rate lv is obtained through a load monitor;

[0017] The reactive power loss rp is obtained by a reactive power meter;

[0018] The voltage fluctuation rate vf is obtained by a voltage sensor;

[0019] The voltage phase difference vp is obtained by a phase meter;

[0020] The vibration displacement zw, vibration frequency zp and vibration amplitude zf are obtained by a vibration sensor;

[0021] The resonant frequency fr of the device is obtained by a resonant frequency measuring instrument;

[0022] The data preprocessing unit is used to perform denoising, data cleaning, smoothing and outlier monitoring preprocessing on the collected operating parameters of the relay protection equipment using sliding average and exponential smoothing techniques, and summarize and classify the preprocessed data to obtain current harmonic data sets, power quality data sets and vibration data sets;

[0023] The current harmonic data set includes current amplitude df, harmonic amplitude xf, fundamental frequency amplitude jf, harmonic frequency xp, and fundamental frequency jp;

[0024] The power quality data set includes load change rate lv, reactive power loss rp, voltage fluctuation rate vf, and voltage phase difference vp;

[0025] The vibration data set includes vibration displacement zw, vibration frequency zp, vibration amplitude zf, and resonance frequency fr of the device.

[0026] Preferably, the relay device analysis module includes a current harmonic distortion calculation unit, a power stability calculation unit and a vibration acceleration calculation unit;

[0027] The current harmonic distortion rate calculation unit is used to perform dimensionless calculation based on the preprocessed current harmonic data set and then calculate and obtain the current harmonic distortion rate THD;

[0028] The current harmonic distortion rate THD is obtained by the following formula:

[0029]

[0030] Where xf h It represents the amplitude of the hth harmonic, and n represents the number of harmonics.

[0031] Preferably, the power stability calculation unit is used to perform dimensionless processing on the pre-processed power quality data set, and then calculate and obtain the power stability index VSI;

[0032] The power stability index VSI is obtained by the following formula:

[0033]

[0034] Where V nom Indicates rated voltage, Q nom Indicates rated reactive power, P nom represents rated power, and cos represents cosine function.

[0035] Preferably, the vibration acceleration calculation unit is used to calculate and obtain the vibration acceleration coefficient ZDJ after dimensionless processing based on the pre-processed vibration data set;

[0036] The vibration acceleration coefficient ZDJ is obtained by the following formula:

[0037]

[0038] Where k represents the damping coefficient of the equipment.

[0039] Preferably, the comprehensive algorithm module includes a comprehensive calculation unit and a comprehensive analysis unit;

[0040] The comprehensive analysis unit is used to normalize the obtained current harmonic distortion rate THD, power stability index VSI and vibration acceleration coefficient ZDJ, and then calculate and obtain the equipment health status comprehensive index HCI;

[0041] The equipment health comprehensive index HCI is obtained by the following formula:

[0042]

[0043] Wherein, k1, k2 and k3 represent the weight coefficients of current harmonic distortion rate THD, vibration acceleration coefficient ZDJ and power stability index VSI respectively, k1+k2+k3=1, and 0<k1<0.33, 0<k2<0.34, 0<k3<0.33.

[0044] Preferably, the comprehensive analysis unit is used to preset the first device health threshold M and the second device health threshold N and the device health status comprehensive index HCI obtained by the comprehensive calculation unit, perform an initial comparative evaluation, and generate optimization information based on the evaluation results. The specific evaluation scheme is as follows;

[0045] When the device health status comprehensive index HCI is less than the preset first device health threshold M, it means that the device is in normal operation, and the second evaluation module is triggered;

[0046] When the second equipment health threshold N> equipment health status comprehensive index HCI ≥ preset first equipment health threshold M, it indicates that the equipment operation status is abnormal. At this time, a yellow warning is generated and the equipment is marked as the second priority for attention and maintenance;

[0047] When the equipment health status comprehensive index HCI ≥ the second equipment health threshold N, it means that the equipment operation status is in a dangerous state. At this time, a red alert is generated. At this time, a stop operation instruction is sent to the relay protection device controller, notifying the relevant staff to immediately stop the operation of the equipment, immediately pay attention to and repair the equipment, and mark this equipment as the first priority for attention and maintenance.

[0048] Preferably, the time series module includes a model building unit and a data prediction unit;

[0049] The model building unit constructs a time series model using an autoregressive integrated moving average (ARIMA) model; divides a historical operating data set into a training set and a test set, uses a time series cross-validation method to split the data into multiple training set and test set combinations by time, gradually trains and tests the model, and learns patterns and features in the data through training parameter learning and model fitting; inputs feature data of the test set into the trained model, and predicts operating data for future time periods based on the patterns and features of the historical operating data to obtain a predicted data set;

[0050] The time series model uses an Excel drawing tool, with the timestamp and date fields as the x-axis and the predicted data set as the y-axis, to draw a partial autocorrelation function PACF line graph and an autocorrelation function ACF line graph; by observing the partial autocorrelation function PACF line graph, the position where the significant value becomes zero or close to zero after the lag period is found, and this position determines the number p of the autoregressive term;

[0051] Use stationarity detection and difference processing to perform a difference processing on the time data, and use ADF test to check whether the sequence after difference is stationary. If it is not stationary, continue to perform secondary difference processing and repeat the steps until the sequence is stationary. Determine the number of differences d;

[0052] By observing the autocorrelation function (ACF) graph, find the position where the value after the lag period becomes zero or close to zero. This position determines the number of sliding average terms q.

[0053] The data prediction unit is used to analyze the autocorrelation function ACF and the partial autocorrelation function PACF graph to determine the number of autoregressive terms p, the number of difference times d and the number of sliding average terms q, and then use the maximum likelihood estimation method MLE to estimate the parameters of the time series model by finding the parameter values ​​that make the observed data under the model;

[0054] The time series model prediction formula YCWL is constructed as follows;

[0055]

[0056] Where, α represents the correction constant; is an autoregressive term, representing the current value Xt The linear relationship between the value of the past p moments, p is the order of the autoregressive term, which indicates how many past moments of data are used. is the autoregressive coefficient, which indicates the value of the past time ti relative to the current value X t the extent of the impact; is the moving average term, indicating the current value X t The linear relationship between the error and the past q moments, q is the order of the moving average term, indicating how many past moments of error are used, θ j is the moving average coefficient, which represents the error of the past time tj to the current value X t The degree of influence of t-j Represents the white noise error at the past time tj; ∈ t Represents the white noise error at the current time t.

[0057] Preferably, the maintenance analysis module includes a risk prediction calculation unit and an assessment and early warning unit;

[0058] The risk prediction calculation unit is used to perform dimensionless processing on the time series model prediction formula YCWL and the equipment health status comprehensive index HCI obtained by the comprehensive calculation unit, and then perform correlation calculation to obtain the fault risk prediction index FRP;

[0059] The fault risk prediction index FRP is obtained by the following formula:

[0060]

[0061] In the formula, RI t Indicates the risk value at the current moment, HCI t represents the comprehensive index of equipment health status at the current moment, and γ represents the influence coefficient of the comprehensive index of equipment health status HCI on risk prediction.

[0062] Preferably, when the evaluation and early warning unit initially measures that the relay protection device is in normal operation, it starts a second evaluation mechanism. The second evaluation mechanism performs a secondary comparative evaluation by using a preset fault risk threshold Z and the obtained fault risk prediction index FRP, and generates corresponding optimization information based on the relevant evaluation results. The specific evaluation scheme is as follows;

[0063] When the FRP ≥ the preset FRP threshold Z, it indicates that the equipment has a potential fault. In this case, the equipment is marked as the third priority for attention and maintenance, and the relevant staff are notified to perform maintenance and inspection on the equipment.

[0064] When the fault risk prediction index FRP is less than the preset fault risk threshold Z, it means that the overall operation of the equipment is normal and it maintains normal operation. This equipment does not need attention and it should be maintained and serviced normally.

[0065] The present invention provides a relay protection status inspection and judgment system based on data analysis. It has the following beneficial effects:

[0066] (1) The system integrates five modules: data acquisition module, equipment analysis module, comprehensive algorithm module, time series module, and maintenance analysis module, to achieve real-time monitoring and intelligent evaluation of the operating status of relay protection equipment. The data acquisition module uses a sensor group and monitoring equipment to accurately obtain operating parameter data and pre-process it to generate an operating data set. The equipment analysis module performs a multi-dimensional analysis of the equipment's operating status by calculating the current harmonic distortion rate THD, the power stability index VSI, and the vibration acceleration coefficient ZDJ. This method not only improves the accuracy and reliability of the data, but also makes the equipment health status assessment more comprehensive.

[0067] (2) The system uses a comprehensive algorithm module to combine the three indicators of current harmonic distortion rate THD, power stability index VSI and vibration acceleration coefficient ZDJ to calculate the equipment health status comprehensive index HCI and conduct an initial comparative evaluation. The time series module uses the autoregressive integral moving average model ARIMA to predict future data trends, and combines the equipment health status comprehensive index HCI to calculate the fault risk prediction index FRP and conduct a secondary evaluation. Through this multi-level and multi-stage evaluation mechanism, the system can promptly identify potential risks of equipment and generate maintenance and analysis information. This method not only effectively improves the accuracy of equipment fault prediction, but also greatly reduces the downtime caused by equipment failure, and improves the operational reliability and safety of the equipment.

[0068] (3) Through data analysis and intelligent algorithms, the system achieves real-time monitoring and prediction of equipment operating status, significantly improving maintenance efficiency and accuracy. In addition, the system uses a variety of sensors and monitoring equipment to improve data collection accuracy. Combined with advanced time series analysis models, it makes the prediction of equipment failure more accurate and timely, thus achieving all-round monitoring of equipment operating status and preventive maintenance, and improving the scientific and intelligent level of equipment management. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 The present invention is a flowchart of a relay protection status inspection and judgment system based on data analysis. DETAILED DESCRIPTION

[0070] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0071] Example 1

[0072] See also Figure 1 The present invention provides a relay protection status inspection and judgment system based on data analysis. To achieve the above purpose, the present invention is implemented through the following technical solutions: including a data acquisition module, a relay equipment analysis module, a comprehensive algorithm module, a time series module and an inspection and judgment module;

[0073] The data acquisition module is used to set detection points around the relay protection equipment and install sensor groups and monitoring equipment to collect operating parameters of the relay protection equipment in real time, and pre-process the collected operating parameters to generate an operating data set;

[0074] The relay equipment analysis module calculates and obtains the current harmonic distortion rate THD, the power stability index VSI and the vibration acceleration coefficient ZDJ based on the operation data set obtained after preprocessing;

[0075] The comprehensive algorithm module is used to perform correlation calculations on the obtained current harmonic distortion rate THD, power stability index VSI and vibration acceleration coefficient ZDJ to obtain the equipment health status comprehensive index HCI, and perform an initial comparative evaluation with the preset first equipment health threshold M and second equipment health threshold N, and initiate a secondary comparative evaluation based on the evaluation results;

[0076] The time series module is used to construct a time series model through the autoregressive integrated moving average model ARIMA, and split the training set and the test set through the cross-validation method, determine the autoregressive term p, the number of differences d and the moving average term q of the time series model by drawing the partial autocorrelation function PACF line graph and the autocorrelation function ACF line graph, use the maximum likelihood estimation method MLE to estimate the parameters of the time series model, and construct the time series model prediction formula YCWL;

[0077] The maintenance analysis module is used to predict the future data sequence based on the time series model and associate it with the equipment health status comprehensive index HCI, obtain the fault risk prediction index FRP through statistical calculation, perform a secondary comparative evaluation of the fault risk prediction index FRP and the preset fault risk threshold Z, and generate maintenance analysis information based on the evaluation results.

[0078] In this embodiment, the data acquisition module implements real-time monitoring of relay protection equipment operating parameters by installing detection equipment and an integrated sensor array. It also preprocesses the collected data, providing a high-quality data foundation for subsequent analysis. The relay equipment analysis module uses this preprocessed data to calculate the current harmonic distortion (THD), power stability index (VSI), and vibration acceleration coefficient (ZDJ) to comprehensively assess the equipment's health status. The comprehensive algorithm module integrates these indicators into the comprehensive equipment health index (HCI) and performs an evaluation, further enhancing the accuracy of fault detection. The time series module uses the autoregressive integrated moving average (ARIMA) model to predict future data and, combined with the comprehensive equipment health index (HCI), calculates the fault risk prediction index (FRP), achieving forward-looking fault prediction. Finally, the maintenance assessment module generates maintenance assessment information based on the FRP, providing timely notification of equipment maintenance needs. This systematic analysis and assessment approach not only improves the real-time and accuracy of equipment monitoring but also optimizes maintenance decisions, enhancing the operational safety and reliability of the equipment.

[0079] Example 2

[0080] This embodiment is explained in Example 1, please refer to Figure 1 ,Specifically: the data acquisition module includes a data acquisition unit and a data preprocessing unit;

[0081] The data acquisition unit is used to collect operating parameters of the relay protection device in real time through a sensor group and a monitoring device, wherein the sensor group includes a current sensor, a voltage sensor and a vibration sensor; the monitoring device includes a harmonic analyzer, a frequency meter, a load monitor, a reactive power meter, a phase meter and a resonance frequency meter;

[0082] The data preprocessing unit is used to perform denoising, data cleaning, smoothing and outlier monitoring preprocessing on the collected operating parameters of the relay protection equipment using sliding average and exponential smoothing techniques, and summarize and classify the preprocessed data to obtain current harmonic data sets, power quality data sets and vibration data sets;

[0083] The current harmonic data set includes current amplitude df, harmonic amplitude xf, fundamental frequency amplitude jf, harmonic frequency xp, and fundamental frequency jp;

[0084] The power quality data set includes load change rate lv, reactive power loss rp, voltage fluctuation rate vf, and voltage phase difference vp;

[0085] The vibration data set includes vibration displacement zw, vibration frequency zp, vibration amplitude zf, and resonance frequency fr of the device.

[0086] In this embodiment, the data acquisition unit, through a sensor array and monitoring equipment, enables the system to comprehensively and accurately collect equipment operating parameters in real time. The data preprocessing unit effectively removes noise and cleans data using sliding average and exponential smoothing techniques, ensuring data accuracy and reliability. The preprocessed data is summarized and classified into current harmonics datasets, power quality datasets, and vibration datasets, providing a clear and reliable data foundation for subsequent equipment analysis. This precise data acquisition and processing method significantly improves the accuracy of equipment status monitoring and the timeliness of fault detection, contributing to more effective equipment management and maintenance.

[0087] Example 3

[0088] This embodiment is explained in Example 2, please refer to Figure 1 ,Specifically: the relay equipment analysis module includes a current harmonic distortion rate calculation unit, a power stability calculation unit and a vibration acceleration calculation unit;

[0089] The current harmonic distortion rate calculation unit is used to perform dimensionless calculation based on the preprocessed current harmonic data set and then calculate and obtain the current harmonic distortion rate THD;

[0090] The current harmonic distortion rate THD is obtained by the following formula:

[0091]

[0092] Where xf h It represents the amplitude of the hth harmonic, and n represents the number of harmonics.

[0093] The power stability calculation unit is used to perform dimensionless processing on the pre-processed power quality data set, and then calculate and obtain the power stability index VSI;

[0094] The power stability index VSI is obtained by the following formula:

[0095]

[0096] Where V nom Indicates rated voltage, Q nom Indicates rated reactive power, P nom represents rated power, and cos represents cosine function.

[0097] The vibration acceleration calculation unit is used to calculate the vibration acceleration coefficient ZDJ after dimensionless processing based on the pre-processed vibration data set;

[0098] The vibration acceleration coefficient ZDJ is obtained by the following formula:

[0099]

[0100] Where k represents the damping coefficient of the equipment.

[0101] In this embodiment, the relay equipment analysis module, through its various calculation units, effectively converts complex equipment operating data into key health indicators, thereby improving the accuracy and reliability of equipment status monitoring. The current harmonic distortion rate calculation unit performs dimensionless processing on the preprocessed current harmonic data set to generate a current harmonic distortion rate (THD) indicator that accurately reflects the harmonic distortion of the equipment current. The power stability calculation unit calculates the power stability index (VSI) from the preprocessed power quality data through dimensionless processing. This index can effectively assess the stability of the power system. The vibration acceleration calculation unit calculates the vibration acceleration coefficient (ZDJ) based on the vibration data set to reflect the mechanical vibration status of the equipment. The comprehensive application of these key health indicators greatly improves the health status of relay protection equipment, providing strong support for the timely detection of potential faults and the implementation of preventive measures, thereby extending equipment life, reducing downtime, and improving the safety and reliability of equipment operation.

[0102] Example 4

[0103] This embodiment is explained in Example 3, please refer to Figure 1 ,Specifically: the comprehensive algorithm module includes a comprehensive calculation unit and a comprehensive analysis unit;

[0104] The comprehensive analysis unit is used to normalize the obtained current harmonic distortion rate THD, power stability index VSI and vibration acceleration coefficient ZDJ, and then calculate and obtain the equipment health status comprehensive index HCI;

[0105] The equipment health comprehensive index HCI is obtained by the following formula:

[0106]

[0107] Wherein, k1, k2 and k3 represent the weight coefficients of current harmonic distortion rate THD, vibration acceleration coefficient ZDJ and power stability index VSI respectively, k1+k2+k3=1, and 0<k1<0.33, 0<k2<0.34, 0<k3<0.33.

[0108] The comprehensive analysis unit is used to preset the first device health threshold M and the second device health threshold N and the device health status comprehensive index HCI obtained by the comprehensive calculation unit, perform an initial comparative evaluation, and generate optimization information based on the evaluation results. The specific evaluation scheme is as follows;

[0109] When the device health status comprehensive index HCI is less than the preset first device health threshold M, it means that the device is in normal operation, and the second evaluation module is triggered;

[0110] When the second equipment health threshold N> equipment health status comprehensive index HCI ≥ preset first equipment health threshold M, it indicates that the equipment operation status is abnormal. At this time, a yellow warning is generated and the equipment is marked as the second priority for attention and maintenance;

[0111] When the equipment health status comprehensive index HCI ≥ the second equipment health threshold N, it means that the equipment operation status is in a dangerous state. At this time, a red warning is generated. At this time, a stop operation instruction is sent to the relay protection device controller, and the equipment is immediately paid attention to and repaired. This equipment is marked as the first priority for attention and maintenance.

[0112] In this embodiment, by normalizing the current harmonic distortion rate THD, the power stability index VSI, and the vibration acceleration coefficient ZDJ, and calculating the comprehensive index of equipment health status HCI, the overall health status of the equipment can be comprehensively reflected. Using the preset first equipment health threshold M and second equipment health threshold N, the comprehensive index of equipment health status HCI is initially compared and evaluated. The system can accurately identify the operating status of the equipment, divide it into three levels: normal, abnormal, and dangerous, and generate corresponding early warning information. In the normal state, further evaluation is triggered; in the abnormal state, a yellow early warning is generated, marked as the second priority and arranged for maintenance; in the dangerous state, a red early warning is generated, a stop operation instruction is issued, and immediate maintenance is arranged. This system can improve the accuracy and timeliness of equipment maintenance, reduce the failure rate, extend the life of the equipment, and improve overall operating efficiency.

[0113] Example 5

[0114] This embodiment is explained in Example 1, please refer to Figure 1 ,Specifically: the time series module includes a model building unit and a data prediction unit;

[0115] The model building unit constructs a time series model using an autoregressive integrated moving average (ARIMA) model; divides a historical operating data set into a training set and a test set, uses a time series cross-validation method to split the data into multiple training set and test set combinations by time, gradually trains and tests the model, and learns patterns and features in the data through training parameter learning and model fitting; inputs feature data of the test set into the trained model, and predicts operating data for future time periods based on the patterns and features of the historical operating data to obtain a predicted data set;

[0116] The time series model uses an Excel drawing tool, with the timestamp and date fields as the x-axis and the predicted data set as the y-axis, to draw a partial autocorrelation function PACF line graph and an autocorrelation function ACF line graph; by observing the partial autocorrelation function PACF line graph, the position where the significant value becomes zero or close to zero after the lag period is found, and this position determines the number p of the autoregressive term;

[0117] Use stationarity detection and difference processing to perform a difference processing on the time data, and use ADF test to check whether the sequence after difference is stationary. If it is not stationary, continue to perform secondary difference processing and repeat the steps until the sequence is stationary. Determine the number of differences d;

[0118] By observing the autocorrelation function (ACF) graph, find the position where the value after the lag period becomes zero or close to zero. This position determines the number of sliding average terms q.

[0119] The data prediction unit is used to analyze the autocorrelation function ACF and the partial autocorrelation function PACF graph to determine the number of autoregressive terms p, the number of difference times d and the number of sliding average terms q, and then use the maximum likelihood estimation method MLE to estimate the parameters of the time series model by finding the parameter values ​​that make the observed data under the model;

[0120] The time series model prediction formula YCWL is constructed as follows;

[0121]

[0122] Where, α represents the correction constant; is an autoregressive term, representing the current value X t The linear relationship between the value of the past p moments, p is the order of the autoregressive term, which indicates how many past moments of data are used. is the autoregressive coefficient, which indicates the value of the past time ti relative to the current value X t the extent of the impact; is the moving average term, indicating the current value X t The linear relationship between the error and the past q moments, q is the order of the moving average term, indicating how many past moments of error are used, θ j is the moving average coefficient, which represents the error of the past time tj to the current value X t The degree of influence of t-j Represents the white noise error at the past time tj; ∈ t Represents the white noise error at the current time t.

[0123] In this embodiment, the model building unit uses the autoregressive integrated moving average model (ARIMA) to build a time series model. By dividing the historical operating data set into multiple training sets and test sets, the model is gradually trained and tested to learn the patterns and features in the data. After the model is trained, the characteristic data of the test set is input to predict the operating data of the future time period. The partial autocorrelation function PACF and autocorrelation function ACF line graphs drawn by the Excel drawing tool help determine the autoregressive term p, the number of differences d, and the moving average term q, and the maximum likelihood estimation method MLE is used for parameter optimization; the time series model is constructed through precise mathematical formulas and can capture the time dependence and pattern characteristics of the operating data of the relay protection equipment. Through accurate prediction of future data, the time series module provides strong support for the prediction of the equipment operating status and fault prevention, thereby improving the initiative of equipment management, reducing the risk of sudden failures, extending the service life of the equipment, and optimizing the equipment maintenance and management strategy to ensure the stable and reliable operation of the power system.

[0124] Example 6

[0125] This embodiment is explained in Example 5, please refer to Figure 1 ,Specifically: the maintenance analysis module includes a risk prediction and calculation unit and an ,assessment and early warning unit;

[0126] The risk prediction calculation unit is used to perform dimensionless processing on the time series model prediction formula YCWL and the equipment health status comprehensive index HCI obtained by the comprehensive calculation unit, and then perform correlation calculation to obtain the fault risk prediction index FRP;

[0127] The fault risk prediction index FRP is obtained by the following formula:

[0128]

[0129] In the formula, RI t Indicates the risk value at the current moment, HCI t represents the comprehensive index of equipment health status at the current moment, and γ represents the influence coefficient of the comprehensive index of equipment health status HCI on risk prediction.

[0130] When the evaluation and early warning unit initially measures that the relay protection device is in normal operation, it starts the second evaluation mechanism. The second evaluation mechanism performs a secondary comparative evaluation by using the preset fault risk threshold Z and the obtained fault risk prediction index FRP, and generates corresponding optimization information based on the relevant evaluation results. The specific evaluation scheme is as follows;

[0131] When the FRP ≥ the preset FRP threshold Z, it indicates that the equipment has a potential fault. In this case, the equipment is marked as the third priority for attention and maintenance, and the relevant staff are notified to perform maintenance and inspection on the equipment.

[0132] When the fault risk prediction index FRP is less than the preset fault risk threshold Z, it means that the overall operation of the equipment is normal and it maintains normal operation. This equipment does not need attention and it should be maintained and serviced normally.

[0133] In this embodiment, the risk prediction calculation unit performs dimensionless processing on the time series model prediction formula YCWL and the comprehensive index of equipment health status HCI, and obtains the fault risk prediction index FRP through correlation calculation, which can accurately predict the failure risk of the equipment in future operation. The evaluation and warning unit realizes a detailed assessment of the equipment status by performing a secondary comparative evaluation of the fault risk prediction index FRP and the preset fault risk threshold Z. When the fault risk prediction index FRP exceeds the preset fault risk threshold Z, the system marks the equipment as the third priority and notifies the relevant staff to perform maintenance and inspection to prevent potential failures. When the fault risk prediction index FRP is lower than the preset fault risk threshold Z, the equipment is considered to be operating normally and maintains normal maintenance and care. This mechanism not only improves the accuracy and timeliness of fault warnings, reduces downtime and repair costs caused by equipment failures, but also effectively optimizes the equipment maintenance and management strategy through an intelligent evaluation mechanism, ensuring the long-term stable operation of relay protection equipment.

[0134] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A relay protection status inspection and judgment system based on data analysis, characterized by: It includes data acquisition module, relay equipment analysis module, comprehensive algorithm module, time series module and maintenance analysis module; The data acquisition module is used to set detection points around the relay protection equipment and install sensor groups and monitoring equipment to collect operating parameters of the relay protection equipment in real time, and pre-process the collected operating parameters to generate an operating data set; The relay equipment analysis module calculates and obtains the current harmonic distortion rate THD, the power stability index VSI and the vibration acceleration coefficient ZDJ based on the operation data set obtained after preprocessing; The comprehensive algorithm module is used to perform correlation calculations on the obtained current harmonic distortion rate THD, power stability index VSI and vibration acceleration coefficient ZDJ to obtain the equipment health status comprehensive index HCI, and perform an initial comparative evaluation with the preset first equipment health threshold M and second equipment health threshold N, and initiate a secondary comparative evaluation based on the evaluation results; The time series module is used to construct a time series model through the autoregressive integrated moving average model ARIMA, and split the training set and the test set through the cross-validation method, determine the autoregressive term p, the number of differences d and the moving average term q of the time series model by drawing the partial autocorrelation function PACF line graph and the autocorrelation function ACF line graph, use the maximum likelihood estimation method MLE to estimate the parameters of the time series model, and construct the time series model prediction formula YCWL; The maintenance analysis module is used to associate the time series model prediction formula YCWL constructed according to the time series model with the equipment health status comprehensive index HCI, obtain the fault risk prediction index FRP through statistical calculation, perform a secondary comparative evaluation with the fault risk prediction index FRP through a preset fault risk threshold Z, obtain the evaluation result, and generate maintenance analysis information based on the evaluation result.

2. The data analysis-based relay protection status inspection and judgment system according to claim 1 is characterized in that: The data acquisition module includes a data acquisition unit and a data preprocessing unit; The data acquisition unit is used to collect operating parameters of the relay protection device in real time through a sensor group and a monitoring device, wherein the sensor group includes a current sensor, a voltage sensor and a vibration sensor; the monitoring device includes a harmonic analyzer, a frequency meter, a load monitor, a reactive power meter, a phase meter and a resonance frequency meter; The data preprocessing unit is used to perform denoising, data cleaning, smoothing and outlier monitoring preprocessing on the collected operating parameters of the relay protection equipment using sliding average and exponential smoothing techniques, and summarize and classify the preprocessed data to obtain current harmonic data sets, power quality data sets and vibration data sets; The current harmonic data set includes current amplitude df, harmonic amplitude xf, fundamental frequency amplitude jf, harmonic frequency xp, and fundamental frequency jp; The power quality data set includes load change rate lv, reactive power loss rp, voltage fluctuation rate vf, and voltage phase difference vp; The vibration data set includes vibration displacement zw, vibration frequency zp, vibration amplitude zf, and resonance frequency fr of the device.

3. The relay protection status inspection and judgment system based on data analysis according to claim 2 is characterized in that: The relay equipment analysis module includes a current harmonic distortion calculation unit, a power stability calculation unit and a vibration acceleration calculation unit; The current harmonic distortion rate calculation unit is used to perform dimensionless calculation based on the preprocessed current harmonic data set and then calculate and obtain the current harmonic distortion rate THD; The current harmonic distortion rate THD is obtained by the following formula: Where xf h It represents the amplitude of the hth harmonic, and n represents the number of harmonics.

4. The data analysis-based relay protection status inspection and judgment system according to claim 3 is characterized by: The power stability calculation unit is used to perform dimensionless processing on the pre-processed power quality data set, and then calculate and obtain the power stability index VSI; The power stability index VSI is obtained by the following formula: Where V nom Indicates rated voltage, Q nom Indicates rated reactive power, P nom represents rated power, and cos represents cosine function.

5. The data analysis-based relay protection status inspection and judgment system according to claim 3 is characterized by: The vibration acceleration calculation unit is used to calculate the vibration acceleration coefficient ZDJ after dimensionless processing based on the pre-processed vibration data set; The vibration acceleration coefficient ZDJ is obtained by the following formula: Where k represents the damping coefficient of the equipment.

6. The relay protection status inspection and judgment system based on data analysis according to claim 1 is characterized in that: The comprehensive algorithm module includes a comprehensive calculation unit and a comprehensive analysis unit; The comprehensive analysis unit is used to normalize the obtained current harmonic distortion rate THD, power stability index VSI and vibration acceleration coefficient ZDJ, and then calculate and obtain the equipment health status comprehensive index HCI; The equipment health comprehensive index HCI is obtained by the following formula: Wherein, k1, k2 and k3 represent the weight coefficients of current harmonic distortion rate THD, vibration acceleration coefficient ZDJ and power stability index VSI respectively, k1+k2+k3=1, and 0<k1<0.33, 0<k2<0.34, 0<k3<0.

33.

7. The data analysis-based relay protection status inspection and judgment system according to claim 6 is characterized in that: The comprehensive analysis unit is used to preset the first device health threshold M and the second device health threshold N and the device health status comprehensive index HCI obtained by the comprehensive calculation unit, perform an initial comparative evaluation, and generate optimization information based on the evaluation results. The specific evaluation scheme is as follows; When the device health status comprehensive index HCI is less than the preset first device health threshold M, it means that the device is in normal operation, and the second evaluation module is triggered; When the second equipment health threshold N> equipment health status comprehensive index HCI ≥ preset first equipment health threshold M, it indicates that the equipment operation status is abnormal. At this time, a yellow warning is generated and the equipment is marked as the second priority for attention and maintenance; When the equipment health status comprehensive index HCI ≥ the second equipment health threshold N, it means that the equipment operation status is in a dangerous state. At this time, a red warning is generated. At this time, a stop operation instruction is sent to the relay protection device controller, and the equipment is immediately paid attention to and repaired. This equipment is marked as the first priority for attention and maintenance.

8. The data analysis-based relay protection status inspection and judgment system according to claim 1 is characterized by: The time series module includes a model building unit and a data prediction unit; The model building unit constructs a time series model using an autoregressive integrated moving average (ARIMA) model; divides a historical operating data set into a training set and a test set, uses a time series cross-validation method to split the data into multiple training set and test set combinations by time, gradually trains and tests the model, and learns patterns and features in the data through training parameter learning and model fitting; inputs feature data of the test set into the trained model, and predicts operating data for future time periods based on the patterns and features of the historical operating data to obtain a predicted data set; The time series model uses an Excel drawing tool, with the timestamp and date fields as the x-axis and the predicted data set as the y-axis, to draw a partial autocorrelation function PACF line graph and an autocorrelation function ACF line graph; by observing the partial autocorrelation function PACF line graph, the position where the significant value becomes zero or close to zero after the lag period is found, and this position determines the number p of the autoregressive term; Use stationarity detection and difference processing to perform a difference processing on the time data, and use ADF test to check whether the sequence after difference is stationary. If it is not stationary, continue to perform secondary difference processing and repeat the steps until the sequence is stationary. Determine the number of differences d; By observing the autocorrelation function (ACF) graph, find the position where the value after the lag period becomes zero or close to zero. This position determines the number of sliding average terms q. The data prediction unit is used to analyze the autocorrelation function ACF and the partial autocorrelation function PACF graph to determine the number of autoregressive terms p, the number of difference times d and the number of sliding average terms q, and then use the maximum likelihood estimation method MLE to estimate the parameters of the time series model by finding the parameter values ​​that make the observed data under the model; The time series model prediction formula YCWL is constructed as follows; Where, α represents the correction constant; is an autoregressive term, representing the current value X t The linear relationship between the value of the past p moments, p is the order of the autoregressive term, which indicates how many past moments of data are used. is the autoregressive coefficient, which indicates the value of the past time ti relative to the current value X t the extent of the impact; is the moving average term, indicating the current value X t The linear relationship between the error and the past q moments, q is the order of the moving average term, indicating how many past moments of error are used, θ j is the moving average coefficient, which represents the error of the past time tj to the current value X t The degree of influence of t-j Represents the white noise error at the past time tj; ∈ t Represents the white noise error at the current time t.

9. The data analysis-based relay protection status inspection and judgment system according to claim 8 is characterized in that: The maintenance analysis module includes a risk prediction and calculation unit and an assessment and early warning unit; The risk prediction calculation unit is used to perform dimensionless processing on the time series model prediction formula YCWL and the equipment health status comprehensive index HCI obtained by the comprehensive calculation unit, and then perform correlation calculation to obtain the fault risk prediction index FRP; The fault risk prediction index FRP is obtained by the following formula: In the formula, RI t Indicates the risk value at the current moment, HCI t represents the comprehensive index of equipment health status at the current moment, and γ represents the influence coefficient of the comprehensive index of equipment health status HCI on risk prediction.

10. The data analysis-based relay protection status inspection and judgment system according to claim 9 is characterized in that: When the evaluation and early warning unit initially measures that the relay protection device is in normal operation, it starts the second evaluation mechanism. The second evaluation mechanism performs a secondary comparative evaluation by using the preset fault risk threshold Z and the obtained fault risk prediction index FRP, and generates corresponding optimization information based on the relevant evaluation results. The specific evaluation scheme is as follows; When the FRP ≥ the preset FRP threshold Z, it indicates that the equipment has a potential fault. In this case, the equipment is marked as the third priority for attention and maintenance, and the relevant staff are notified to perform maintenance and inspection on the equipment. When the fault risk prediction index FRP is less than the preset fault risk threshold Z, it means that the overall operation of the equipment is normal and it maintains normal operation. This equipment does not need attention and it should be maintained and serviced normally.

Images