State monitoring method for monitoring energy consumption of secondary equipment of power grid and related equipment

Abstract

The invention discloses a state monitoring method for power grid secondary equipment energy consumption monitoring and related equipment, which can effectively remove noise and redundant information in data through feature extraction and dimension reduction processing of a principal component analysis method so as to retain key information, remarkably improve the accuracy of signal feature extraction and reduce the influence of external interference. Through a principal component analysis method and grid processing, data optimization is realized, and power consumption monitoring is realized under the condition that no extra hardware equipment is added, so that the cost and the operation and maintenance burden are reduced; through the PCA dimension reduction and grid processing technology, the data dimension and the data volume are reduced, the requirement for computing resources is lowered, the local outlier factor algorithm is low in computing complexity, the power grid secondary equipment can be monitored under the condition of limited resources, and the monitoring efficiency and accuracy are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid secondary equipment status monitoring, and in particular to a status monitoring method and related equipment for monitoring energy consumption of power grid secondary equipment. Background Art

[0002] In the power grid system, the stable operation of secondary equipment plays a vital role in ensuring the safety and reliable power supply of the power system. Therefore, secondary equipment status monitoring technology has always been a research focus in the power field.

[0003] Currently, a wide range of research has been conducted on the status monitoring of power grid secondary equipment, with research based on energy consumption monitoring being particularly prominent. For example, the paper "Discussion on Status Monitoring Content of Power Grid Secondary Equipment" proposes an innovative method for status monitoring using secondary equipment power consumption data and for correlating monitoring across multiple data sources. This method uses power consumption data collection equipment to compare the real-time power consumption of secondary equipment with the average power consumption over different time periods. If the real-time power consumption consistently exceeds the average power consumption over different time periods, the system will immediately issue an alarm signal indicating an abnormality in the secondary equipment power consumption data, allowing operations and maintenance personnel to promptly detect and address potential issues.

[0004] In addition to energy consumption monitoring methods, other monitoring technologies are also evolving. Signal feature analysis-based methods also play a key role in secondary equipment status monitoring. This method precisely collects and deeply analyzes electrical and optical signals generated by secondary equipment during operation, extracting characteristic quantities closely related to the equipment's status. For example, it closely monitors the dynamic changes in key signal parameters such as amplitude, frequency, and phase. Any deviations from normal ranges in these parameters indicate potential equipment failure. Furthermore, with the rapid development of artificial intelligence (AI) technology, monitoring methods based on AI algorithms have emerged. Advanced algorithms such as neural networks and support vector machines are widely used in secondary equipment status monitoring. Through deep learning and training on massive amounts of secondary equipment operating data, a highly accurate equipment status prediction model is constructed. Simply inputting real-time monitoring data, the model leverages its powerful computing power and intelligent algorithms to accurately predict the equipment's future operating status, proactively identifying potential failure risks and saving valuable time for equipment maintenance.

[0005] However, the existing secondary equipment status monitoring technology still has obvious limitations. In terms of the energy consumption monitoring method, the literature fails to clearly define to what extent the power consumption of secondary equipment deviates from the average power consumption to trigger a warning message. This lack of a precise quantitative judgment standard in actual application scenarios seriously affects the accuracy and reliability of monitoring results. At the same time, the current power grid secondary equipment itself does not have the function of directly monitoring power consumption. To achieve this function, additional power consumption acquisition devices need to be added to the secondary equipment, which will undoubtedly greatly increase the equipment cost and operation and maintenance burden, and thus largely limit the large-scale promotion and application of energy consumption monitoring technology. Although the method based on artificial intelligence algorithms has powerful prediction capabilities, it has extremely high requirements for data volume and computing resources, and it is difficult to fully exert its advantages in some practical application scenarios with limited resources.

[0006] In summary, in order to meet the growing demand for the status monitoring of power grid secondary equipment, it is urgent to develop a more perfect, efficient and cost-controlled innovative solution for the status monitoring of secondary equipment based on energy consumption monitoring to effectively overcome the many problems existing in the existing technology, improve the operation stability and reliability of power grid secondary equipment, and ensure the safe and stable operation of the power system. Summary of the Invention

[0007] Based on the problems raised in the above background technology, the purpose of the present invention is to provide a status monitoring method and related equipment for the energy consumption monitoring of power grid secondary equipment, which solves the problems in the existing technology that additional hardware equipment is required to obtain data, the signal characteristics are easily interfered by the outside world, and it is difficult to be applied to application scenarios with limited resources.

[0008] The present invention is realized through the following technical solutions:

[0009] The first aspect of the present invention provides a status monitoring method for the energy consumption monitoring of power grid secondary equipment, including the following steps:

[0010] Obtain device energy consumption data, extract features from the device energy consumption data to obtain energy consumption features;

[0011] Use the principal component analysis method to perform dimensionality reduction processing on the energy consumption features to obtain dimensionality-reduced energy consumption features;

[0012] Based on the grid processing technology, perform data screening on the dimensionality-reduced energy consumption features to obtain low-density energy consumption features;

[0013] Calculate the local outlier factor for the low-density energy consumption features to obtain the local outlier factor value. If the local outlier factor value is greater than the anomaly threshold, the power grid secondary equipment corresponding to the local outlier factor value is defined as an abnormal power grid secondary equipment.

[0014] In the above technical solution, the energy consumption of secondary power grid equipment is represented by the average hourly energy consumption. The energy consumption of each secondary power grid equipment can be expressed as a multi-dimensional vector. Through feature extraction, the eigenvectors of the energy consumption characteristics can be extracted from the multi-dimensional vector representing the energy consumption. These eigenvectors can comprehensively describe the energy consumption characteristics of the equipment and provide a basis for subsequent anomaly detection.

[0015] Since there is a strong correlation between the various eigenvalues, which means that there may be overlapping information between different features. Therefore, in order to remove redundancy and simplify the feature data, the principal component analysis method is adopted in this method to effectively merge the highly correlated features into a few unrelated principal components, generating the reduced-dimensional energy consumption characteristics. The reduced-dimensional energy consumption characteristics are the most representative features of the energy consumption of secondary power grid equipment. By performing dimensionality reduction on the energy consumption characteristics, data redundancy is reduced, the data structure is simplified, and a more concise and effective input is provided for subsequent anomaly detection. Through feature extraction and dimensionality reduction by the principal component analysis method, the noise and redundant information in the data can be effectively removed to retain the key information, significantly improving the accuracy of signal feature extraction and reducing the influence of external interference.

[0016] Since this method uses the LOF algorithm (Local Outlier Factor) to calculate the feature data, and for large-scale data sets, the running time of the LOF algorithm will be very long. Usually, the outlier objects in the data set only account for a small number. Therefore, in order to improve the calculation efficiency and reduce the amount of calculation, this method is based on the grid processing technology to divide the reduced-dimensional energy consumption characteristics into grids, and select the data points in the low-density areas from the reduced-dimensional energy consumption characteristics after grid division as the detection features. Through data screening, the calculation efficiency is significantly improved and unnecessary calculation amount is reduced. Through the principal component analysis method and grid processing, the optimization of data is realized, and power consumption monitoring is realized without additional hardware equipment, thereby reducing the cost and operation and maintenance burden.

[0017] Finally, calculate the local outlier factor of the low-density energy consumption characteristics, calculate the local outlier factor value of each device to quantify its degree of anomaly. The secondary power grid equipment corresponding to the local outlier factor value greater than the anomaly threshold has an obvious difference compared with the normal equipment. Therefore, these secondary power grid equipment are defined as abnormal secondary power grid equipment. Through the calculation of the local outlier factor, a clear quantitative basis can be provided for the anomaly detection of secondary power grid equipment, thus solving the problem of the lack of quantitative judgment criteria. At the same time, through PCA dimensionality reduction and grid processing technology, the data dimension and data volume are reduced, the demand for computing resources is reduced, and the calculation complexity of the local outlier factor algorithm itself is relatively low, so that the monitoring of secondary power grid equipment can be realized under limited resource conditions, improving the monitoring efficiency and accuracy.

[0018] In an alternative embodiment, the energy consumption characteristics include: a trend index, a variability index, and a volatility index;

[0019] The trend index includes an energy consumption increase rate, an energy consumption decrease rate, and an energy consumption slope; wherein, the energy consumption increase rate and the energy consumption decrease rate are obtained by calculating the average growth value and the average decrease value of the device energy consumption data; the energy consumption slope is obtained by performing linear regression fitting on the energy consumption data;

[0020] The variability index includes an energy consumption mean difference and an energy consumption mean ratio; wherein, the energy consumption mean difference is obtained by calculating the difference value between the hourly energy consumption mean of the device energy consumption data and the current energy consumption mean; the energy consumption mean ratio is obtained by calculating the ratio of the hourly energy consumption mean of the device energy consumption data to the energy consumption mean over the monitoring duration;

[0021] The volatility index includes an energy consumption standard deviation, a frequency characteristic, and a median correlation coefficient; wherein, the frequency characteristic is obtained by calculating the modulus of the difference of the Fourier coefficients of the device energy consumption data; the median correlation coefficient is obtained by calculating the correlation between the device energy consumption data and the median of the device energy consumption data.

[0022] In an alternative embodiment, the principal component analysis method is used to perform dimensionality reduction processing on the energy consumption characteristics, including the following steps:

[0023] Perform data standardization on the energy consumption characteristics to obtain standardized energy consumption characteristics;

[0024] Calculate the covariance matrix of the standardized energy consumption characteristics, perform eigenvalue decomposition on the covariance matrix to obtain eigenvalues and eigenvectors;

[0025] Sort the eigenvalues from largest to smallest, and select the n eigenvectors corresponding to the top n eigenvalues from the sorted eigenvalues as the main energy consumption characteristics;

[0026] Project the main energy consumption characteristics into the principal component space to generate a scatter plot.

[0027] In an alternative embodiment, data screening is performed on the dimensionality-reduced energy consumption characteristics based on grid processing technology, including the following steps:

[0028] Divide the scatter plot into a number of grids of equal size;

[0029] Calculate the number of data points of the dimensionality-reduced energy consumption characteristics in each grid;

[0030] Divide the number of grids into internal cells and boundary cells according to the number of data points, delete the dimensionality-reduced energy consumption characteristics within the internal cells, and retain the dimensionality-reduced energy consumption characteristics within the boundary cells to generate low-density energy consumption characteristics.

[0031] In an alternative embodiment, a number of grids are divided into internal cells and boundary cells according to the number of the data points, including:

[0032] Select any one of the grids as the grid to be judged;

[0033] If any adjacent grid of the grid to be judged does not contain the dimensionality-reduced energy consumption feature, the grid to be judged is a boundary cell; otherwise, the grid to be judged is an internal cell.

[0034] In an alternative embodiment, calculating the local outlier factor for the low-density energy consumption feature includes the following steps:

[0035] Determine the k-nearest neighbor distance of the low-density energy consumption feature, and divide the k-distance neighborhood of the low-density energy consumption feature based on the k-nearest neighbor distance;

[0036] Determine the reachable distance of the low-density energy consumption feature, and calculate the local reachable density and the average local reachable density of the low-density energy consumption feature within the k-distance neighborhood based on the reachable distance;

[0037] Calculate the ratio of the average local reachable density to the local reachable density to obtain the local outlier factor value.

[0038] In an alternative embodiment, the calculation formula of the local reachable density is as follows:

[0039]

[0040] In the above formula, lrd MinPts (p) is the local reachable density of the low-density energy consumption feature data point p, and reach-dist k (p, o) is the k-th reachable distance from the low-density energy consumption feature data point p to the low-density energy consumption feature data point o, and N MinPts (p) is the number of low-density energy consumption feature data points within the k-distance neighborhood of the low-density energy consumption feature data point p.

[0041] The second aspect of the present invention provides a state monitoring system for power grid secondary equipment energy consumption monitoring, including:

[0042] A feature extraction module, configured to obtain device energy consumption data, extract features from the device energy consumption data to obtain energy consumption features;

[0043] A dimensionality reduction module, configured to perform dimensionality reduction processing on the energy consumption features by using the principal component analysis method to obtain dimensionality-reduced energy consumption features;

[0044] A data screening module, configured to screen the dimensionality-reduced energy consumption features based on grid processing technology to obtain low-density energy consumption features;

[0045] A local outlier factor calculation module, configured to calculate the local outlier factor for the low-density energy consumption features to obtain a local outlier factor value. If the local outlier factor value is greater than the anomaly threshold, the secondary power grid equipment corresponding to the local outlier factor value is defined as abnormal secondary power grid equipment.

[0046] A third aspect of the present invention provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, it implements a status monitoring method for energy consumption monitoring of secondary power grid equipment.

[0047] A fourth aspect of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements a status monitoring method for energy consumption monitoring of secondary power grid equipment.

[0048] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0049] 1. Through feature extraction and dimensionality reduction processing by the principal component analysis method, it is possible to effectively remove noise and redundant information in the data to retain key information, significantly improve the accuracy of signal feature extraction, and reduce the influence of external interference;

[0050] 2. Through the principal component analysis method and grid processing, the optimization of data is realized, and power consumption monitoring is realized without additional hardware devices, thereby reducing costs and operation and maintenance burdens;

[0051] 3. Through PCA dimensionality reduction and grid processing technology, the data dimension and data volume are reduced, the demand for computing resources is reduced, and the local outlier factor algorithm itself has a low computational complexity, so that the monitoring of secondary power grid equipment can be realized under limited resource conditions, improving the monitoring efficiency and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following will briefly introduce the drawings required in the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings. In the drawings:

[0053] Figure 1 It is a schematic flowchart of a status monitoring method for energy consumption monitoring of secondary power grid equipment provided in Embodiment 1 of the present invention;

[0054] Figure 2 A schematic structural diagram of an electronic device provided in Embodiment 3 of the present invention. Detailed implementation manners

[0055] To make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with embodiments and drawings. The illustrative embodiments and descriptions thereof of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0056] Embodiment 1

[0057] Figure 1 A schematic flowchart of a state monitoring method for power grid secondary equipment energy consumption monitoring provided in Embodiment 1 of the present invention. As Figure 1 shown, the state monitoring method for power grid secondary equipment energy consumption monitoring includes the following steps:

[0058] Obtain equipment energy consumption data, perform feature extraction on the equipment energy consumption data to obtain energy consumption features;

[0059] Use the principal component analysis method to perform dimensionality reduction processing on the energy consumption features to obtain dimensionality-reduced energy consumption features;

[0060] Based on the grid processing technology, perform data screening on the dimensionality-reduced energy consumption features to obtain low-density energy consumption features;

[0061] Calculate the local outlier factor for the low-density energy consumption features to obtain a local outlier factor value. If the local outlier factor value is greater than the anomaly threshold, the power grid secondary equipment corresponding to the local outlier factor value is defined as an abnormal power grid secondary equipment.

[0062] It should be noted that the energy consumption of power grid secondary equipment is represented by the hourly average energy consumption. The energy consumption of each power grid secondary equipment can be represented as a multi-dimensional vector. Through feature extraction, the feature vectors of energy consumption features can be extracted from the multi-dimensional vector representing energy consumption. These feature vectors can comprehensively describe the energy consumption characteristics of the equipment and provide a basis for subsequent anomaly detection.

[0063] Due to the strong correlation among various eigenvalues, which means there may be overlapping information between different features. Therefore, in order to remove redundancy and simplify the feature data, the principal component analysis method is adopted in this method to effectively merge highly correlated features into a few uncorrelated principal components, generating dimensionality-reduced energy consumption features. The dimensionality-reduced energy consumption features are the most representative features of the energy consumption of secondary equipment in the power grid. By performing dimensionality reduction on the energy consumption features, data redundancy is reduced, the data structure is simplified, and more concise and effective input is provided for subsequent anomaly detection. Through feature extraction and dimensionality reduction processing by the principal component analysis method, noise and redundant information in the data can be effectively removed to retain key information, significantly improving the accuracy of signal feature extraction and reducing the influence of external interference.

[0064] Since this method uses the LOF algorithm (Local Outlier Factor) to calculate the feature data, and the running time of the LOF algorithm for large-scale data sets will be very long. Usually, only a small number of outlier objects exist in the data set. Therefore, in order to improve the calculation efficiency and reduce the amount of calculation, this method is based on grid processing technology. The dimensionality-reduced energy consumption features are divided into grids, and data points in the low-density regions are selected from the dimensionality-reduced energy consumption features after grid division as detection features. The calculation efficiency is significantly improved through data screening, and unnecessary calculation amount is reduced. Through the principal component analysis method and grid processing, data optimization is achieved, and power consumption monitoring is realized without additional hardware devices, thus reducing costs and operation and maintenance burdens.

[0065] Finally, the local outlier factor is calculated for the low-density energy consumption features, and the local outlier factor value of each device is calculated to quantify its degree of anomaly. There are obvious differences between the secondary equipment of the power grid corresponding to the local outlier factor value greater than the anomaly threshold and normal equipment. Therefore, these secondary equipment of the power grid are defined as abnormal secondary equipment of the power grid. Calculating the local outlier factor can provide a clear quantitative basis for the anomaly detection of secondary equipment in the power grid, thus solving the problem of the lack of a quantitative judgment standard. At the same time, through PCA dimensionality reduction and grid processing technology, the data dimension and data volume are reduced, the demand for computing resources is reduced, and the local outlier factor algorithm itself has a relatively low computational complexity, enabling the monitoring of secondary equipment in the power grid to be realized under limited resource conditions, improving the monitoring efficiency and accuracy.

[0066] In this embodiment, the anomaly threshold is taken as 1.

[0067] In an alternative embodiment, the energy consumption features include: trend index, variability index, and volatility index;

[0068] The trend indicators include the energy consumption rising rate, the energy consumption falling rate, and the energy consumption slope. Among them, the energy consumption rising rate and the energy consumption falling rate are obtained by calculating the average growth value and the average decline value of the equipment energy consumption data. The energy consumption slope is obtained by performing linear regression fitting on the energy consumption data.

[0069] The variability indicators include the energy consumption mean difference and the energy consumption mean ratio. Among them, the energy consumption mean difference is obtained by calculating the difference value between the hourly energy consumption mean of the equipment energy consumption data and the current energy consumption mean. The energy consumption mean ratio is obtained by calculating the ratio of the hourly energy consumption mean of the equipment energy consumption data to the energy consumption mean during the monitoring duration.

[0070] The volatility indicators include the energy consumption standard deviation, the frequency feature, and the median correlation coefficient. Among them, the frequency feature is obtained by calculating the modulus of the difference of the Fourier coefficients of the equipment energy consumption data. The median correlation coefficient is obtained by calculating the correlation between the equipment energy consumption data and the median of the equipment energy consumption data.

[0071] It should be noted that the equipment energy consumption data is denoted as {x1, x2, …, x i , …, x H}, where x i is the energy consumption value at the i-th hour, and H is the length of the equipment energy consumption data. N pieces of power consumption data of the same type of secondary equipment for H hours are obtained. In this embodiment, taking 12 hours as a cycle, feature extraction is performed based on the hourly average energy consumption of each equipment for 12 hours, that is, H = 12.

[0072] Furthermore, the energy consumption rising rate V1 represents the load rising rate in the equipment energy consumption data and is used to measure the increasing trend of energy consumption within a certain period of time. The specific calculation process is as follows:

[0073]

[0074] The energy consumption falling rate V2 represents the energy consumption falling rate in the equipment energy consumption data and is used to measure the decreasing trend of energy consumption within a certain period of time. The specific calculation process is as follows:

[0075]

[0076] The energy consumption slope represents the linear change trend of the equipment energy consumption data and reflects the change trend of energy consumption over time. By performing linear regression fitting on the equipment energy consumption data, the slope is obtained. Specifically, the fitting formula is y = ax + b, where the slope a is the change rate of the equipment energy consumption data, that is, the energy consumption slope.

[0077] In this embodiment, the mean differences between the energy consumption means of the previous 1 hour, 3 hours, and 6 hours and the current energy consumption mean are calculated respectively as the energy consumption mean differences to measure the change of the load. Taking the energy consumption mean difference V3 of the previous 1 hour as an example, the specific calculation process is as follows:

[0078]

[0079] The ratios of the energy consumption means of the subsequent 3 hours, 6 hours, and 9 hours to the mean of the energy consumption of all hours are calculated respectively as the energy consumption mean ratios. Taking the energy consumption mean ratio V9 of the subsequent 3 hours as an example, the specific calculation process is as follows:

[0080]

[0081] In this embodiment, the standard deviations of the energy consumption for 1 hour, the previous 6 hours, and the subsequent 6 hours are as follows. The specific calculation process is as follows:

[0082]

[0083] In the above formula, μ represents the mean of the energy consumption per hour.

[0084] It should be noted that the standard deviation V6 of the energy consumption for 1 hour is used to measure the volatility of the energy consumption, the standard deviation V7 of the energy consumption for the previous 6 hours is used to measure the long-term energy consumption fluctuation, and the standard deviation V8 of the energy consumption for the previous 6 hours is used to measure the long-term trend of the energy consumption fluctuation.

[0085] In this embodiment, the frequency characteristics of the equipment energy consumption data are extracted by Fourier transform, and the modulus of the difference of the Fourier coefficients of the energy consumption data for 12 hours is calculated. The specific calculation process of the frequency characteristic V13 is as follows:

[0086]

[0087] In the above formula, F i 前 [[ID=3)); and F i 后 are the Fourier coefficients of the previous 6 hours and the subsequent 6 hours respectively.

[0088] The median correlation coefficient V14 reflects the difference in the energy consumption characteristics through the correlation between the energy consumption data of each device and the median sequence of the energy consumption of all devices. The specific calculation process is as follows:

[0089]

[0090] In the above formula, μ x and μ m are the means of the device energy consumption and the median sequence of the energy consumption of all devices respectively, x i and m iThey are the energy consumption values of the i-th device and the median sequence respectively.

[0091] In an alternative embodiment, the principal component analysis method is used to reduce the dimension of the energy consumption characteristics, including the following steps:

[0092] Perform data standardization on the energy consumption characteristics to obtain standardized energy consumption characteristics;

[0093] Calculate the covariance matrix of the standardized energy consumption characteristics, perform eigenvalue decomposition on the covariance matrix to obtain eigenvalues and eigenvectors;

[0094] Sort the eigenvalues from largest to smallest, and select the n eigenvectors corresponding to the first n eigenvalues from the sorted eigenvalues as the main energy consumption characteristics;

[0095] Project the main energy consumption characteristics into the principal component space to generate a scatter plot.

[0096] It should be noted that the core objective of principal component analysis is to retain the maximum variability in the data. In this embodiment, by decomposing the covariance matrix by eigenvalues, the first two principal components are selected. These two principal components represent most of the variance information in the data. Mapping the data onto the two-dimensional plane of these two principal components, where the abscissa is the first main energy consumption characteristic (principal component 1) and the ordinate is the second main energy consumption characteristic (principal component 2), can more clearly display the characteristics of the normal secondary device energy consumption and the abnormal secondary device energy consumption. The points of abnormal devices usually distribute in the low-density area, while the points of normal devices concentrate in the high-density area, which makes the subsequent anomaly detection more intuitive and efficient.

[0097] It should be noted that the manifestation form of the dimension-reduced energy consumption characteristics in this method is the scatter plot.

[0098] In an alternative embodiment, data screening is performed on the dimension-reduced energy consumption characteristics based on the grid processing technology, including the following steps:

[0099] Divide the scatter plot into several grids of equal size;

[0100] Calculate the number of data points of the dimension-reduced energy consumption characteristics in each grid;

[0101] Divide the several grids into internal cells and boundary cells according to the number of data points, delete the dimension-reduced energy consumption characteristics in the internal cells, and retain the dimension-reduced energy consumption characteristics in the boundary cells to generate low-density energy consumption characteristics.

[0102] It should be noted that for large-scale data sets, the running time of the LOF algorithm will be very long. Usually, the outlier objects in the data set only account for a small number. To improve the computing efficiency and reduce the amount of calculation, grid processing excludes the data points in the internal cells that account for the majority, filters out the data points in the boundary region, and then performs the next step of calculation.

[0103] In this embodiment, the two-dimensional plane obtained in the previous step is divided into w grids of equal size, and the two-dimensional coordinates of each dimensionality-reduced energy consumption feature are calculated to obtain the grid number where each dimensionality-reduced energy consumption feature is located, and the count within the grid is updated. By scanning all the dimensionality-reduced energy consumption features, the number of data points of the dimensionality-reduced energy consumption features contained in each grid can be obtained.

[0104] Check the adjacent grids of each grid to determine whether the current grid is a boundary region. Examine all cells and find all internal cells, delete the data points contained in the internal cells, and retain the data points in the boundary cells to reduce the scale of the data set for easy calculation.

[0105] In an alternative embodiment, several grids are divided into internal cells and boundary cells according to the number of the data points, including:

[0106] Select any one of the several grids as the grid to be judged;

[0107] If any adjacent grid of the grid to be judged does not contain the dimensionality-reduced energy consumption feature, the grid to be judged is a boundary cell; otherwise, the grid to be judged is an internal cell.

[0108] In an alternative embodiment, the local outlier factor of the low-density energy consumption feature is calculated, including the following steps:

[0109] Determine the k-nearest neighbor distance of the low-density energy consumption feature, and divide the k-distance neighborhood of the low-density energy consumption feature based on the k-nearest neighbor distance;

[0110] Determine the reachable distance of the low-density energy consumption feature, and calculate the local reachable density and the average local reachable density of the low-density energy consumption feature within the k-distance neighborhood based on the reachable distance;

[0111] Calculate the ratio of the average local reachable density to the local reachable density to obtain the local outlier factor value.

[0112] It should be noted that taking the low-density energy consumption characteristic data point p as an example, among the several data points closest to the low-density energy consumption characteristic data point p, the distance between the k-th closest point and the low-density energy consumption characteristic data point p becomes the k-nearest distance of the low-density energy consumption characteristic data point p, denoted as k-distance(p). Taking the low-density energy consumption characteristic data point p as the center and the k-nearest distance as the radius to draw a circle, the area within this circle is divided into the k-distance neighborhood of the low-density energy consumption characteristic.

[0113] Among them, the local reachability density of the low-density energy consumption characteristic data point P is the reciprocal of the average distance of the reachable distances from the low-density energy consumption characteristic data point P to the points within its k-neighborhood, that is, the greater the distance, the smaller the density. Therefore, the k-th reachable distance from the low-density energy consumption characteristic data point p to the point o is defined as:

[0114] reach-dist k (p, o) = max{k-dist(o), d(p, o)}

[0115] Based on the reachable distance, calculate the local reachability density lrd MinPts (p) within the k-distance neighborhood, and calculate the average local reachability density of all points within the k-distance neighborhood. The ratio of the average local reachability density to the local reachability density is the local outlier factor value required in this embodiment.

[0116] Among them, if the low-density energy consumption characteristic data point p is an energy consumption characteristic with a greater degree of outlier, its local reachability density is smaller while the local reachability density of the energy consumption characteristics within its neighborhood is larger, that is, the local outlier factor value of the low-density energy consumption characteristic data point p is larger. Therefore, the low-density energy consumption characteristics with a higher local outlier factor value are considered outlier energy consumption characteristics.

[0117] In an alternative embodiment, the calculation formula of the local reachability density is as follows:

[0118]

[0119] In the above formula, lrd MinPts (p) is the local reachability density of the low-density energy consumption characteristic data point p, reach-dist k (p, o) is the k-th reachable distance from the low-density energy consumption characteristic data point p to the low-density energy consumption characteristic data point o, N MinPts (p) is the number of low-density energy consumption characteristic data points within the k-distance neighborhood of the low-density energy consumption characteristic data point p.

[0120] Embodiment 2

[0121] Embodiment 2 of the present invention provides a status monitoring system for power grid secondary equipment energy consumption monitoring, including:

[0122] A feature extraction module, configured to obtain device energy consumption data, perform feature extraction on the device energy consumption data, and obtain energy consumption features;

[0123] A dimensionality reduction module, configured to perform dimensionality reduction processing on the energy consumption features by using the principal component analysis method to obtain dimensionality-reduced energy consumption features;

[0124] A data screening module, configured to perform data screening on the dimensionality-reduced energy consumption features based on grid processing technology to obtain low-density energy consumption features;

[0125] A local outlier factor calculation module, configured to calculate the local outlier factor for the low-density energy consumption features to obtain a local outlier factor value. If the local outlier factor value is greater than the anomaly threshold, the secondary power grid device corresponding to the local outlier factor value is defined as an abnormal secondary power grid device.

[0126] Embodiment 3

[0127] Figure 2 The following is a schematic structural diagram of an electronic device provided in Embodiment 3 of the present invention. As Figure 2 shown, the electronic device includes a processor 21, a memory 22, an input device 23, and an output device 24. The number of processors 21 in the computer device may be one or more. Figure 2 Here, one processor 21 is taken as an example. The processor 21, the memory 22, the input device 23, and the output device 24 in the electronic device may be connected through a bus or other means. Figure 2 Here, the connection through the bus is taken as an example.

[0128] The memory 22, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules. The processor 21 executes various functional applications and data processing of the electronic device by running the software programs, instructions, and modules stored in the memory 22, that is, implements the state monitoring method for power grid secondary device energy consumption monitoring in Embodiment 1.

[0129] The memory 22 may mainly include a program storage area and a data storage area. Among them, the program storage area may store an operating system and application programs required for at least one function. The data storage area may store data created according to the use of the terminal, etc. In addition, the memory 22 may include a high-speed random access memory, and may further include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state storage devices. In some instances, the memory 22 may further include a memory remotely set relative to the processor 21, and these remote memories may be connected to the electronic device through a network. Examples of the above network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0130] The input device 23 can be used to receive user inputs such as IDs and passwords. The output device 24 is used to output the network configuration page.

[0131] Embodiment 4

[0132] Embodiment 4 of the present invention further provides a computer-readable storage medium. The computer-executable instructions, when executed by a computer processor, are used to implement the status monitoring method for power grid secondary equipment energy consumption monitoring provided in Embodiment 1.

[0133] A storage medium containing computer-executable instructions provided in the embodiments of the present invention. The computer-executable instructions are not limited to the method operations provided in Embodiment 1, and can also execute relevant operations in the status monitoring methods for power grid secondary equipment energy consumption monitoring provided in any embodiment of the present invention.

[0134] The specific embodiments described above further elaborate on the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A status monitoring method for power grid secondary equipment energy consumption monitoring, characterized in that, It includes the following steps: Obtain the device energy consumption data, extract features from the device energy consumption data to obtain energy consumption features; Use the principal component analysis method to perform dimensionality reduction processing on the energy consumption features to obtain dimensionality-reduced energy consumption features; Based on the grid processing technology, perform data screening on the dimensionality-reduced energy consumption features to obtain low-density energy consumption features; Calculate the local outlier factor for the low-density energy consumption features to obtain the local outlier factor value. If the local outlier factor value is greater than the anomaly threshold, define the power grid secondary device corresponding to the local outlier factor value as an abnormal power grid secondary device.

2. The state monitoring method for power grid secondary equipment energy consumption monitoring according to claim 1, characterized in that, The energy consumption features include: trend index, variability index, and volatility index; The trend index includes the energy consumption increase rate, energy consumption decrease rate, and energy consumption slope; among them, the energy consumption increase rate and the energy consumption decrease rate are obtained by calculating the average growth value and average decrease value of the device energy consumption data; the energy consumption slope is obtained by performing linear regression fitting on the energy consumption data; The variability index includes the energy consumption mean difference and the energy consumption mean ratio; among them, the energy consumption mean difference is obtained by calculating the difference value between the hourly energy consumption mean and the current energy consumption mean of the device energy consumption data; the energy consumption mean ratio is obtained by calculating the ratio of the hourly energy consumption mean of the device energy consumption data to the energy consumption mean during the monitoring period; The volatility index includes the energy consumption standard deviation, frequency feature, and median correlation coefficient; among them, the frequency feature is obtained by calculating the modulus of the difference of the Fourier coefficients of the device energy consumption data; the median correlation coefficient is obtained by calculating the correlation between the device energy consumption data and the median of the device energy consumption data.

3. The state monitoring method for power grid secondary equipment energy consumption monitoring according to claim 1, wherein, Using the principal component analysis method to perform dimensionality reduction processing on the energy consumption features includes the following steps: Perform data standardization on the energy consumption features to obtain standardized energy consumption features; Calculate the covariance matrix of the standardized energy consumption features, perform eigenvalue decomposition on the covariance matrix to obtain eigenvalues and eigenvectors; Sort the eigenvalues from largest to smallest, and select the n eigenvectors corresponding to the first n eigenvalues from the sorted eigenvalues as the main energy consumption features; Project the main energy consumption features into the principal component space to generate a scatter plot.

4. The state monitoring method for power grid secondary equipment energy consumption monitoring according to claim 3, characterized in that, Based on the grid processing technology, performing data screening on the dimensionality-reduced energy consumption features includes the following steps: Divide the scatter plot into several grids of equal size; Calculate the number of data points of the dimensionality-reduced energy consumption features in each grid; According to the number of data points, divide the several grids into internal cells and boundary cells, delete the dimensionality-reduced energy consumption features in the internal cells, and retain the dimensionality-reduced energy consumption features in the boundary cells to generate low-density energy consumption features.

5. The state monitoring method for power grid secondary equipment energy consumption monitoring according to claim 4, characterized in that, Dividing the several grids into internal cells and boundary cells according to the number of data points includes: Select any one of the several grids as the grid to be judged; If any adjacent grid of the grid to be judged does not contain the dimensionality-reduced energy consumption features, the grid to be judged is a boundary cell; otherwise, the grid to be judged is an internal cell.

6. The state monitoring method for power grid secondary equipment energy consumption monitoring according to claim 1, wherein Calculating the local outlier factor for the low-density energy consumption features includes the following steps: Determine the k-nearest neighbor distance of the low-density energy consumption characteristics, and divide the k-distance neighborhood of the low-density energy consumption characteristics based on the k-nearest neighbor distance; Determine the reachable distance of the low-density energy consumption characteristics, and calculate the local reachability density and the average local reachability density of the low-density energy consumption characteristics within the k-distance neighborhood based on the reachable distance; Calculate the ratio of the average local reachability density to the local reachability density to obtain the local outlier factor value.

7. The state monitoring method for power grid secondary equipment energy consumption monitoring according to claim 6, characterized in that The calculation formula of the local reachability density is as follows: In the above formula, lrd MinPts (p) is the local reachability density of the low-density energy consumption characteristic data point p, reach-dist k (p, o) is the k-th reachable distance from the low-density energy consumption characteristic data point p to the low-density energy consumption characteristic data point o, N MinPts (p) is the number of low-density energy consumption characteristic data points within the k-distance neighborhood of the low-density energy consumption characteristic data point p.

8. A state monitoring system for power grid secondary equipment energy consumption monitoring, characterized in that, Including: A feature extraction module, configured to obtain device energy consumption data, perform feature extraction on the device energy consumption data to obtain energy consumption characteristics; A dimensionality reduction module, configured to perform dimensionality reduction processing on the energy consumption characteristics by using the principal component analysis method to obtain dimensionality-reduced energy consumption characteristics; A data screening module, configured to perform data screening on the dimensionality-reduced energy consumption characteristics based on the grid processing technology to obtain low-density energy consumption characteristics; A local outlier factor calculation module, configured to calculate the local outlier factor of the low-density energy consumption characteristics to obtain a local outlier factor value. If the local outlier factor value is greater than the anomaly threshold, the power grid secondary device corresponding to the local outlier factor value is defined as an abnormal power grid secondary device.

9. An electronic device, characterized in that, Including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the state monitoring method for power grid secondary device energy consumption monitoring according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the state monitoring method for power grid secondary device energy consumption monitoring according to any one of claims 1 to 7 is implemented.

Images