Power equipment operation and maintenance monitoring method and system based on data analysis

Abstract

The invention discloses a power equipment operation and maintenance monitoring method and system based on data analysis, relates to the technical field of power data analysis, and comprehensively knows the fault interaction and propagation trend between equipment by calculating the fault propagation state S of the power equipment. Through comprehensive analysis of the distance D between the power equipment and the fault propagation state S, the fault propagation range R is calculated, the equipment range possibly influenced by the fault is determined, and data support is further provided for adjusting the operation state of the power equipment. Finally, according to the analysis result of the fault propagation state S and the influence range R, the equipment state Z needing to be adjusted is judged, and through a triggering mechanism of a self-adaptive matching power supply scheduling strategy, the intelligence and accuracy of operation and maintenance management of the power equipment are remarkably improved, the limitation that correlation and propagation influence between the equipment cannot be comprehensively considered in a traditional method is overcome, and the method is suitable for popularization and application. Therefore, the complex fault propagation problem of a modern power grid can be better solved.

Description

Technical Field

[0001] The present invention relates to the technical field of power data analysis, and particularly to a power equipment operation and maintenance monitoring method and system based on data analysis. Background Technique

[0002] With the rapid development of intelligent technologies, data analysis has gradually become an indispensable part of the power industry. In the power system, the stability and operation efficiency of equipment are directly related to the energy supply security of society. Power equipment operation and maintenance monitoring is precisely an important link to ensure the stable operation of the system. This field covers a wide range of technologies, including monitoring, fault diagnosis, predictive maintenance, and many other aspects.

[0003] In this process, real-time data collection and analysis have become the core means. Specifically, power equipment operation and maintenance monitoring involves a wide variety of equipment, including transformers, switchgear, cables, and transmission lines in substations. These equipment need to monitor their health status in real time during daily operation. Especially in a complex power network, the failure of each equipment may trigger a chain reaction, thereby affecting the operation stability of the entire power grid.

[0004] Existing power equipment monitoring methods mostly focus on the real-time monitoring of single equipment and local fault warning. Traditional monitoring systems often rely on simple sensor data collection and fixed threshold alarm mechanisms. Although this method can capture the anomalies of individual equipment in a timely manner, it is difficult to deeply analyze the interaction between different equipment and the global effects of fault propagation. For example, when a transformer fails, traditional methods may only warn of the anomaly of the equipment, while ignoring whether the fault will cause abnormal loads or systematic failures of other related equipment. Moreover, existing methods usually do not fully consider spatio-temporal factors, such as the time, location of equipment failures, and changes in surrounding loads, which leads to relatively lagged fault prediction and risk assessment and cannot take effective preventive measures before the occurrence of faults. Although some advanced algorithms have begun to attempt multi-equipment fault diagnosis, most systems still cannot achieve global spatio-temporal analysis and lack the ability to predict fault propagation paths. Summary of the Invention

[0005] Aiming at the deficiencies of the existing technology, the present invention provides a power equipment operation and maintenance monitoring method and system based on data analysis, which solves the problems mentioned in the background technique.

[0006] To achieve the above objectives, the present invention is realized through the following technical solutions: A power equipment operation and maintenance monitoring method based on data analysis, comprising the following steps:

[0007] S1. Collect the real-time operation status data of the device through the sensor and device monitoring system, and at the same time collect the spatio-temporal information of the device. After preprocessing the real-time operation status data and spatio-temporal information, form the spatio-temporal feature vector F;

[0008] S2. Extract features from the spatio-temporal feature vector F to obtain the feature vectors associated with the occurrence of faults, and form the fault prediction feature vector Ffail;

[0009] S3. Based on the obtained spatio-temporal feature vector F and fault prediction feature vector Ffail, and combined with the historical fault data and network connection relationship stored in the power equipment, establish the propagation matrix M between power equipment, and obtain the fault propagation state S of the power equipment;

[0010] S4. Based on the spatio-temporal information in the spatio-temporal feature vector F, obtain the distance D between power equipment, and calculate the propagation range of the fault with the obtained fault propagation state S of the power equipment to obtain the fault propagation influence range R;

[0011] S5. Analyze according to the obtained fault propagation state S and fault propagation influence range R, obtain the power equipment state Z that needs to be adjusted, and regulate the trigger mechanism of the power supply dispatching strategy adaptively matched to the power equipment state Z.

[0012] Preferably, S1 includes S11 and S12;

[0013] S11. Through the sensors and monitoring system of the device, collect the real-time operation status data and spatio-temporal information of the device in real time. The real-time operation status data includes the load L, power P, temperature W, and vibration V data of the power equipment, and the spatio-temporal information includes the position coordinates (x, y) of the power equipment;

[0014] S12. Preprocess the real-time operation status data. The preprocessing includes filling in missing data and standardization processing to obtain the preprocessed standard load CL, standard power CP, standard temperature CW, and standard vibration CV, and integrate them with the spatio-temporal information to obtain the spatio-temporal feature vector F.

[0015] Preferably, S2 includes S21 and S22;

[0016] S21. Extract features from the spatio-temporal feature vector F to obtain the feature vectors associated with the occurrence of faults, including the load-fault correlation coefficient RLF, temperature-fault correlation coefficient RWF, vibration-fault correlation coefficient RVF, and power-fault correlation coefficient RPF;

[0017] S22. Recombine through the load-fault correlation coefficient RLF, temperature-fault correlation coefficient RWF, vibration-fault correlation coefficient RVF, and power-fault correlation coefficient RPF with the position coordinates (x, y) and timestamp t in the spatio-temporal feature vector F to obtain the fault prediction feature vector Ffail.

[0018] Preferably, through the load-fault correlation coefficient RLF, temperature-fault correlation coefficient RWF, vibration-fault correlation coefficient RVF, and power-fault correlation coefficient RPF, the correlation degrees between the standard load CL, standard power CP, standard temperature CW, and standard vibration CV and the fault label F can be understood.

[0019] Among them, the correlation degrees are as follows:

[0020] When the load-fault correlation coefficient RLF, temperature-fault correlation coefficient RWF, vibration-fault correlation coefficient RVF, and power-fault correlation coefficient RPF = 1, it indicates a positive correlation, meaning that when the standard load CL, standard power CP, standard temperature CW, and standard vibration CV increase, the probability of a fault occurring also increases, and there is a strict linear relationship between the two.

[0021] When the load-fault correlation coefficient RLF, temperature-fault correlation coefficient RWF, vibration-fault correlation coefficient RVF, and power-fault correlation coefficient RPF = 0, it indicates no linear relationship, meaning that when the standard load CL, standard power CP, standard temperature CW, and standard vibration CV change, the probability of a fault occurring does not change, and the changes of the two are not correlated.

[0022] When the load-fault correlation coefficient RLF, temperature-fault correlation coefficient RWF, vibration-fault correlation coefficient RVF, and power-fault correlation coefficient RPF = -1, it indicates a negative correlation, meaning that when the standard load CL, standard power CP, standard temperature CW, and standard vibration CV increase, the probability of a fault occurring decreases, and there is a strict negative linear relationship between the two.

[0023] Preferably, the load-fault correlation coefficient RLF is obtained through the following calculation formula:

[0024] ;

[0025] In the formula, CL(p) represents the standard load of the p-th data point, F(p) represents the fault label of the power equipment at the p-th data point, μ(F) represents the mean of the fault label data, reflecting the average probability of a fault occurring during the sampling time period, μ(CL) represents the mean of the standard load data, specifically representing the average value of all standard loads in the standard load sequence, and n represents the total number of data points, specifically representing the number of samples of the power equipment participating in the calculation.

[0026] The temperature and fault correlation coefficient RWF is obtained through the following calculation formula:

[0027] ;

[0028] In the formula, CW(p) represents the standard temperature of the p-th data point, μ(CW) represents the mean value of the standard temperature data, specifically representing the average value of all standard loads in the standard temperature sequence;

[0029] The vibration and fault correlation coefficient RVF is obtained through the following calculation formula:

[0030] ;

[0031] In the formula, CV(p) represents the standard vibration of the p-th data point, μ(CV) represents the mean value of the standard vibration data, specifically representing the average value of all standard loads in the standard vibration sequence;

[0032] The power and fault correlation coefficient RPF is obtained through the following calculation formula:

[0033] ;

[0034] In the formula, CP(p) represents the standard power of the p-th data point, μ(CP) represents the mean value of the standard power data, specifically representing the average value of all standard loads in the standard power sequence.

[0035] Preferably, S3 includes S31 and S32;

[0036] S31, based on the obtained spatio-temporal feature vector F, combines the historical fault data stored in the power equipment and the connection relationships between various power equipment;

[0037] Among them, the historical fault data includes the fault occurrence time, type and influence range; the connection relationships include the power transmission connection between power equipment, the physical connection between power equipment, and the data exchange connection between power equipment;

[0038] Based on the historical fault data, the connection relationships between power equipment, and the spatio-temporal feature vector F of the power equipment, analyze the influence of different power equipment on other power equipment when a fault occurs, including that a power equipment fault causes a power outage, which in turn affects other connected power equipment; establish a propagation matrix M based on the obtained influence. Each row and each column in the propagation matrix M represents a power equipment, and the element m(i, j) in the propagation matrix M represents the degree of influence of power equipment i on power equipment j.

[0039] Preferably, in S32, based on the propagation matrix M and the fault prediction feature vector Ffail, the fault propagation state S of each power device is analyzed, where the fault propagation state S is used to reflect that the fault of each power device is not only determined by the fault probability of the power device itself, but also related to the fault states of the adjacent devices of the power device and the propagation path between the power devices;

[0040] The fault propagation state S is obtained through the following calculation formula:

[0041] ;

[0042] In the formula, S(i, t) represents the fault propagation state of power device i at time t, α(i) represents the weighting coefficient of power device i, specifically representing the fault sensitivity of power device i, Ffail(i, t) represents the fault prediction feature vector of power device i at time t, w represents the total number of power devices, β(i, j) represents the fault influence weight of power device i on power device j, Ffail(j, t) represents the fault prediction feature vector of power device j at time t, λ(i, j) represents the attenuation factor of fault propagation, specifically representing the rate at which the propagation effect decays with the spatial distance and the passage of time when the fault propagates from power device i to power device j, △t represents the fault propagation time difference, and exp represents the exponential decay function.

[0043] Preferably, S4 includes S41;

[0044] S41. Based on the spatio-temporal information in the spatio-temporal feature vector F, the distance D between power devices is obtained, and the propagation range of the fault is calculated by combining with the obtained fault propagation state S of the power device to obtain the fault propagation influence range R;

[0045] Among them, the distance D is calculated and obtained through the position coordinates (x, y) in the spatio-temporal information of the spatio-temporal feature vector F;

[0046] The fault propagation influence range R is obtained through the following calculation formula:

[0047] ;

[0048] In the formula, R(i) represents the fault propagation influence range of power device i, D(i, j) represents the distance between power device i and power device j, and Dmax represents the upper limit influence distance of power device fault propagation, specifically used for normalizing the distance D(i, j) between power device i and power device j, for simulating the distance attenuation effect.

[0049] Preferably, S5 includes S51;

[0050] S51. Analyze according to the obtained fault propagation state S and fault propagation influence range R, obtain the power equipment state Z that needs to be adjusted, and regulate the triggering mechanism of the power supply scheduling strategy adaptively matched to the power equipment state Z;

[0051] The power equipment state Z is obtained through the following analysis method:

[0052] ;

[0053] Where Z(i, t) represents the power equipment state of power equipment i at time t, Sthe represents the preset fault propagation state trigger threshold, Rthe represents the preset fault propagation influence range trigger threshold, and Hthe represents the preset power equipment health index threshold;

[0054] When the power equipment state Z(i, t) of power equipment i at time t = 1, it means that power equipment i needs to be adjusted, triggers the matching power supply scheduling strategy of power equipment i, and at the same time prompts the relevant inspection and maintenance personnel that power equipment i needs maintenance, and generates a maintenance task to be sent to the pending task list of the relevant inspection and maintenance personnel for processing;

[0055] When the power equipment state Z(i, t) of power equipment i at time t = 0, it means that power equipment i does not need to be adjusted;

[0056] Among them, the matching power supply scheduling strategy is used to detect the number of power equipment that needs to be adjusted. When the total number of power equipment that needs to be adjusted reaches the preset distributed fault threshold, it prompts to perform fault separation on the faulty equipment.

[0057] A power equipment operation and maintenance monitoring system based on data analysis includes a power equipment data acquisition module, a fault prediction correlation module, a power equipment fault propagation module, a propagation influence range module, and a power equipment operation and maintenance decision-making module;

[0058] The power equipment data acquisition module collects the real-time operation status data of the equipment through sensors and equipment monitoring systems, and at the same time collects the spatio-temporal information of the equipment. After preprocessing the real-time operation status data and spatio-temporal information, it forms a spatio-temporal feature vector F;

[0059] The fault prediction correlation module extracts features from the spatio-temporal feature vector F to obtain the feature vector associated with the occurrence of the fault, and forms a fault prediction feature vector Ffail;

[0060] The power equipment fault propagation module is based on the obtained spatio-temporal feature vector F and fault prediction feature vector Ffail, and combines the historical fault data and network connection relationship stored in the power equipment to establish a propagation matrix M between the power equipment, and obtains the fault propagation state S of the power equipment;

[0061] The propagation influence range module obtains the distance D between power devices based on the spatio-temporal information in the spatio-temporal feature vector F, and calculates the propagation range of the fault by calculating with the obtained fault propagation state S of the power devices to obtain the fault propagation influence range R;

[0062] The power device operation and maintenance decision-making module analyzes based on the obtained fault propagation state S and fault propagation influence range R, obtains the power device state Z that needs to be adjusted, and controls the triggering mechanism of the power supply scheduling strategy adaptively matched to the power device state Z.

[0063] The present invention provides a power device operation and maintenance monitoring method and system based on data analysis, having the following beneficial effects:

[0064] (1) By calculating the fault propagation state S of the power devices, the fault interaction and propagation trend between the devices can be comprehensively understood. Through the comprehensive analysis of the distance D between the power devices and the fault propagation state S, the propagation range R of the fault is calculated, the device range that may be affected by the fault is clarified, and further data support is provided for adjusting the operation state of the power devices. Finally, according to the analysis results of the fault propagation state S and the influence range R, the device state Z that needs to be adjusted is determined, and by adaptively matching the triggering mechanism of the power supply scheduling strategy, the intelligence and accuracy of the power device operation and maintenance management are significantly improved, overcoming the limitations in the traditional method that cannot comprehensively consider the association and propagation influence between devices, so as to better cope with the complex fault propagation problems in the modern power grid.

[0065] (2) By establishing the propagation matrix M, the influence degree between each power device is quantified, and further reveals how a device failure affects other power devices through different ways such as power transmission, physical connection or data exchange. Each element m(i, j) in this propagation matrix M represents the influence of power device i on power device j, helping to identify potential fault chains. On this basis, by calculating the fault propagation state S, it is further revealed that the fault propagation of power devices is not only affected by the probability of its own failure, but also jointly affected by the association with neighboring devices and the propagation path. It can effectively prevent large-scale fault spread, improve the emergency response ability of power device operation, especially in the early stage of the fault, the affected range of the fault can be accurately predicted, and the stability of power device operation and maintenance is improved.

[0066] (3) By combining the spatio-temporal feature vector F and the fault propagation state S of power equipment, as well as the distance D between power equipment, the accurate calculation of the fault propagation influence range R is achieved. By normalizing the distance between equipment and applying an exponential decay function, the spatial decay effect of fault propagation can be effectively simulated, ensuring the accurate definition of the fault propagation influence range. Subsequently, based on the fault propagation state S and the influence range R, the state Z of the power equipment that needs to be adjusted can be intelligently analyzed, and the corresponding power equipment scheduling strategy can be triggered. This scheduling mechanism can contain the further spread of the fault by adjusting the equipment state at the early stage of fault spread. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 Schematic diagram of the steps of a power equipment operation and maintenance monitoring method based on data analysis according to the present invention;

[0068] Figure 2 Block diagram schematic diagram of a power equipment operation and maintenance monitoring system based on data analysis according to the present invention;

[0069] Figure 3 Schematic diagram of the fluctuation of the correlation degree; DETAILED DESCRIPTION OF THE EMBODIMENTS

[0070] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0071] Embodiment 1

[0072] The present invention provides a power equipment operation and maintenance monitoring method based on data analysis. Please refer to Figure 1 and includes the following steps:

[0073] S1. Collect the real-time operation status data of the equipment through sensors and equipment monitoring systems, and at the same time collect the spatio-temporal information of the equipment. After preprocessing the real-time operation status data and spatio-temporal information, form the spatio-temporal feature vector F;

[0074] S2. Extract features from the spatio-temporal feature vector F to obtain the feature vector associated with the occurrence of a fault, and form the fault prediction feature vector Ffail;

[0075] S3. Based on the obtained spatio-temporal feature vector F and fault prediction feature vector Ffail, and in combination with the historical fault data and network connection relationships stored in the power equipment, establish the propagation matrix M between the power equipment, and obtain the fault propagation state S of the power equipment;

[0076] S4. Based on the spatio-temporal information in the spatio-temporal feature vector F, obtain the distance D between power equipment, and calculate the propagation range of the fault by calculating with the obtained fault propagation state S of the power equipment, and obtain the fault propagation influence range R;

[0077] S5. Analyze according to the obtained fault propagation state S and fault propagation influence range R, obtain the power equipment state Z that needs to be adjusted, and regulate the triggering mechanism of the power supply scheduling strategy adaptively matching the power equipment state Z.

[0078] In this embodiment, by forming the fault prediction feature vector Ffail, the accuracy and sensitivity of fault prediction are improved. Then, based on the spatio-temporal feature vector F and the fault prediction feature vector Ffail, combined with historical fault data and the network connection relationship of power equipment, a fault propagation matrix M is constructed, and the fault propagation state S of the power equipment is calculated, so as to comprehensively understand the fault interaction and propagation trend between equipment. Through the comprehensive analysis of the distance D between power equipment and the fault propagation state S, the propagation range R of the fault is calculated, the range of equipment that may be affected by the fault is clarified, and further data support is provided for adjusting the operation state of power equipment. Finally, according to the analysis results of the fault propagation state S and the influence range R, the equipment state Z that needs to be adjusted is determined, and by adaptively matching the triggering mechanism of the power supply scheduling strategy, the intelligence and accuracy of the operation and maintenance management of power equipment are significantly improved, overcoming the limitations of the traditional method that cannot comprehensively consider the association between equipment and propagation influence, so as to better cope with the complex fault propagation problems in modern power grids.

[0079] Embodiment 2

[0080] This embodiment is an explanatory description based on Embodiment 1, please refer to Figure 1 and Figure 3 , specifically: S1 includes S11 and S12;

[0081] S11. Through the sensors and monitoring systems of the equipment, real-time collect the real-time operation state data and spatio-temporal information of the equipment. The real-time operation state data includes the load L, power P, temperature W and vibration V data of the power equipment, and the spatio-temporal information includes the position coordinates (x, y) of the power equipment;

[0082] S12. Preprocess the real-time operation state data. The preprocessing includes missing data filling and standardization processing, obtain the preprocessed standard load CL, standard power CP, standard temperature CW and standard vibration CV, and integrate them with the spatio-temporal information to obtain the spatio-temporal feature vector F.

[0083] S2 includes S21 and S22;

[0084] S21. Extract features from the spatio-temporal feature vector F to obtain a feature vector associated with the occurrence of a fault, including the load-fault correlation coefficient RLF, the temperature-fault correlation coefficient RWF, the vibration-fault correlation coefficient RVF, and the power-fault correlation coefficient RPF;

[0085] S22. Recombine the load-fault correlation coefficient RLF, the temperature-fault correlation coefficient RWF, the vibration-fault correlation coefficient RVF, and the power-fault correlation coefficient RPF with the position coordinates (x, y) and the timestamp t in the spatio-temporal feature vector F to obtain the fault prediction feature vector Ffail.

[0086] Through the load-fault correlation coefficient RLF, the temperature-fault correlation coefficient RWF, the vibration-fault correlation coefficient RVF, and the power-fault correlation coefficient RPF, the correlation degrees between the standard load CL, the standard power CP, the standard temperature CW, and the standard vibration CV and the fault marker F can be understood, as shown in Table 1;

[0087] Among them, the correlation degrees are as follows:

[0088] When the load-fault correlation coefficient RLF, the temperature-fault correlation coefficient RWF, the vibration-fault correlation coefficient RVF, and the power-fault correlation coefficient RPF = 1, it indicates a positive correlation, meaning that when the standard load CL, the standard power CP, the standard temperature CW, and the standard vibration CV increase, the probability of the fault occurrence also increases, and there is a strict linear relationship between the two;

[0089] When the load-fault correlation coefficient RLF, the temperature-fault correlation coefficient RWF, the vibration-fault correlation coefficient RVF, and the power-fault correlation coefficient RPF = 0, it indicates no linear relationship, meaning that when the standard load CL, the standard power CP, the standard temperature CW, and the standard vibration CV change, the probability of the fault occurrence will not change, and the changes of the two are not correlated;

[0090] When the load-fault correlation coefficient RLF, the temperature-fault correlation coefficient RWF, the vibration-fault correlation coefficient RVF, and the power-fault correlation coefficient RPF = -1, it indicates a negative correlation, meaning that when the standard load CL, the standard power CP, the standard temperature CW, and the standard vibration CV increase, the probability of the fault occurrence decreases, and there is a strict negative linear relationship between the two;

[0091] Table 1:

[0092] Correlation coefficient R (specifically: RLF, RWF, RVF, RPF) Load - fault correlation coefficient RLF Temperature - fault correlation coefficient RWF Vibration - fault correlation coefficient RVF Power - fault correlation coefficient RPF Positive correlation R = 1 When the standard load CL increases, the probability of failure increases When the standard temperature CW increases, the probability of failure increases When the standard vibration CV increases, the probability of failure increases When the standard power CP increases, the probability of failure increases No correlation R = 0 When the standard load CL changes, the probability of failure remains unchanged When the standard temperature CW changes, the probability of failure remains unchanged When the standard vibration CV changes, the probability of failure remains unchanged When the standard power CP changes, the probability of failure remains unchanged Negative correlation R = 0 When the standard load CL increases, the probability of failure decreases When the standard temperature CW increases, the probability of failure decreases When the standard vibration CV increases, the probability of failure decreases When the standard power CP increases, the probability of failure decreases

[0093] .

[0094] The load-fault correlation coefficient RLF is obtained through the following calculation formula:

[0095] ;

[0096] Wherein, CL(p) represents the standard load of the p-th data point, F(p) represents the fault mark of the power equipment at the p-th data point, μ(F) represents the mean value of the fault mark data, reflecting the average probability of faults occurring during the sampling time period, μ(CL) represents the mean value of the standard load data, specifically representing the average value of all standard loads in the standard load sequence, and n represents the total number of data points, specifically representing the number of samples of the power equipment participating in the calculation;

[0097] The temperature-fault correlation coefficient RWF is obtained through the following calculation formula:

[0098] ;

[0099] Wherein, CW(p) represents the standard temperature of the p-th data point, and μ(CW) represents the mean value of the standard temperature data, specifically representing the average value of all standard loads in the standard temperature sequence;

[0100] The vibration-fault correlation coefficient RVF is obtained through the following calculation formula:

[0101] ;

[0102] Wherein, CV(p) represents the standard vibration of the p-th data point, and μ(CV) represents the mean value of the standard vibration data, specifically representing the average value of all standard loads in the standard vibration sequence;

[0103] The power-fault correlation coefficient RPF is obtained through the following calculation formula:

[0104] ;

[0105] Wherein, CP(p) represents the standard power of the p-th data point, and μ(CP) represents the mean value of the standard power data, specifically representing the average value of all standard loads in the standard power sequence.

[0106] In this embodiment, through the preprocessing of real-time data and the integration of spatio-temporal information, a spatio-temporal feature vector F is formed, providing comprehensive basic data for subsequent fault prediction. During the feature extraction process, by calculating the load-fault correlation coefficient RLF, temperature-fault correlation coefficient RWF, vibration-fault correlation coefficient RVF, and power-fault correlation coefficient RPF, the linear relationship between each operating state and the occurrence of faults can be quantified, and the probability of fault occurrence is determined through the fault flag F(p). This process can accurately evaluate the relevance between different operating state parameters and faults, help early warning of potential fault risks of equipment, and provide an accurate basis for subsequent fault propagation prediction. Finally, combining these correlation coefficients, the formed fault prediction feature vector Ffail will further promote the intelligent operation and maintenance management of power equipment, achieve the accuracy and forward-looking of fault prediction, avoid the over-reliance on a single indicator and potential prediction blind spots in traditional operation and maintenance, thus improving the fault prevention ability of power equipment.

[0107] Embodiment 3

[0108] This embodiment is an explanatory description carried out in Embodiment 2. Please refer to Figure 1 , specifically: S3 includes S31 and S32;

[0109] S31. Based on the obtained spatio-temporal feature vector F, combine the historical fault data stored in the power equipment and the connection relationships between various power equipment;

[0110] Among them, the historical fault data includes the fault occurrence time, type, and influence range; the connection relationships include power transmission connections between power equipment, physical connections between power equipment, and data exchange connections between power equipment;

[0111] Based on the historical fault data, the connection relationships between power equipment, and the spatio-temporal feature vector F of the power equipment, analyze the influence of different power equipment on other power equipment when a fault occurs, including that a power equipment failure causes a power outage, which in turn affects other connected power equipment; establish a propagation matrix M based on the obtained influence. Each row and each column in the propagation matrix M represents a power equipment, and the element m(i, j) in the propagation matrix M represents the degree of influence of power equipment i on power equipment j. Among them, the propagation matrix M is mainly determined by analyzing the transmission paths between equipment and the patterns of historical faults.

[0112] S32. Based on the propagation matrix M and the fault prediction feature vector Ffail, analyze the fault propagation state S of each power equipment. Among them, the fault propagation state S is used to reflect the situation that the fault of each power equipment is not only determined by the fault probability of the power equipment itself, but also related to the fault states of the neighboring equipment of the power equipment and the propagation paths between power equipment;

[0113] The fault propagation state S is obtained through the following calculation formula:

[0114] ;

[0115] In the formula, S(i, t) represents the fault propagation state of power equipment i at time t, α(i) represents the weighted coefficient of power equipment i, specifically representing the fault sensitivity of power equipment i, which is usually set according to the importance of the equipment, fault tolerance, etc. Ffail(i, t) represents the fault prediction feature vector of power equipment i at time t, w represents the total number of power equipment, β(i, j) represents the fault influence weight of power equipment i on power equipment j, Ffail(j, t) represents the fault prediction feature vector of power equipment j at time t, λ(i, j) represents the attenuation factor of fault propagation, specifically representing the rate at which the propagation effect decays with the spatial distance and the passage of time when the fault propagates from power equipment i to power equipment j, △t represents the fault propagation time difference, and exp represents the exponential decay function.

[0116] In this embodiment, by comprehensively considering the spatio-temporal feature vector F of power equipment, historical fault data, and the connection relationship between equipment, the fault propagation path and influence degree between power equipment can be accurately analyzed, so as to achieve more accurate fault propagation prediction. First, by establishing a propagation matrix M, the influence degree between each power equipment is quantified, further revealing how a fault in one equipment affects other power equipment through different ways such as power transmission, physical connection, or data exchange. Each element m(i, j) in this propagation matrix M represents the influence of power equipment i on power equipment j, helping to identify potential fault chains. On this basis, by calculating the fault propagation state S, it is further revealed that the fault propagation of power equipment is not only affected by its own fault probability, but also jointly affected by the association with neighboring equipment and the propagation path. It can effectively prevent large-scale fault spread, improve the emergency response ability of power equipment operation, especially in the early stage of the fault, the affected range of the fault can be accurately predicted, and the stability of power equipment operation and maintenance can be improved.

[0117] Embodiment 4

[0118] This embodiment is an explanatory description based on Embodiment 3. Please refer to Figure 1 , specifically: S4 includes S41;

[0119] S41. Based on the spatio-temporal information in the spatio-temporal feature vector F, obtain the distance D between power equipment, and calculate the propagation range of the fault with the obtained fault propagation state S of the power equipment to obtain the fault propagation influence range R;

[0120] Among them, the distance D is calculated and obtained through the position coordinates (x, y) in the spatio-temporal information of the spatio-temporal feature vector F;

[0121] The fault propagation influence range R is obtained through the following calculation formula:

[0122] ;

[0123] In the formula, R(i) represents the fault propagation influence range of power equipment i, D(i, j) represents the distance between power equipment i and power equipment j, and Dmax represents the upper limit influence distance of power equipment fault propagation, which is specifically used to normalize the distance D(i, j) between power equipment i and power equipment j, for simulating the distance attenuation effect, that is, the farther the distance between power equipment, the smaller the influence of fault propagation. The rate of reduction of this influence is calculated through an exponential attenuation function.

[0124] S5 includes S51;

[0125] S51. Analyze according to the obtained fault propagation state S and fault propagation influence range R, obtain the power equipment state Z that needs to be adjusted, and regulate the triggering mechanism of the power supply scheduling strategy adaptively matched to the power equipment state Z;

[0126] The power equipment state Z is obtained through the following analysis method:

[0127] ;

[0128] In the formula, Z(i, t) represents the power equipment state of power equipment i at time t, Sthe represents the preset fault propagation state trigger threshold, Rthe represents the preset fault propagation influence range trigger threshold, and Hthe represents the preset power equipment health index threshold;

[0129] When the power equipment state Z(i, t) of power equipment i at time t = 1, it means that power equipment i needs to be adjusted, triggers the power supply scheduling strategy of power equipment i, and at the same time prompts the relevant inspection and maintenance personnel that power equipment i needs maintenance, and generates a maintenance task to be sent to the pending task list of the relevant inspection and maintenance personnel for processing;

[0130] When the power equipment state Z(i, t) of power equipment i at time t = 0, it means that power equipment i does not need to be adjusted;

[0131] Among them, the power supply scheduling strategy is used to detect the number of power equipment that needs to be adjusted. When the total number of power equipment that needs to be adjusted reaches the preset distributed fault threshold, it prompts to separate the faulty equipment.

[0132] In this embodiment, by combining the spatio-temporal feature vector F of power equipment, the fault propagation state S, and the distance D between power equipment, the accurate calculation of the fault propagation influence range R is realized. By normalizing the distance between equipment and applying the exponential decay function, the spatial decay effect of fault propagation can be effectively simulated, ensuring the accurate definition of the fault propagation influence range. Subsequently, based on the fault propagation state S and the influence range R, the state Z of the power equipment that needs to be adjusted can be intelligently analyzed, and the corresponding power equipment scheduling strategy can be triggered. This scheduling mechanism can contain the further spread of the fault by adjusting the equipment state in the early stage of the fault spread. When the state Z of the power equipment is 1, an adjustment instruction is issued in a timely manner, and a maintenance task is automatically generated, improving the response speed and working efficiency of the power equipment operation and maintenance. Through this method, not only can the maintenance scheduling of power equipment be optimized, but also the spread of faults can be effectively reduced through accurate monitoring of the power equipment state at critical moments, enhancing the overall reliability and recovery ability of the power system.

[0133] Embodiment 5

[0134] A power equipment operation and maintenance monitoring system based on data analysis, please refer to Figure 2 , specifically: including a power equipment data acquisition module, a fault prediction correlation module, a power equipment fault propagation module, a propagation influence range module, and a power equipment operation and maintenance decision-making module;

[0135] The power equipment data acquisition module collects the real-time operation state data of the equipment through sensors and equipment monitoring systems, and at the same time collects the spatio-temporal information of the equipment. After preprocessing the real-time operation state data and spatio-temporal information, a spatio-temporal feature vector F is formed;

[0136] The fault prediction correlation module extracts features from the spatio-temporal feature vector F to obtain the feature vector associated with the occurrence of the fault, and forms a fault prediction feature vector Ffail;

[0137] The power equipment fault propagation module is based on the obtained spatio-temporal feature vector F and fault prediction feature vector Ffail, and combines the historical fault data and network connection relationships stored in the power equipment to establish a propagation matrix M between the power equipment, and obtains the fault propagation state S of the power equipment;

[0138] The propagation influence range module obtains the distance D between power equipment based on the spatio-temporal information in the spatio-temporal feature vector F, and calculates the propagation range of the fault with the obtained fault propagation state S of the power equipment to obtain the fault propagation influence range R;

[0139] The power equipment operation and maintenance decision-making module analyzes based on the obtained fault propagation state S and fault propagation influence range R, obtains the power equipment state Z that needs to be adjusted, and controls the triggering mechanism of the power supply scheduling strategy adaptively matched to the power equipment state Z.

[0140] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A power equipment operation and maintenance monitoring method based on data analysis, characterized in that: It includes the following steps: S1. Collect the real-time operation status data of the device through the sensor and device monitoring system, and at the same time collect the spatio-temporal information of the device. After preprocessing the real-time operation status data and spatio-temporal information, form the spatio-temporal feature vector F; S2. Extract features from the spatio-temporal feature vector F to obtain the feature vector associated with the occurrence of faults, and form the fault prediction feature vector Ffail; S3. Based on the obtained spatio-temporal feature vector F and fault prediction feature vector Ffail, and combined with the historical fault data and network connection relationship stored in the power equipment, establish the propagation matrix M between power equipment, and obtain the fault propagation state S of the power equipment; S4. Based on the spatio-temporal information in the spatio-temporal feature vector F, obtain the distance D between power equipment, and calculate the propagation range of the fault with the obtained fault propagation state S of the power equipment to obtain the fault propagation influence range R; S5. Analyze according to the obtained fault propagation state S and fault propagation influence range R, obtain the power equipment state Z that needs to be adjusted, and control the trigger mechanism of the power supply scheduling strategy adaptively matched to the power equipment state Z.

2. A power equipment operation and maintenance monitoring method based on data analysis according to claim 1, characterized in that: S1 includes S11 and S12; S11. Through the sensors and monitoring system of the device, collect the real-time operation status data and spatio-temporal information of the device in real time. The real-time operation status data includes the load L, power P, temperature W, and vibration V data of the power equipment, and the spatio-temporal information includes the position coordinates (x, y) of the power equipment; S12. Preprocess the real-time operation status data. The preprocessing includes filling in missing data and standardization processing to obtain the preprocessed standard load CL, standard power CP, standard temperature CW, and standard vibration CV, and integrate them with the spatio-temporal information to obtain the spatio-temporal feature vector F.

3. A power equipment operation and maintenance monitoring method based on data analysis according to claim 2, characterized in that: S2 includes S21 and S22; S21. Extract features from the spatio-temporal feature vector F to obtain the feature vector associated with the occurrence of faults, including the load-fault correlation coefficient RLF, temperature-fault correlation coefficient RWF, vibration-fault correlation coefficient RVF, and power-fault correlation coefficient RPF; S22. Recombine the load-fault correlation coefficient RLF, temperature-fault correlation coefficient RWF, vibration-fault correlation coefficient RVF, and power-fault correlation coefficient RPF with the position coordinates (x, y) and timestamp t in the spatio-temporal feature vector F to obtain the fault prediction feature vector Ffail.

4. A power equipment operation and maintenance monitoring method based on data analysis according to claim 3, characterized in that: Through the load-fault correlation coefficient RLF, temperature-fault correlation coefficient RWF, vibration-fault correlation coefficient RVF, and power-fault correlation coefficient RPF, the correlation degree between the standard load CL, standard power CP, standard temperature CW, standard vibration CV, and the fault label F can be understood; Among them, the correlation degree is as follows: When the load-fault correlation coefficient RLF, temperature-fault correlation coefficient RWF, vibration-fault correlation coefficient RVF, and power-fault correlation coefficient RPF = 1, it indicates a positive correlation, indicating that when the standard load CL, standard power CP, standard temperature CW, and standard vibration CV increase, the probability of fault occurrence also increases, and there is a strict linear relationship between the two; When the load-fault correlation coefficient RLF, temperature-fault correlation coefficient RWF, vibration-fault correlation coefficient RVF, and power-fault correlation coefficient RPF = 0, it indicates no linear relationship. That is, when the standard load CL, standard power CP, standard temperature CW, and standard vibration CV change, the probability of a fault occurring does not change, and the changes of the two are not correlated. When the load-fault correlation coefficient RLF, temperature-fault correlation coefficient RWF, vibration-fault correlation coefficient RVF, and power-fault correlation coefficient RPF = -1, it indicates a negative correlation. That is, when the standard load CL, standard power CP, standard temperature CW, and standard vibration CV increase, the probability of a fault occurring decreases, and there is a strict negative linear relationship between the two.

5. A power equipment operation and maintenance monitoring method based on data analysis according to claim 3, characterized in that: The load-fault correlation coefficient RLF is obtained through the following calculation formula: ； In the formula, CL(p) represents the standard load of the p-th data point, F(p) represents the fault mark of the power equipment at the p-th data point, μ(F) represents the mean value of the fault mark data, which reflects the average probability of a fault occurring during the sampling time period, μ(CL) represents the mean value of the standard load data, specifically representing the average value of all standard loads in the standard load sequence, n represents the total number of data points, specifically representing the number of samples of the power equipment participating in the calculation. The temperature-fault correlation coefficient RWF is obtained through the following calculation formula: ； In the formula, CW(p) represents the standard temperature of the p-th data point, μ(CW) represents the mean value of the standard temperature data, specifically representing the average value of all standard loads in the standard temperature sequence. The vibration-fault correlation coefficient RVF is obtained through the following calculation formula: ； In the formula, CV(p) represents the standard vibration of the p-th data point, μ(CV) represents the mean value of the standard vibration data, specifically representing the average value of all standard loads in the standard vibration sequence. The power-fault correlation coefficient RPF is obtained through the following calculation formula: ； In the formula, CP(p) represents the standard power of the p-th data point, μ(CP) represents the mean value of the standard power data, specifically representing the average value of all standard loads in the standard power sequence.

6. A power equipment operation and maintenance monitoring method based on data analysis according to claim 5, characterized in that: S3 includes S31 and S32; S31, based on the obtained spatio-temporal feature vector F, combined with the historical fault data stored in the power equipment and the connection relationships between various power equipment; Among them, the historical fault data includes the fault occurrence time, type, and influence range; the connection relationships include the power transmission connection between power equipment, the physical connection between power equipment, and the data exchange connection between power equipment; Based on the historical fault data, the connection relationships between power equipment, and the spatio-temporal feature vector F of the power equipment, analyze the influence of different power equipment on other power equipment when a fault occurs, including that a power equipment failure leads to a power outage, which in turn affects other connected power equipment; establish a propagation matrix M based on the obtained influence. Each row and each column in the propagation matrix M represents a power equipment, and the element m(i, j) in the propagation matrix M represents the degree of influence of power equipment i on power equipment j.

7. A power equipment operation and maintenance monitoring method based on data analysis according to claim 6, characterized in that: S32. Analyze the fault propagation state S of each power device based on the propagation matrix M and the fault prediction feature vector Ffail. The fault propagation state S is used to reflect the situation that the fault of each power device is not only determined by the fault probability of the power device itself, but also related to the fault states of the adjacent devices of the power device and the propagation path between the power devices. The fault propagation state S is obtained through the following calculation formula: ； In the formula, S(i, t) represents the fault propagation state of power device i at time t, α(i) represents the weighted coefficient of power device i, specifically representing the fault sensitivity of power device i, Ffail(i, t) represents the fault prediction feature vector of power device i at time t, w represents the total number of power devices, β(i, j) represents the fault influence weight of power device i on power device j, Ffail(j, t) represents the fault prediction feature vector of power device j at time t, λ(i, j) represents the attenuation factor of fault propagation, specifically representing the rate at which the propagation effect decays with the spatial distance and the passage of time when the fault propagates from power device i to power device j, △t represents the fault propagation time difference, and exp represents the exponential decay function.

8. A power equipment operation and maintenance monitoring method based on data analysis according to claim 7, characterized in that: S4 includes S41; S41. Based on the spatio-temporal information in the spatio-temporal feature vector F, obtain the distance D between power devices, and calculate the propagation range of the fault with the obtained fault propagation state S of the power device to obtain the fault propagation influence range R; Among them, the distance D is calculated and obtained through the position coordinates (x, y) in the spatio-temporal information of the spatio-temporal feature vector F; The fault propagation influence range R is obtained through the following calculation formula: ； Wherein, R(i) represents the fault propagation influence range of power equipment i, D(i, j) represents the distance between power equipment i and power equipment j, and Dmax represents the upper limit influence distance of power equipment fault propagation, which is specifically used for normalizing the distance D(i, j) between power equipment i and power equipment j. It is used to simulate the distance attenuation effect.

9. A power equipment operation and maintenance monitoring method based on data analysis according to claim 1, characterized in that: S5 includes S51; S51. Analyze according to the obtained fault propagation state S and the fault propagation influence range R, obtain the power device state Z that needs to be adjusted, and control the triggering mechanism of the power supply scheduling strategy adaptively matched to the power device state Z; The power device state Z is obtained through the following analysis method: ； In the formula, Z(i, t) represents the power device state of power device i at time t, Sthe represents the preset fault propagation state trigger threshold, Rthe represents the preset fault propagation influence range trigger threshold, and Hthe represents the preset power device health index threshold; When the power device state Z(i, t) of power device i at time t = 1, it means that power device i needs to be adjusted, triggers the power supply scheduling strategy matching power device i, and at the same time prompts the relevant inspection and maintenance personnel that power device i needs maintenance, and generates a maintenance task to be sent to the pending task list of the relevant inspection and maintenance personnel for processing; When the power device state Z(i, t) of power device i at time t = 0, it means that power device i does not need to be adjusted; Among them, the power supply scheduling strategy matching is used to detect the number of power devices that need to be adjusted. When the total number of power devices that need to be adjusted reaches the preset distributed fault threshold, it prompts to perform fault separation on the faulty devices.

10. A power equipment operation and maintenance monitoring system based on data analysis, which is applied to a power equipment operation and maintenance monitoring method based on data analysis according to any one of claims 1 to 9, and is characterized in that: It includes a power device data acquisition module, a fault prediction correlation module, a power device fault propagation module, a propagation influence range module, and a power device operation and maintenance decision module; The power equipment data acquisition module collects the real-time operation status data of the equipment through sensors and the equipment monitoring system, and at the same time collects the spatio-temporal information of the equipment. After preprocessing the real-time operation status data and spatio-temporal information, a spatio-temporal feature vector F is formed. The fault prediction correlation module extracts features from the spatio-temporal feature vector F to obtain a feature vector associated with the occurrence of a fault, and forms a fault prediction feature vector Ffail. The power equipment fault propagation module is based on the obtained spatio-temporal feature vector F and fault prediction feature vector Ffail, and combines the historical fault data and network connection relationships stored in the power equipment to establish a propagation matrix M between the power equipment, and obtains the fault propagation state S of the power equipment. The propagation influence range module obtains the distance D between power equipment based on the spatio-temporal information in the spatio-temporal feature vector F, and calculates the propagation range of the fault with the obtained fault propagation state S of the power equipment to obtain the fault propagation influence range R. The power equipment operation and maintenance decision-making module analyzes according to the obtained fault propagation state S and fault propagation influence range R, obtains the power equipment state Z that needs to be adjusted, and controls the trigger mechanism of the power supply scheduling strategy adaptively matched to the power equipment state Z.

Images