An energy-saving diagnosis method for abnormal electrical load

By constructing U-I trajectory diagrams and analyzing the difference in characteristic data, and calculating the similarity in combination with the shape context algorithm, the accuracy problem of electrical load abnormal detection in the prior art is solved, and more efficient electrical load abnormal monitoring is achieved.

CN115600837BActive Publication Date: 2024-11-05STATE GRID SHANDONG ELECTRIC POWER CO MARKETING SERVICE CENT (MEASURING CENT)
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202211361650.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-11-02
Publication Date
2024-11-05
Estimated Expiration
2042-11-02

AI Technical Summary

Technical Problem

Existing non-invasive load monitoring technology is difficult to cope with the replacement and transformation of complex equipment, and it is impossible to promptly and accurately identify abnormal electrical load conditions.

Method used

By collecting current and voltage data, building U-I trajectory diagrams, calculating symmetry indicators and area, analyzing characteristic data differences in combination with historical data, using improved shape context algorithms to calculate trajectory diagram similarity, calculating load abnormality indicators, and determining electrical load abnormality conditions .

Benefits of technology

It improves the accuracy and reliability of electrical load abnormality detection, and can make preliminary judgments in the steady-state operation state of electrical appliances, saves calculations, reduces errors, and enhances monitoring efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115600837B_ABST
    Figure CN115600837B_ABST
Patent Text Reader

Abstract

The present invention relates to the technical field of electrical load detection, and in particular to an energy-saving diagnosis method for electrical load anomaly. When the electrical appliance is running, the method collects current and voltage within a set period to obtain active power, steady-state reactive power, and a symmetry index and area of ​​a U-I trajectory diagram; when the state of the electrical appliance operation within the set period is confirmed to be a steady-state state and it is preliminarily judged that the electrical appliance is suspected of having an abnormal load, a sensitivity index and a correlation index of each characteristic data are obtained according to characteristic data of historical electrical load states, and then a first abnormality evaluation value is obtained; a second abnormality evaluation value is obtained according to the U-I trajectory diagram; the first abnormality evaluation value and the second abnormality evaluation value are added to obtain a load abnormality index within the set period; and an electrical load result is determined by combining the load abnormality index within the set period with the average value between the load abnormality index within the adjacent historical set periods, thereby increasing the reliability of electrical load abnormality detection.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of electrical load detection, and in particular to an energy-saving diagnosis method for electrical load abnormality. Background Art

[0002] Non-intrusive load monitoring (NILM) has the advantages of low installation cost and high user acceptance compared to intrusive monitoring. NILM mainly includes three steps: event detection, feature extraction and load identification. Most of the non-intrusive load monitoring in the prior art uses data collection, data processing, feature extraction, and inputs the extracted features into the prediction neural network of the corresponding equipment to identify the abnormal load of the equipment. However, this method has limitations and cannot cope with the replacement and transformation of complex equipment. When new equipment appears, the corresponding prediction neural network is not well trained, and the abnormal load of the equipment cannot be identified in time and accurately. Summary of the invention

[0003] In order to solve the above technical problems, the purpose of the present invention is to provide an energy-saving diagnosis method for abnormal electrical loads, and the technical solution adopted is as follows:

[0004] An embodiment of the present invention provides an energy-saving diagnosis method for abnormal electrical load, the method comprising:

[0005] When the appliance is running, the current and voltage at each sampling moment in the set period are collected, and the active power and steady-state reactive power in the set period are calculated by combining the current and voltage; and a UI trajectory diagram is constructed with the voltage as the horizontal coordinate and the current as the vertical coordinate; and the symmetry index and area of ​​the UI trajectory in the UI trajectory diagram are obtained;

[0006] When the state of the electrical appliance operation within the set period is confirmed to be a steady state, a preliminary judgment is made on the electrical appliance load abnormality based on the current collected within the set period; when it is preliminarily judged that the electrical appliance load is suspected to be abnormal, the sensitivity index and correlation index of each characteristic data are obtained respectively according to the difference between each characteristic data under the historical electrical appliance load abnormal state and the corresponding characteristic data under the electrical appliance load normal state, and the characteristic data include active power, steady-state reactive power, symmetry index and area; the absolute value of the difference between each characteristic data within the set period and the corresponding characteristic data under the electrical appliance load normal state is calculated, and the product of the sensitivity index, correlation index and absolute value of the difference of each characteristic data within the set period is added to obtain the first abnormal evaluation value;

[0007] The improved shape context algorithm is used to calculate the similarity between the UI trajectory graph and the standard UI trajectory graph, and the similarity is used as the second abnormality evaluation value; the first abnormality evaluation value and the second abnormality evaluation value are added to obtain the load abnormality index within the set period; the average value between the load abnormality index within the set period and the load abnormality index in the adjacent historical set period is calculated as the final evaluation index, and the electrical load result within the set period is determined according to the final evaluation index.

[0008] Furthermore, the method for calculating the similarity between the UI trajectory graph and the standard UI trajectory graph using the improved shape context algorithm includes:

[0009] Perform contour extraction and contour point sampling on the UI trajectory in the UI trajectory graph and the standard UI trajectory graph respectively;

[0010] For each contour point, a polar logarithmic coordinate system centered on the point is constructed. The polar logarithmic coordinate system includes a set number of regions, and the contour points around the point are mapped into each region.

[0011] Randomly select a contour point in the UI trajectory graph, based on the polar logarithm coordinate system of the contour point, obtain the minimum distance from the contour point in each area in the polar logarithm coordinate system to the edge of the corresponding area, calculate the average distance of the minimum distance in each area according to the minimum distance of all contour points in the same area, substitute the opposite of the average distance into the exponential function with the natural constant e as the base to obtain the exponential function result, and take the difference between the constant 1 and the exponential function result as the fuzzy value of the corresponding area; count the total number of first contour points contained in each area of ​​the contour point, randomly select any contour point in the standard UI trajectory graph as the target contour point, count the total number of second contour points contained in each area of ​​the target contour point, calculate the difference between the total number of first contour points and the total number of second contour points between the current area of ​​the contour point and the area corresponding to the same position of the target contour point, multiply the result of subtracting the fuzzy value of the current area from the constant 1 by the difference to obtain the product of the current area, and take the sum of the products of all areas of the contour point as the similarity value between the contour point and the target contour point;

[0012] The similarity value between any two contour points between the UI trajectory map and the standard UI trajectory map is obtained to form a description matrix of size m*n, where m is the number of contour points in the UI trajectory map and n is the number of contour points in the standard UI trajectory map. Based on the description matrix, the contour points in the UI trajectory map and the standard UI trajectory map are matched using the Hungarian algorithm to obtain the similarity value of each pair of contour points after matching. The average of the similarity values ​​is calculated based on the similarity value of each pair of contour points after matching as the similarity between the UI trajectory map and the standard UI trajectory map.

[0013] Furthermore, the method for making a preliminary judgment on the abnormality of the electrical load according to the current collected within a set period includes:

[0014] The historical current mean is calculated based on the historical current during the operation of the electrical appliance without load abnormality in a steady state; the historical load abnormality index of the previous set period before the set period is obtained, the absolute value of the difference between the mean current in the set period and the historical current mean is calculated, the opposite of the absolute value of the difference is substituted into the exponential function with the natural constant e as the base to obtain the first result, the difference between the constant 1 and the first result is calculated, and the difference is multiplied by the historical load abnormality index to obtain the evaluation index of the suspected load abnormality of the electrical appliance in the set period; an evaluation threshold is set, and when the evaluation index is greater than the evaluation threshold, it is preliminarily confirmed that the electrical appliance in the set period has a suspected load abnormality.

[0015] Furthermore, the method for obtaining the sensitivity index includes:

[0016] Obtain each characteristic data under M historical electrical load abnormal states and each characteristic data under electrical load normal states, where M is a positive integer; calculate the difference between any characteristic data under each historical electrical load abnormal state and the corresponding characteristic data under electrical load normal states, obtain M differences of the characteristic data, and calculate the sum of the M differences as a sensitivity index of the characteristic data.

[0017] Furthermore, the method for obtaining the correlation index includes:

[0018] Obtain each characteristic data under M historical electrical load abnormal states and each characteristic data under electrical load normal states, where M is a positive integer; respectively calculate the difference between each characteristic data under each historical electrical load abnormal state and the corresponding characteristic data under electrical load normal states, and obtain the difference between the corresponding characteristic data under the corresponding historical electrical load abnormal state;

[0019] Calculate the absolute values ​​of the differences between the difference of the u-th characteristic data under the current historical electrical load abnormality state and the differences of other characteristic data, add up all the absolute values ​​of the differences to obtain the deviation value of the u-th characteristic data under the current historical electrical load abnormality state; add up the deviation values ​​of the u-th characteristic data under M historical electrical load abnormalities and calculate the mean, and use the mean as the correlation index between the u-th characteristic data and other characteristic data.

[0020] Furthermore, the method for confirming the operating status of the electrical appliance within the set period includes:

[0021] The sliding window double-sided CUSUM detection algorithm is used to confirm the operating status of the appliance based on the current within the set period.

[0022] Furthermore, the method for determining the electrical load result within a set period according to the final evaluation index includes:

[0023] Set a threshold. When the final evaluation index is greater than the threshold, it is confirmed that the electrical load within the set period is abnormal.

[0024] The present invention has the following beneficial effects: when the electrical appliance is running, the present invention collects the current and voltage within a set period, and obtains the active power, steady-state reactive power and UI trajectory diagram to more intuitively reflect the operation characteristics of the electrical appliance. In order to quantify the trajectory characteristics of the UI trajectory diagram, the symmetry index of the UI trajectory in the UI trajectory diagram and the area of ​​the enclosed region are obtained, so as to obtain various characteristic data of the electrical appliance within the set period; in order to ensure that the abnormal load detection of the electrical appliance is carried out under the steady-state operation state of the electrical appliance, when it is confirmed that the operation state of the electrical appliance within the set period is a steady-state state, a preliminary judgment is made on the load abnormality of the electrical appliance based on the current change, so as to save the amount of calculation and improve the monitoring efficiency; when it is preliminarily judged that the load of the electrical appliance is suspected to be abnormal, the correlation between various characteristic data is analyzed according to historical data, that is, according to the difference between each characteristic data under the historical abnormal load state of the electrical appliance and the corresponding characteristic data under the normal load state of the electrical appliance, the sensitivity index and correlation index of each characteristic data are respectively obtained, so as to avoid the detection result being Error, the correlation index and sensitivity index of various characteristic data under abnormal electrical load are used as the weight for judging whether the electrical load is abnormal within the set period, and then the product of the sensitivity index, correlation index and absolute value of the difference of each characteristic data within the set period is added to obtain the first abnormal evaluation value; since the UI trajectory diagram can well reflect the coupling relationship between current and voltage, the similarity between the UI trajectory diagram and the standard UI trajectory diagram is calculated as the second abnormal evaluation value for judging the electrical load abnormality, and the similarity between the UI trajectory diagram and the standard UI trajectory diagram is calculated; since the electrical load abnormality within a single period of time cannot accurately reflect whether the electrical appliance has actually been abnormal, the average of the load abnormality index within the set period and the load abnormality index within the adjacent historical set period is calculated, and the electrical load result within the set period is finally determined according to the average, thereby greatly increasing the reliability of the detection result, and avoiding detection errors caused by abnormal or inapplicable situations of any feature. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0026] Figure 1A flowchart of a method for energy-saving diagnosis of abnormal electrical load provided by an embodiment of the present invention;

[0027] Figure 2 Schematic diagram of a polar logarithmic coordinate system in an embodiment of the present invention. DETAILED DESCRIPTION

[0028] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the energy-saving diagnosis method for abnormal electrical loads proposed by the present invention, its specific implementation, structure, features and effects, in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0029] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0030] A specific scheme of an energy-saving diagnosis method for abnormal electrical load provided by the present invention is described in detail below with reference to the accompanying drawings.

[0031] See also Figure 1 , which shows a flowchart of a method for energy-saving diagnosis of abnormal electrical load provided by an embodiment of the present invention, the method comprising:

[0032] Step S001, when the appliance is running, collect the current and voltage at each sampling moment in a set period, calculate the active power and steady-state reactive power in the set period by combining the current and voltage; and construct a UI trajectory diagram with voltage as the horizontal axis and current as the vertical axis; obtain the symmetry index and area of ​​the UI trajectory in the UI trajectory diagram.

[0033] Specifically, the present invention collects the current and voltage of the electrical appliance during operation through a current sensor and a voltage sensor, and obtains the current and voltage of the electrical appliance at the corresponding sampling moment based on the sampling frequency within a set period, wherein the preferred setting period of this scheme is 1 minute.

[0034] Calculate the product of current and voltage at each sampling moment, and take the average of all products within the set period as the active power P; calculate the steady-state reactive power Q of the appliance within the set period according to the current and voltage at each sampling moment, and the calculation formula for steady-state reactive power is:

[0035]

[0036] Where N is the number of current or voltage samples collected within a set period; is the square of the kth voltage; is the square of the kth current.

[0037] It should be noted that the acquisition of active power and steady-state reactive power is a well-known technology and will not be described in detail in this solution.

[0038] The active power and steady-state reactive power obtained based on current and voltage are used as the first type of features for abnormal electrical load detection, and then a UI trajectory diagram is constructed with voltage as the horizontal axis and current as the vertical axis to reflect the coupling relationship between current and voltage. Compared with the first type of features, the UI trajectory diagram has better distinguishability and can better solve the problem of feature overlap. In order to quantify the trajectory features of the UI trajectory diagram, the symmetry index of the UI trajectory and the area of ​​the enclosed area in the UI trajectory diagram are obtained.

[0039] The method for obtaining the symmetry index is as follows: rotate the UI trajectory in the UI trajectory graph 180 degrees, calculate the contour similarity value of the UI trajectory after rotation and the UI trajectory before rotation using the shape context algorithm, and use the contour similarity value as the symmetry index. The larger the contour similarity value, the greater the symmetry of the UI trajectory graph. The area acquisition method and the shape context algorithm are both well-known technologies and will not be repeated in this solution.

[0040] Step S002, when the state of operation of the electrical appliance within the set period is confirmed to be a steady state, a preliminary judgment is made on the abnormality of the electrical load based on the current collected within the set period; when it is preliminarily judged that the electrical appliance load is suspected to be abnormal, a sensitivity index and a correlation index of each characteristic data are obtained respectively according to the difference between each characteristic data under the historical abnormal electrical load state and the corresponding characteristic data under the normal electrical load state, and the characteristic data include active power, steady-state reactive power, symmetry index and area; the absolute value of the difference between each characteristic data within the set period and the corresponding characteristic data under the normal electrical load state is calculated, and the product of the sensitivity index, the correlation index and the absolute value of the difference of each characteristic data within the set period is added to obtain a first abnormal evaluation value.

[0041] Specifically, since transient states may occur in electrical appliances during the state conversion process, when performing electrical load abnormality detection, it is necessary to analyze the operating state of the electrical appliance within a set period to determine whether it is in a steady state or a transient state, so as to ensure that the electrical load abnormality detection is performed under the steady-state operating state of the electrical appliance to avoid errors in the detection results. Therefore, the present invention uses a sliding window bilateral CUSUM detection algorithm to determine the operating state of the electrical appliance based on the current within the set period, wherein the method of detecting steady-state characteristics using the sliding window bilateral CUSUM detection algorithm is a well-known technology and will not be repeated in this solution.

[0042] When the operating state of the electrical appliance within the set period is confirmed to be a steady state, since current is the most basic descriptive indicator of the operating state of the electrical appliance, when performing load abnormality monitoring, in order to save calculation amount and improve monitoring efficiency, a preliminary judgment is first made on the load abnormality of the electrical appliance based on the current change. When the preliminary judgment is that the load of the electrical appliance is abnormal, the detection result of the load abnormality of the electrical appliance is further determined based on the data collected within the set period.

[0043] Among them, the method for making a preliminary judgment on the load abnormality of the electrical appliance based on the current change is: calculate the historical current average according to the historical current during the operation of the electrical appliance without load abnormality in a steady state; obtain the historical load abnormality index of the previous set period before the set period, calculate the absolute value of the difference between the mean current in the set period and the historical current average, substitute the opposite of the absolute value of the difference into the exponential function with the natural constant e as the base to obtain the first result, calculate the difference between the constant 1 and the first result, multiply the difference by the historical load abnormality index to obtain an evaluation index of the suspected load abnormality of the electrical appliance in the set period; set an evaluation threshold, when the evaluation index is greater than the evaluation threshold, it means that the electrical appliance is suspected to be in a load abnormal state in the set period, and it is confirmed that the load abnormality of the electrical appliance in the set period needs to be further confirmed.

[0044] Preferably, the evaluation threshold in the embodiment of the present invention takes an empirical value of 0.7.

[0045] As an example, the calculation formula of the evaluation index E of suspected abnormal load of electrical appliances is:

[0046] E=R′*[1-exp(-|Ia|)]

[0047] Among them, R' is the historical load abnormality index; exp is an exponential function with the natural constant e as the base; I is the mean current in the set period; a is the historical current mean; |*| is the absolute value function.

[0048] Since the historical load abnormality index R' reflects the load abnormality of the previous set cycle before the set cycle, and the time between the previous set cycle and the set cycle is adjacent, when the load abnormality of the appliance occurs in the previous set cycle, it is very likely that the load is abnormal in the next adjacent set cycle. Therefore, the larger the historical load abnormality index is, the larger the evaluation index is, and there is a positive correlation between the two. Since the historical current mean is the current index under normal load conditions of the appliance, the greater the difference between the current mean in the set cycle and the historical current mean, the more abnormal the operating state of the appliance in the set cycle is, and the larger the corresponding evaluation index is, and there is a positive correlation between the two.

[0049] After confirming that the electrical appliances within the set period are suspected to be in an abnormal load state, the correlation between various characteristic data is analyzed based on historical data, where the characteristic data includes active power, steady-state reactive power, symmetry index and area.

[0050] Obtain each characteristic data under M historical electrical load abnormal states and each characteristic data under electrical load normal states, where M is a positive integer; respectively calculate the difference between any characteristic data under each historical electrical load abnormal state and the corresponding characteristic data under electrical load normal states, obtain M differences of the characteristic data, and calculate the sum of the M differences as a sensitivity index of the characteristic data. The larger the difference between the characteristic data under the electrical load abnormal state and the characteristic data under the electrical load normal state, the more accurate the judgment of the corresponding characteristic data on the electrical load abnormal state is, and the larger the corresponding sensitivity index is, so the sensitivity index of each characteristic data can be obtained.

[0051] Since the abnormality of electrical load often affects the acquired characteristic data, the correlation index between each characteristic data can be analyzed based on each characteristic data under M historical electrical load abnormal states: obtain each characteristic data under M historical electrical load abnormal states and each characteristic data under electrical load normal states; calculate the difference between each characteristic data under each historical electrical load abnormal state and the corresponding characteristic data under electrical load normal state, and obtain the difference of the corresponding characteristic data under the corresponding historical electrical load abnormal state; calculate the absolute value of the difference between the difference of the u-th characteristic data under the current historical electrical load abnormal state and the difference of other characteristic data, and add all the absolute values ​​of the difference to obtain the deviation value of the u-th characteristic data under the current historical electrical load abnormal state; add the deviation values ​​of the u-th characteristic data under the M historical electrical load abnormal states and calculate the mean, and use the mean as the correlation index between the u-th characteristic data and other characteristic data; similarly, obtain the correlation index between each characteristic data and other characteristic data.

[0052] As an example,

[0053]

[0054] Among them, W u is the correlation index of the u-th feature data; Δf j,u is the difference between the uth characteristic data in the jth historical electrical load abnormal state and the corresponding characteristic data in the electrical load normal state; Δf j,v is the difference between the vth characteristic data under the jth historical electrical load abnormal state and the corresponding characteristic data under the electrical load normal state; A is the number of types of other characteristic data except the uth characteristic data.

[0055] Based on the difference between the corresponding same kind of characteristic data under the historical electrical load abnormal state and the electrical load normal state, the difference between the differences of each characteristic data is used to analyze the influence of other characteristic data on one of the characteristic data when confirming the electrical load abnormal state, so as to reflect the correlation between the characteristic data. In order to more accurately analyze the correlation between the characteristic data, the characteristic data under multiple historical electrical load abnormal states are used to calculate the mean, and the mean is used as the correlation indicator between each characteristic data and other characteristic data.

[0056] The correlation index and sensitivity index of various characteristic data under abnormal load of electrical appliances are used as weights for judging whether the load of the electrical appliances is abnormal within a set period to improve the judgment accuracy. Therefore, the absolute value of the difference between each characteristic data within the set period and the corresponding characteristic data under normal load of the electrical appliances is first calculated, and the product of the sensitivity index, correlation index and absolute value of the difference of each characteristic data within the set period is added to obtain the first abnormal evaluation value.

[0057] As an example, the first abnormal evaluation value R 1 The calculation formula is:

[0058]

[0059] Among them, H u is the sensitivity index corresponding to the u-th characteristic data in the set period; W u is the correlation index corresponding to the u-th characteristic data within the set period; L u is the absolute value of the difference corresponding to the u-th characteristic data within the set period; v is the number of characteristic data.

[0060] When an abnormal electrical load is detected, the corresponding characteristic data is more sensitive when the load is abnormal, that is, the sensitivity index H u The larger the value is, the closer the change trend is to other characteristic data when load abnormality occurs, that is, the correlation index W u The larger the absolute value of the difference between the characteristic data and the corresponding characteristic data under normal load state of the electrical appliance, the more it can be explained that the load of the electrical appliance is abnormal, and the larger the first abnormality evaluation value obtained by combining the sensitivity index, correlation index and the product of the absolute value of the difference of each characteristic data.

[0061] Step S003, using the improved shape context algorithm to calculate the similarity between the UI trajectory diagram and the standard UI trajectory diagram, and taking the similarity as the second abnormality evaluation value; calculating the similarity between the UI trajectory diagram and the standard UI trajectory diagram; calculating the average between the load abnormality index within the set period and the load abnormality index within the adjacent historical set period, and finally determining the electrical load result within the set period according to the average.

[0062] Specifically, according to the UI trajectory diagram within the set period, a standard UI trajectory diagram under normal load operation of the appliance is obtained, and the UI trajectory diagram is compared with the standard UI trajectory diagram to obtain similarity, wherein the comparison is mainly based on the UI trajectory in the diagram.

[0063] Preferably, the present invention uses a shape context algorithm to obtain similarity.

[0064] When the shape context algorithm is used to compare the UI trajectory graph with the standard UI trajectory graph, the current and voltage fluctuate, which will lead to a low comparison result when the UI trajectory graph is compared for similarity. Therefore, in order to improve the accuracy of the comparison result, the present invention improves the shape context algorithm and designs a fuzzy shape context algorithm based on the shape context algorithm, which is as follows:

[0065] 1. Perform contour extraction and contour point sampling on the UI trajectories in the UI trajectory graph and the standard UI trajectory graph respectively.

[0066] 2. For each contour point, a polar logarithmic coordinate system centered on it is constructed. The polar logarithmic coordinate system includes 12 angle regions and 5 distance regions, such as Figure 2 As shown, the contour points around it are mapped into each area.

[0067] 3. Taking a contour point in the UI trajectory diagram as an example, based on the polar logarithmic coordinate system of the contour point, obtain the minimum distance from the contour point in each area in the polar logarithmic coordinate system to the edge of the corresponding area, calculate the average distance of the minimum distance in each area according to the minimum distance of all contour points in the same area, substitute the opposite of the average distance into the exponential function with the natural constant e as the base to obtain the exponential function result, and take the difference between the constant 1 and the exponential function result as the fuzzy value of the corresponding area; count the total number of first contour points contained in each area of ​​the contour point, and randomly select any contour point in the standard UI trajectory diagram as the target contour point. Similarly, count the total number of second contour points contained in each area corresponding to the target contour point, calculate the difference between the total number of first contour points and the total number of second contour points between the current area of ​​the contour point and the area at the same position corresponding to the target contour point, multiply the result of subtracting the fuzzy value of the current area by the difference from the constant 1 to obtain the product of the current area, and add the products of all areas of the contour point as the similarity value between the contour point and the target contour point.

[0068] As an example, the calculation formula of the fuzzy value F is:

[0069]

[0070] Where, exp is an exponential function with the natural constant e as the base; d i is the minimum distance corresponding to the i-th contour point in any region; p is the number of contour points in any region.

[0071] The farther the contour point is from the edge of the area, the smaller the possibility of the position change of the contour point in the adjacent contour is, and the larger the fuzzy value of the corresponding area is. The larger the fuzzy value is, the greater the change of the contour point in the corresponding area is. The fuzzy value calculated by the average value of the minimum distance of all contour points in the area can better reflect the fuzziness of the corresponding area.

[0072] As an example, the similarity value is calculated as:

[0073]

[0074] Where J is the similarity value; Z is the number of regions corresponding to the contour points, in this solution, Z = 60; F j is the fuzzy value of the jth region; Δh j is the difference in the total number of contour points between the jth region and the corresponding region of the target contour point.

[0075] Since each area of ​​each contour point in the standard UI trajectory map does not have blur, that is, there is no blur value, the larger the blur value, the blurrier the corresponding area, the greater the difference between the target contour point in the standard UI trajectory map, the more dissimilar the two corresponding contour points are, and the smaller the corresponding similarity value is; at the same time, the greater the difference in the total number of contour pixels contained between the corresponding areas, the more dissimilar the two corresponding areas are, the more dissimilar the corresponding contour points are, and the smaller the similarity value is. Therefore, the difference between all areas of any contour point in the UI trajectory map and the target contour point in the standard UI trajectory map is combined to obtain the similarity value between the corresponding contour point in the UI trajectory map and the target contour point. The greater the difference, the smaller the similarity value.

[0076] 4. Use the method in step (3) to obtain the similarity between any two contour points in the UI trajectory map and the standard UI trajectory map, and then form a description matrix of size m*n, where m is the number of contour points in the UI trajectory map and n is the number of contour points in the standard UI trajectory map. Based on the description matrix, the Hungarian algorithm is used to match the contour points in the UI trajectory map and the standard UI trajectory map, and then the similarity value of each pair of contour points after matching is obtained. The average of the similarity values ​​is calculated based on the similarity value of each pair of contour points after matching as the similarity between the UI trajectory map and the standard UI trajectory map.

[0077] Based on the improved shape context algorithm, the similarity between the UI trajectory map and the standard UI trajectory map is obtained, and the similarity is used as the second abnormal evaluation value of the electrical load abnormality within the set period. The first abnormal evaluation value and the second abnormal evaluation value are added to obtain the load abnormality index R within the set period.

[0078] Since the electrical load abnormality may not be an abnormality at only a certain moment, it is necessary to continuously analyze the continuously acquired current and voltage characteristics to determine whether the electrical load abnormality has occurred. Otherwise, it may be considered that the detected abnormality may be caused by noise factors. Therefore, the abnormality in the adjacent set period of the set period is analyzed in combination: the load abnormality index in the t adjacent historical set periods before the set period is obtained, t is a positive integer, and the average of the load abnormality index in the t adjacent historical set periods and the load abnormality index in the set period is calculated as the final evaluation index of the set period. The calculation formula for the final evaluation index is:

[0079]

[0080] Among them, R' c is the load anomaly index within the cth adjacent historical setting cycle; T is the final evaluation index of the setting cycle; R is the load anomaly index within the setting cycle; t is the number of adjacent historical setting cycles.

[0081] Preferably, in the present invention, t is set empirically, and the empirical value t=5 is taken, that is, the electrical appliance operating status in the set period and the previous 5 historical set periods is referred to to judge whether the electrical appliance load in the set period is abnormal.

[0082] It should be noted that when an abnormal load occurs in an electrical appliance in an adjacent set period before the set period, it will also affect the operating status of the electrical appliance in the set period, and will also cause the electrical appliance in the set period to be in an abnormal load state to a large extent. Therefore, calculating the average of the load abnormality index in the adjacent historical set period and the load abnormality index in the set period can more accurately evaluate the load situation of the electrical appliances in the set period. The larger the average value, the more likely it is that the electrical appliance load is abnormal.

[0083] A threshold is set, and the preferred threshold in the present invention is 0.7. When the final evaluation index is greater than the threshold, it is confirmed that the load of the electrical appliance has an abnormal load within the set period. Since the load of the electrical appliance is abnormal, the electrical appliance will cause unnecessary energy waste. Therefore, when it is confirmed that the load of the electrical appliance is abnormal, the operator is notified in time to detect the electrical appliance to reduce unnecessary power use, greatly save the energy consumed by the electrical appliance, and thus achieve the purpose of energy saving.

[0084] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0085] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0086] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An energy-saving diagnosis method for abnormal electrical load, characterized in that: The method comprises the following steps: When the appliance is running, the current and voltage at each sampling moment within a set period are collected, and the active power and steady-state reactive power within the set period are calculated by combining the current and voltage; a UI trajectory graph is constructed with voltage as the horizontal axis and current as the vertical axis; and the symmetry index and area of ​​the UI trajectory in the UI trajectory graph are obtained; When the state of operation of the electrical appliance within the set period is confirmed to be a steady state, a preliminary judgment is made on the abnormality of the electrical load based on the current collected within the set period; when it is preliminarily judged that the electrical appliance load is suspected to be abnormal, the sensitivity index and correlation index of each characteristic data are obtained respectively based on the difference between each characteristic data under the historical abnormal electrical load state and the corresponding characteristic data under the normal electrical load state, and the characteristic data include active power, steady-state reactive power, symmetry index and area; the absolute value of the difference between each characteristic data within the set period and the corresponding characteristic data under the normal electrical load state is calculated, and the product of the sensitivity index, correlation index and absolute value of the difference of each characteristic data within the set period is added to obtain a first abnormality evaluation value; The improved shape context algorithm is used to calculate the similarity between the UI trajectory graph and the standard UI trajectory graph, and the similarity is used as the second anomaly evaluation value. The first anomaly evaluation value and the second anomaly evaluation value are added together to obtain the load anomaly index within the set period. The average of the load anomaly index within the set period and the load anomaly index within the adjacent historical set periods is calculated as the final evaluation index, and the electrical load result within the set period is determined based on the final evaluation index. The method for calculating the similarity between the UI trajectory graph and the standard UI trajectory graph using the improved shape context algorithm includes: Perform contour extraction and contour point sampling on the UI trajectory in the UI trajectory graph and the standard UI trajectory graph respectively; For each contour point, a polar logarithmic coordinate system is constructed with it as the center. The polar logarithmic coordinate system includes a set number of regions, and the contour points around it are mapped into each region; Randomly select a contour point in the UI trajectory graph, and based on the pole logarithm coordinate system of the contour point, obtain the minimum distance from the contour point in each area in the pole logarithm coordinate system to the edge of the corresponding area, calculate the average distance of the minimum distance in each area based on the minimum distance of all contour points in the same area, substitute the opposite of the average distance into the exponential function with the natural constant e as the base to obtain the exponential function result, and subtract the constant 1 from the exponential function result as the fuzzy value of the corresponding area; count the total number of first contour points contained in each area of ​​the contour point, randomly select any contour point in the standard UI trajectory graph as the target contour point, count the total number of second contour points contained in each area of ​​the target contour point, calculate the difference between the total number of first contour points and the total number of second contour points between the current area of ​​the contour point and the area corresponding to the same position of the target contour point, multiply the result of subtracting the fuzzy value of the current area by the difference from the constant 1 to obtain the product of the current area, and add the products of all areas of the contour point as the similarity value between the contour point and the target contour point; Obtain the similarity between any two contour points in the UI trajectory map and the standard UI trajectory map to form an m*n description matrix, where m is the number of contour points in the UI trajectory map and n is the number of contour points in the standard UI trajectory map. Based on the description matrix, use the Hungarian algorithm to match the contour points in the UI trajectory map and the standard UI trajectory map to obtain the similarity value of each pair of contour points after matching. Calculate the average of the similarity values ​​based on the similarity value of each pair of contour points after matching as the similarity between the UI trajectory map and the standard UI trajectory map. The method for obtaining the sensitivity index includes: Obtain each type of characteristic data under M historical electrical load abnormal conditions and each type of characteristic data under M historical electrical load normal conditions, where M is a positive integer; calculate the difference between any type of characteristic data under each historical electrical load abnormal condition and the corresponding characteristic data under each historical electrical load normal condition, obtain M differences of the characteristic data, and calculate the sum of the M differences as a sensitivity index of the characteristic data; The method for obtaining the correlation index includes: Obtain each characteristic data under M historical electrical load abnormal conditions and each characteristic data under M historical electrical load normal conditions, where M is a positive integer; calculate the difference between each characteristic data under each historical electrical load abnormal condition and the corresponding characteristic data under each historical electrical load normal condition, and obtain the difference between the corresponding characteristic data under the corresponding historical electrical load abnormal condition; Calculate the absolute values ​​of the differences between the difference of the u-th characteristic data under the current historical electrical load abnormality state and the differences of other characteristic data, add up all the absolute values ​​of the differences to obtain the deviation value of the u-th characteristic data under the current historical electrical load abnormality state; add up the deviation values ​​of the u-th characteristic data under M historical electrical load abnormalities and calculate the mean, and use the mean as the correlation index between the u-th characteristic data and other characteristic data.

2. The energy-saving diagnosis method for abnormal electrical load according to claim 1, characterized in that: The method for preliminarily judging electrical load abnormality based on current collected within a set period includes: The historical current mean is calculated based on the historical current during the operation of the electrical appliance in a steady state without load abnormality; the historical load abnormality index of the previous set period before the set period is obtained, the absolute value of the difference between the mean current in the set period and the historical current mean is calculated, the opposite of the absolute value of the difference is substituted into the exponential function with the natural constant e as the base to obtain the first result, the difference between the constant 1 and the first result is calculated, and the difference is multiplied by the historical load abnormality index to obtain an evaluation index for suspected load abnormality of the electrical appliance in the set period; an evaluation threshold is set, and when the evaluation index is greater than the evaluation threshold, it is preliminarily confirmed that the electrical appliance in the set period has suspected load abnormality.

3. The energy-saving diagnosis method for abnormal electrical load according to claim 1, characterized in that: The method for confirming the operating status of the electrical appliance within the set period includes: The operating status of the appliance is confirmed using a sliding window double-sided CUSUM detection algorithm based on the current within a set period.

4. The energy-saving diagnosis method for abnormal electrical load according to claim 1, characterized in that: The method for determining the electrical load result within a set period according to the final evaluation index includes: Set a threshold. When the final evaluation index is greater than the threshold, it is confirmed that the electrical load within the set period is abnormal.