Load identification method, computer readable storage medium, and device

Abstract

A load identification method, comprising: sampling a to-be-identified load, determining a characteristic ratio combination parameter of the to-be-identified load according to short-time load characteristic information of the to-be-identified load, and detecting a load switching event in an operation cycle of the to-be-identified load according to the characteristic ratio combination parameter. After the load switching event is acquired, a load characteristic ratio of the to-be-identified load is obtained, the to-be-identified load is identified from a known characteristic library, in a case where the to-be-identified load is not identified from the known characteristic library, event characteristic information of the to-be-identified load is sent to a server, and an identification result of the to-be-identified load fed back by the server is received.

Description

Technical Field

[0001] This application relates to the field of electrical information acquisition, and in particular to a load identification method, a computer-readable storage medium, and a device. Background Technology

[0002] The statements herein are provided only as background information in connection with this application and do not necessarily constitute exemplary technology.

[0003] Since the advent of alternating current (AC), both industrial and household electricity have been powered by AC, and most electrical equipment is also AC-powered. For power supply systems, knowing the electrical equipment connected to the grid allows for effective energy management. Especially in energy storage systems, knowing the electrical information of the equipment allows the system to pre-store appropriate amounts of electricity to meet equipment needs. It can also manage power distribution based on equipment priority when power is insufficient, providing the optimal power supply solution.

[0004] Therefore, knowing the electrical information of electrical equipment is of great significance to power systems, especially energy storage systems. Accurate identification of electrical loads can not only achieve optimal power management, improve the user experience, and reduce carbon emissions, but also effectively detect the operating status of electrical equipment, providing greater assurance for safe electricity use. Summary of the Invention

[0005] According to various embodiments of this application, a load identification method, a computer-readable storage medium, and an energy storage device are provided.

[0006] The technical solution is as follows: Firstly, a load identification method is provided, the method comprising: The load to be identified is sampled to obtain the target load characteristic information of the load to be identified; Based on the target load characteristic information of the load to be identified, the characteristic ratio combination parameter of the load to be identified is determined, wherein the characteristic ratio combination parameter includes the characteristic ratio of the first parameter and the characteristic ratio of the second parameter of the load to be identified; Based on the feature ratio of the first parameter and the feature ratio of the second parameter, load switching events in the operating cycle of the load to be identified are detected, and the load switching events are used to determine the time point when the load to be identified is turned on or off. Obtain the load characteristic ratio of the load to be identified after the load switching event; Based on the load characteristic ratio of the load to be identified and the known feature library including the feature samples of the data source of known loads, the characteristic ratio combination minimum residual optimization algorithm is used to identify the load to be identified, and the identification result of the load to be identified is obtained.

[0007] Secondly, a load identification device is provided, wherein the load identification device employs the above-described load identification method.

[0008] Thirdly, a computer-readable storage medium is provided, wherein instructions are stored in the storage medium, and when the instructions are executed, the above-described load identification method is implemented.

[0009] Fourthly, a device is provided, the energy storage device including a processor coupled to a communication interface, the processor being used to run computer programs or instructions to implement the above-described method.

[0010] Details of one or more embodiments of this application are set forth in the following drawings and description. Other features, objects, and advantages of this application will become apparent from the specification, drawings, and claims. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a flowchart of a load identification method provided in an embodiment of this application.

[0013] Figure 2 This is a flowchart of a load identification method provided in another embodiment of this application.

[0014] Figure 3 This is a flowchart of a load identification method based on a cloud-based feature library, provided in another embodiment of this application.

[0015] Figure 4 This is a flowchart of a load switching detection algorithm provided in another embodiment of this application.

[0016] Figure 5 This is a basic architecture diagram of cloud-based collaborative load identification provided in another embodiment of this application. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0018] It should be understood that "multiple" as mentioned in this application refers to two or more. In the description of this application, unless otherwise stated, " / " indicates "or," for example, A / B can mean A or B; "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist, for example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Furthermore, to facilitate a clear description of the technical solutions of this application, the terms "first," "second," etc., are used to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first," "second," etc., do not limit the quantity or execution order, and that "first," "second," etc., do not necessarily imply differences.

[0019] Before providing a detailed explanation of the embodiments of this application, the application scenarios of these embodiments will be described first.

[0020] Currently, the vast majority of electricity used in both industrial and residential applications is powered by alternating current (AC), and correspondingly, most electrical devices are also powered by AC. For power supply systems, knowing the devices connected to the grid allows for effective energy management. Especially in energy storage systems, knowing the electrical information of the devices allows the system to pre-store appropriate amounts of electricity to meet device demands, and can also manage power distribution based on device priority when power is insufficient, providing the optimal power supply solution.

[0021] Therefore, knowing the electrical information of electrical equipment is of great significance to power systems, especially energy storage systems. It not only enables optimal energy management, improves the user experience, and reduces carbon emissions, but also effectively monitors the operating status of electrical equipment, providing greater assurance for safe electricity use.

[0022] The total electrical power drawn by an electricity user's electrical equipment from the power system at a given moment is called the electrical load. During operation and switching on / off, electrical loads exhibit different operating characteristics and start-stop characteristics due to their different working principles. Measurement data of the electrical load is acquired through terminal acquisition devices installed in specific areas, allowing for effective identification and decomposition of the load. This application's embodiment achieves cloud-based collaborative load identification based on multi-type feature ratio combination optimization by actively detecting voltage or current information of the electrical load in real time. The algorithm automatically identifies the type of electrical load. The cloud-based collaborative optimization method is based on the existing terminal load identification system. For uncertain loads, a feature ratio algorithm is used to detect load event switching, extracting event feature information and uploading it to the cloud. The cloud uses historical feature data and a feature ratio combination minimum residual optimization algorithm, employing the feature ratio minimum proximity principle classification method to achieve load identification. After identification, the cloud feeds back personalized information to the terminal to assist in real-time identification.

[0023] Figure 1 This application provides a method for load identification according to an embodiment, including: Step 101: Sample the load to be identified to obtain the target load characteristic information of the load to be identified.

[0024] The terminal acquisition device obtains measurement data of the load to be identified through high-frequency sampling. Because the load to be identified exhibits different operating characteristics and start-up / shutdown characteristics due to different working principles during operation and switching off, the load to be identified is effectively identified and decomposed based on the measurement data. In this application, the load to be identified is also referred to as electrical load, power load, etc.

[0025] Step 102: Determine the characteristic ratio combination parameters of the load to be identified based on the target load characteristic information. The characteristic ratio combination parameters include the characteristic ratio of the first parameter and the characteristic ratio of the second parameter of the load to be identified. The first parameter is current, and the second parameter is active power.

[0026] Step 103: Based on the characteristic ratio of the first parameter and the characteristic ratio of the second parameter, detect the load switching events in the operating cycle of the load to be identified. The load switching events are used to determine the time point when the load to be identified is turned on or off.

[0027] The signal is defined by Formula 1. In time period Valid values: Formula 1 in, The time window length represents the sampled characteristic signal of the load to be identified, including current or active power. It is a fixed time period. Therefore, the load characteristic ratio based on switching event detection satisfies the following formula 2: Formula 2 in, This corresponds to the midpoint of the time window. From the signal The time window is moved from the starting point, and the load characteristic ratio is recorded. If a load switching event occurs, the load characteristic ratio will change, thus the switching point and the switching event can be detected.

[0028] Step 104: Obtain the load characteristic ratio of the load to be identified after the load switching event.

[0029] Step 105: Based on the load characteristic ratio of the load to be identified and the known feature library, the characteristic ratio combination minimum residual optimization algorithm is used to identify the load to be identified, and the identification result of the load to be identified is obtained. The known feature library includes data source feature samples of known loads, which will be described below.

[0030] Figure 2 This is a flowchart illustrating a load identification method provided in an embodiment of this application. The construction of the feature ratio library mainly includes collecting electrical load measurement data using a terminal, detecting load switching events based on a feature ratio algorithm, obtaining the load feature ratio after the load switching event, constructing the load feature ratio library, and generating a classified load model based on a feature ratio combination minimum residual algorithm. Specifically, real-time on-site load identification of the load to be identified mainly includes: installing a terminal on-site to collect real-time measurement data of the load to be identified; the terminal detecting switching events based on the measurement data and constructing load feature ratios; generating load types based on the algorithm model; and obtaining the identification result of the load to be identified. In this embodiment, this feature ratio library is also referred to as the known feature library or the terminal's own load feature library.

[0031] This application obtains target load characteristic information by performing high-frequency sampling on the load to be identified, and determines the characteristic ratio combination parameters based on the target load characteristic information. In this embodiment, the target load characteristic information is the short-term characteristic information of the load to be identified, hence it is also called short-term load characteristic information. In the embodiment provided by this application, load switching events are detected based on the characteristic ratio method according to the characteristic ratio combination parameters, which improves the sensitivity to prevent false detection events during the detection process. The load characteristic ratio of the load to be identified after the load switching event is obtained, and combined with a known feature library, a minimum residual optimization algorithm is used to identify the load to be identified. The known feature library includes the load characteristic ratios of one or more loads of known load types. Based on these characteristic ratios, the load type is identified. The minimum residual optimization algorithm can more accurately determine the type of the load to be identified, improving the accuracy of load identification. Specifically, the improvement in detection sensitivity is due to the load switching event detection and identification based on the characteristic ratio method provided in this application. This application uses dual index information of current and active power for identification. Load switching event checking can effectively improve the detection process and prevent false detection events, but it is not limited to this. Since the characteristic ratio participates in the identification process of load switching event checking, it brings convenience to the subsequent construction of the characteristic ratio library and the calculation of the minimum residual of characteristic ratio combination.

[0032] In one embodiment of this application, the load identification method further includes: sending event feature information of the load to be identified to a server when the load to be identified cannot be identified from a known feature library; the event feature information includes the time features and statistical features of the load to be identified; and receiving the identification result of the load to be identified from the server.

[0033] As an example, in this embodiment, the server is also referred to as the cloud or a cloud server. Server-based identification can be understood as load identification based on a cloud-based feature library. This process includes terminal load identification. First, data is collected at high frequency from the terminal. Based on the collected data, a terminal load feature ratio library is constructed. Load identification is performed using conventional methods, and the terminal load feature ratio is extracted. The load is identified for the first time. If the load type is confirmed, the terminal load feature library is expanded; if the terminal cannot confirm the load type, load identification is performed via the cloud.

[0034] In one possible implementation of this application, Figure 3 This document provides a flowchart for load identification based on a cloud-based feature library, as an embodiment of this application. Figure 3The load identification method, which combines a cloud-based feature library located on the server side, includes: the terminal collecting high-frequency measurement data of known loads; constructing a terminal load feature ratio library based on the measurement data of the known loads; when it is necessary to identify a load to be identified, collecting the target load feature information of the load to be identified, and applying the load identification method in the above embodiment to identify the load. In this embodiment, the measurement data of the known load corresponds to the target feature information of the load to be identified. When the terminal cannot determine the identification result of the load to be identified, the relevant data of the load to be identified is uploaded to the cloud, and load switching events are detected in the cloud based on the feature ratio algorithm. The cloud obtains the load feature ratio of the load to be identified after the switching event, identifies the load based on the feature ratio combined minimum residual algorithm, and finally feeds back the identification result to the terminal, thereby realizing the fine-grained identification of loads that the terminal cannot confirm in the cloud. After the cloud server feeds back the identification result to the terminal, it updates its own load feature library based on the load to be identified and its identification result. In this embodiment, the cloud server's own load feature library is also referred to as the cloud feature library.

[0035] It's worth noting that after cloud-based identification, if the terminal encounters the same load again, it still needs to rely on the cloud to prevent the inability to achieve short-term online identification. Therefore, after cloud identification, the terminal receives feedback on the specific characteristics of this load, such as the specific active step and normally-on time. The terminal then updates its own specific feature database in its load feature database, thus enabling specific identification of this load in the next load identification.

[0036] In one embodiment of this application, by actively detecting the voltage and current information of the power supply load output in real time, the function of cloud-based collaborative load identification based on multi-type feature ratio combination optimization is realized. The load identification method provided by this application embodiment can automatically identify the type of power load.

[0037] As an example, the terminal performs high-frequency measurement data sampling on the load to be identified, extracts load feature information, and implements load switching event detection based on the feature ratio algorithm to improve the sensitivity of preventing false detection events during the detection process. The lightweight feature ratio combination minimum residual optimization method is applied to identify the load and send the feature information of uncertain loads to the cloud.

[0038] The cloud-based collaborative load identification mainly involves constructing a cloud-based historical feature database containing inherent characteristics, spatiotemporal characteristics, and statistical characteristics, and establishing a cloud-based closed-loop upgrade mechanism. The cloud sends back differential feature information to improve the terminal load feature database. The terminal updates its load feature database with the differential feature information returned from the cloud, so that it can be effectively identified on the terminal the next time the load to be identified is identified, reducing the identification process of uploading to the cloud. It should be noted that in Figure 3In this process, after the terminal builds a load characteristic comparison database, it can identify the load to be identified. However, if the identification process for the load to be identified cannot be completed on the terminal, all data of the load to be identified is uploaded to the cloud. All data of the load to be identified includes target load characteristic information, etc. The cloud can then perform identification based on the data of the load to be identified.

[0039] In one embodiment of this application, obtaining the load characteristic ratio of the load to be identified after a load switching event includes: obtaining a set of load characteristic ratios of one or more loads to be identified after a load switching event.

[0040] As an example, high-frequency sampling of known loads can effectively collect load data and construct a feature sample library for electricity load data sources. A set of feature samples S from the electricity load data source consists of nine attribute variables from the measurement data and one flag variable, providing comprehensive data support for the subsequent construction of the load feature library. The data source feature samples can be expressed as follows: (Formula 3) Formula 3 In formula 3: For the first The load of the first One feature; For the first Each load type; Take the numbers 1, 2, ..., n; Take values ​​1, 2, ..., 9. Based on formula 3 above, generate a set of data source feature samples, forming test samples within each feature ratio space. For a random, individual load, the set of feature ratios obtained after the load switching event detection satisfies the following formula 4: Formula 4 in, Defined as load The Each feature ratio, This indicates the number of load characteristic ratios.

[0041] Based on the load characteristic ratios of the loads to be identified, the minimum residual optimization algorithm for characteristic ratio combination is used to identify the loads from the known characteristic library. This includes: determining a common characteristic ratio set for one or more loads to be identified based on the load characteristic ratio set; classifying the common characteristic ratio set into several subsets to obtain a common characteristic ratio subset library, where each subset includes a cluster center; constructing an objective function and constraints based on the common characteristic ratio set, subsets, and cluster centers; determining the fit between the load characteristic ratios of the loads to be identified and the cluster centers based on the objective function and constraints; determining the target subset in the common characteristic ratio subset library based on the fit, where the target subset is the subset with the highest fit between the load characteristic ratio and one or more subsets; and updating the target subset in the pre-constructed characteristic ratio subset library to obtain the updated characteristic ratio subset library. Compare the data source feature samples in the known feature library with each target subset in the feature ratio set subset library. If the data source feature samples and each target subset in the feature ratio set subset library are consistent, determine the load type of the load to be identified based on multiple flag values ​​in the data source feature samples. Each flag value corresponds to a load type of the data source feature samples.

[0042] In one possible implementation, considering the random switching of multiple loads during electrical load operation, the overall characteristic ratio can be described by the following formula 5: Formula 5 In the formula: Choose 1, 2, ..., 9. This represents the total number of loads in operation after random load switching. The set of characteristic ratios shared by each load is: The purpose of load identification is to... Classified as Subset The cluster center of each subset can be described as Then, the objective function as shown in Equation 6 and the constraints as shown in Equation 7 can be constructed: Formula 6 Formula 7 In the formula: ; The shared feature is described by the Euclidean distance between each target subset in the subset library and its corresponding cluster center. This is the clustering shrinkage coefficient, and the degree of shrinkage in the classification varies with... Increase and strengthen The value range is 1.5 to 2.5. In this embodiment, the objective function corresponding to Formula 6, which describes the constraints shown in Formula 7, is solved using Lagrange multipliers, as shown in Formula 8: Formula 8 in, ,… The Lagrange coefficients are given by equation (9). Solving for the Lagrange function mathematically yields the following formula: Formula 9 The optimal solution is: Formula 10 Eliminate using the constraints in Formula 7 Then, after transformation, the result shown in Formula 11 is obtained: Formula 11 Similarly, using Formula 8... Taking the derivative and setting it equal to 0, the cluster centers are obtained as shown in Formula 11: Formula 11 Update repeatedly using the above formula and Determine the actual conformity ratio and The fit, as shown in Formula 12: Formula 12 In the formula, For this feature ratio set subset library, for The target subset it belongs to for The degree of closeness between the middle element and its corresponding characteristic ratio. If it satisfies Formula 13: Formula 13 Then determine and closest Classified in In the middle, find The set of markers for each sample Finally, the type of the load to be identified is determined according to the principle of least proximity, and the identification result of the load to be identified is obtained.

[0043] In one embodiment of this application, such as Figure 4This is a flowchart of a load switching detection method. Specifically, firstly, measurement data of the load to be identified is collected through a terminal. The collected measurement data is then denoised, and the load characteristic ratio is calculated from the denoised measurement data. The load characteristic ratio calculation includes a first parameter and a second parameter, where the first parameter is current and the second parameter is active power. Based on the characteristic ratios of the first and second parameters, load switching events are detected during the load's operating cycle, including: determining the second norm constraint condition of the current characteristic ratio based on the current characteristic ratio, and determining the second norm constraint condition of the active power characteristic ratio based on the active power characteristic ratio.

[0044] Among them, when the characteristic ratio of the load current to the second norm constraint condition is greater than the first value, the load switching point is determined, such as... Figure 4 As shown, the first value is 1. When the active power characteristic ratio of the load exceeds the second value under the constraint condition, a load switching event is recorded, such as... Figure 4 As shown, the second value is 1. When the active power characteristic ratio of the load is less than or equal to the second value under the constraint condition, the load shedding point is determined and the occurrence of the load shedding event is recorded.

[0045] In one possible implementation, the specific calculation process is as follows: define the signal. In time period The effective value X satisfies Formula 1. The load characteristic ratio based on switching event detection can be designed as shown in Formula 2, where, This indicates either the first parameter (current) or the second parameter (active power). This is the midpoint of the time window. The time window is moved from the starting point of the signal, and the load characteristic ratio is recorded. If a load switching event occurs, the load characteristic ratio will change, thus allowing the switching point and event to be detected, and subsequent calculations to be performed based on this information.

[0046] As an example, for load characteristic information that satisfies the superposition principle, i.e., the load characteristic ratio that satisfies Formula 14, load identification can be performed by obtaining an optimal combination that minimizes the residuals of this characteristic ratio and each load characteristic ratio in the known characteristic library through an optimization algorithm. Formula 14 is as follows: Formula 14 In the formula: The feature ratio extracted when the j-th feature of the i-th event is run separately; For a moment The corresponding system characteristic ratio; To measure noise signals; For the i-th event and the j-th feature in The characteristic ratio corresponding to the time step after being added to the system. It is obvious that the transient segment used in Formula 14... This not only prevents the "curse of dimensionality" caused by selecting too small a time window to achieve high accuracy in event detection, but also avoids the defect of an overly large time window causing the event detection cut point to be indistinct.

[0047] Therefore, the eigenvector of the i-th event and j-th feature ratio of the total M loads can be simplified to the following formula 15: Formula 15 In the formula: , .

[0048] Wherein, the feature ratio vector to be decomposed is The m-th load characteristic ratio vector is Since the load switching action results in two states for the load: "on" and "off," the load feature vector can be decomposed into two sequences: "on" and "off," and then classified and sorted. Therefore, the "on" sequence is... The "resection" sequence is Based on the actual situation, all elements in the "removal" sequence are "0", while all elements in the "input" sequence are actual load characteristic ratio data. According to the load switching action, the original load characteristic ratio vector can be reordered as follows: ,at the same time and The relationship between them is as shown in Formula 16 below: Formula 16 In the formula: The transformation matrix obtained by multiplying N binary vectors of 0 and 1 achieves the transformation of the original characteristic ratio vector of the load. Perform row and column transformations to obtain The purpose is to achieve this. Based on the above analysis, the load characteristic ratio vector of the "input" sequence can be transformed into the following formula: Formula 17 In the formula, Because the load "removal" action causes the "removal" sequence to satisfy... Therefore, the above formula can be further simplified to: Formula 18 In the formula: , and The matrix element values ​​are uniformly distributed and easily known. However, in reality, the m-th load may exist. l If there are several states, then formula 18 can be rewritten as: Formula 19 The load "putting in" action in this application embodiment can be understood as "putting in" in a broad sense. For different operating states of the m-th load, formula 19 can be modified as follows: Formula 20 Therefore, the embodiments of this application are applicable to the detection of two-state or multi-state switching events. Based on this, a custom adjacency matrix is ​​introduced, and its calculation method is as follows: Formula 21 In the formula Therefore, formula 18 can be simplified to the following formula: B T CB A Formula 22 According to the theory of unsorted graph signal processing (GSP), Formula 23 In the formula: , representing the number of nodes connected to the i-th node. Let be the set of neighboring nodes of the i-th node. This is the number of all neighboring nodes connected to node i. Therefore, formula 23 can be expressed as follows in this embodiment: Formula 24 Wherein, the average value of the characteristic ratio of the m-th load, the i-th event, and the j-th characteristic is defined as Since the load is in the steady state range It can be assumed that the constant value is stable, therefore also The values ​​are very close, so subtracting formula 22 from formula 24 reveals: Formula 25 Wherein, the embodiments of this application define Let be the characteristic ratio error of the m-th load, the i-th event, and the j-th characteristic ratio. Therefore, the characteristic ratio of the m-th load, the i-th event, and the j-th characteristic ratio described by Equation 25 can be converted into the corresponding L2 norm constraint condition for the characteristic ratio error: Formula 26 Therefore, for a load with three or more operating states, assuming that the characteristic ratio error of the i-th event generated within the stable section is basically stable, then Formula 26 can be considered as the upper limit of the characteristic ratio error. That is, when multiple states exist, the characteristic ratio error during load switching operations does not exceed [a certain value]. .

[0049] In one embodiment of this application, after receiving the identification result of the load to be identified from the server, the method further includes: receiving the specific features of the load to be identified from the server. Based on the specific features of the load to be identified and the identification result, the known feature library is updated. The updated known feature library includes the specific features of the load to be identified.

[0050] In one embodiment of this application, before performing high-frequency sampling on the load to be identified to obtain target load characteristic information, the method further includes: performing high-frequency sampling on a known load to obtain typical measurement parameters of the known load, including the cycle and harmonic components of current and voltage. Based on the typical measurement parameters, basic measurement data source information is constructed and obtained through distributed computing. The basic measurement data source information includes: active power variation, reactive power variation, and 3rd and 2nd harmonic variations. Based on the basic measurement data source information, a data source feature sample library is constructed, including load characteristic ratio and load type. The data source feature sample library is used as a known feature library.

[0051] As an example, this embodiment uses a high frequency of 6.4 kHz to sample the load to be identified and obtain measurement data. In actual work, the measurement data are all discrete values, so the active power calculation formula is as shown in Formula 27: Formula 27 In formula 27, The number of power frequency cycles contained in the calculation time window of the active power sequence P; The length of the time window; The number of current and voltage sampling points contained within one power frequency cycle ( ), The sampling frequency of the current and voltage; , The voltage and current sequences are discrete. The steady-state reactive power calculation formula (28) is as follows: Formula 28 In addition, it can also obtain identifiable characteristic quantities of the start-up and shutdown characteristics of electrical loads. Therefore, relying on measurement data sources obtained through high-frequency sampling can achieve effective collection of load data and construct a known characteristic database of electrical loads in advance.

[0052] As an example, the known feature library of the constructed electrical load is of the type of short-time characteristics of the load terminal, which includes 11 features. The serial numbers of the 11 features are ON, OFF, F1, F2, F3, F4, F5, F6, F7, F8, and F9, respectively, representing load input, load disconnection, active power change when load starts, reactive power change when load starts, third harmonic change when load starts, second harmonic change when load starts, number of consecutive starts and stops within a unit window, active power change rate when load starts, reactive power change rate when load starts, third harmonic change rate when load starts, and second harmonic change rate when load starts.

[0053] In one embodiment of this application, the time features in the cloud feature library include at least one of the following: average time interval between continuous start-stop operations, load running time, and average number of load operations per day. The statistical features include at least one of the following: coefficient of variation and feature mean.

[0054] In one possible implementation, Figure 5 This paper presents a basic architecture diagram for cloud-based collaborative load identification. The terminal module in the diagram comprises four key parts: a data sampling module, a data upload module, a feature upload module, and a task time management module. It can upload key load features to the cloud. The cloud utilizes a distributed database to perform long-term feature calculations based on historical data. The cloud comprehensively identifies the load type based on temporal and statistical features in addition to inherent features, and then sends a more detailed feature description back to the terminal to assist in identification. Specifically, the basic principle of terminal-cloud collaboration for load type identification is as follows: if the terminal cannot identify the load, it uses feature upload to build a massive feature database in the cloud to assist in identification. Figure 5 As shown, the cloud database constructed in this application embodiment contains three types of features. The first type is the inherent features of the load base, such as active power, reactive power, third harmonic, and second harmonic variations at startup; the second type is the time features of the load, such as the average time interval between continuous load start-stop and running time; the third type is the statistical features of the load, such as the coefficient of variation (CV) of the fluctuation degree, which represents the state coefficient of the number of load state changes within a window.

[0055] As an example, suppose the window signal is , For the k-th measurement data, and m as the window size, the coefficient of variation is: Formula 29 In the formula, The standard deviation of the sample data. Let be the mean of the sample data. The standard deviation and mean of the sample data satisfy the following formula: Formula 30 Formula 31 It is worth noting that the preset mutation threshold is... The calculation based on the difference algorithm first involves traversing the signals within the window. to And calculate their difference sequences respectively. , Secondly, count the points of positive and negative value changes, and calculate whether the absolute value of the difference between the two regions is greater than the threshold. Finally, the number of times the threshold was broken was counted and used as the state coefficient.

[0056] The overall process of the load identification method provided in the above embodiments of this application is as follows: 1) Short-term data collection is performed on the load to be identified, which is described in this application as target load characteristic information.

[0057] 2) Calculate the characteristic ratio of the short-term data of the load to be identified to obtain the characteristic ratio combination parameters.

[0058] 3) Based on the characteristic ratio combination parameters, the load switching event is detected and identified using the characteristic ratio algorithm.

[0059] 4) The feature ratio combination minimum residual algorithm is adopted to determine the load type according to the minimum proximity principle. At the same time, the feature ratio library is constructed and the distributed database in the cloud is updated in real time.

[0060] 5) Due to the limitations of terminal memory, when the unknown type of load to be identified has long-term characteristic information (including the time characteristics and statistical characteristics mentioned above), the cloud collaborative load identification mode is enabled.

[0061] 6) Cloud-based collaborative load identification.

[0062] 7) After the cloud completes the load identification, it feeds back the identification results to the terminal and updates the terminal's known feature library.

[0063] In this embodiment, the load switching event detection algorithm is applicable to all loads that need to be detected. Based on the measurement data acquired through high-frequency sampling, after denoising processing, the load characteristic ratio algorithm is applicable to the calculation of nine characteristic ratios in the measurement data. Unknown types of loads to be identified each have their own characteristic ratio patterns. Different characteristic ratios mean that when a switching event of this unknown type of load occurs, the characteristic ratio algorithm can obtain the characteristic ratio combination of the load to be identified based on the acquired target load characteristic information. The load switching event detection algorithm only utilizes information related to current and active power from the load characteristic ratio combination.

[0064] This application provides a load identification device that employs the load identification method described above.

[0065] This application provides a computer-readable storage medium storing instructions that, when executed, implement the above-described load identification method.

[0066] This application provides a device, an energy storage device including a processor and a communication interface coupled together. The processor is used to run computer programs or instructions to implement the load identification method described above. The communication interface is used to communicate with electrical devices outside the device. Specifically, the device may be an energy storage device capable of providing electrical energy to connected electrical devices through a related output interface. In some embodiments, the device may also be a distribution box, with the electrical devices connected to it. The distribution box is capable of identifying the electrical load of the electrical devices and providing corresponding control strategies based on the identification results of the electrical load.

[0067] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above method embodiments of this application can be implemented by a computer program instructing related hardware. This computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate form. The computer-readable medium can include at least: any entity or device capable of carrying the computer program code to a photographing device / terminal device, a recording medium, a computer memory, ROM (Read-Only Memory), RAM (Random Access Memory), CD-ROM (Compact Disc Read-Only Memory), magnetic tape, floppy disk, and optical data storage devices. The computer-readable storage medium mentioned in this application can be a non-volatile storage medium; in other words, it can be a non-transient storage medium.

[0068] It should be understood that all or part of the steps of the above embodiments can be implemented by software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented in whole or in part as a computer program product. The computer program product includes one or more computer instructions. The computer instructions can be stored in the above-described computer-readable storage medium.

[0069] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0070] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0071] In the embodiments provided in this application, it should be understood that the disclosed apparatus / computer devices and methods can be implemented in other ways. For example, the apparatus / computer device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0072] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0073] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A load identification method, the method comprising: The target load characteristic information of the load to be identified is obtained by sampling the load to be identified. Based on the target load characteristic information of the load to be identified, the characteristic ratio combination parameter of the load to be identified is determined, wherein the characteristic ratio combination parameter includes the characteristic ratio of the first parameter and the characteristic ratio of the second parameter of the load to be identified; Based on the feature ratio of the first parameter and the feature ratio of the second parameter, load switching events in the operating cycle of the load to be identified are detected, and the load switching events are used to determine the time point when the load to be identified is turned on or off. Obtain the load characteristic ratio of the load to be identified after the load switching event; Based on the load characteristic ratio of the load to be identified and a known feature library including the load characteristic ratios of known loads, the characteristic ratio combination minimum residual optimization algorithm is used to identify the load to be identified, and the identification result of the load to be identified is obtained. Wherein, the first parameter is current, the second parameter is active power, and the step of detecting load switching events during the load's operating cycle based on the characteristic ratio of the first parameter and the characteristic ratio of the second parameter includes: The characteristic ratio of the current is determined according to the characteristic ratio of the current, and the characteristic ratio of the active power is determined according to the characteristic ratio of the active power. The characteristic ratio of the current refers to the ratio of the effective value of the second half of the current to the effective value of the first half of the current within a fixed length time window, and the characteristic ratio of the active power refers to the ratio of the effective value of the second half of the active power to the effective value of the first half of the active power within a fixed length time window. When the characteristic ratio of the current of the load is greater than the first value under the second constraint condition, the switching point of the load is determined. When the active power characteristic ratio of the load is greater than the second value under the second constraint condition, the load switching event is recorded. When the active power characteristic ratio of the load is less than or equal to the second value under the second constraint condition, the load shedding point is determined and the occurrence of the load shedding event is recorded.

2. The method according to claim 1, wherein, The load identification method further includes: If the load to be identified cannot be identified from the known feature library, the event feature information of the load to be identified is sent to the server. The event feature information includes the time features and statistical features of the load to be identified. Receive the identification result of the load to be identified from the server.

3. The method according to claim 1, wherein, The process of obtaining the load characteristic ratio of the load to be identified after the load switching event includes: Obtain a set of load characteristic ratios for one or more of the loads to be identified after the load switching event. The load characteristic ratio refers to the ratio of the effective value of the second half of the sampled characteristic signal of the load to be identified to the effective value of the first half within a fixed-length time window. The step of identifying the load to be identified from the known feature library using the feature ratio combination minimum residual optimization algorithm based on the load characteristic ratio of the load to be identified includes: Based on the load characteristic ratio set, determine one or more common characteristic ratio sets of the loads to be identified; The common feature ratio set is classified into several subsets to obtain a common feature ratio set subset library, and each subset includes a cluster center; Based on the set of common feature ratios, the subset, and the cluster center, construct the objective function and constraints; The fit between the load characteristic ratio of the load to be identified and the cluster center is obtained based on the objective function and the constraints. Based on the fit, a target subset is determined in the common feature ratio set subset library, wherein the target subset is the subset with the highest fit between the load feature ratio and one or more of the subsets; The target subset is updated to the pre-built feature ratio set subset library to obtain the updated feature ratio set subset library; Compare the data source feature samples in the known feature library with each target subset in the feature ratio set subset library; When the data source feature samples and each target subset in the feature ratio set subset library are consistent, the load type of the load to be identified is determined according to multiple flag values ​​in the data source feature samples, and each flag value corresponds to a load type of the data source feature samples.

4. The method according to claim 2, wherein, After receiving the identification result of the payload to be identified from the server, the process further includes: Receive the specific characteristics of the load to be identified fed back by the server; The known feature library is updated based on the specific characteristics of the load to be identified and the identification result, wherein the updated known feature library includes the specific characteristics of the load to be identified.

5. The method according to claim 1, wherein, Before sampling the load to be identified to obtain the target load characteristic information of the load to be identified, the method further includes: The characteristic samples of the data source are sampled to obtain typical measurement parameters of the known load. The typical measurement parameters include the frequency and harmonic components of the current and the frequency and harmonic components of the voltage. Based on the typical measurement parameters, basic measurement data source information is obtained through distributed computing. The basic measurement data source information includes: active power change, reactive power change, third harmonic change and second harmonic change. Based on the basic measurement data source information, a data source feature sample library is constructed, which includes load characteristic ratio and load type. The data source feature sample library is used as the known feature library.

6. The method of claim 2, wherein, The time characteristics include at least one of the following: the average time interval between continuous start-stop cycles, the load operating time, and the average number of load operations per day.

7. The method of claim 2, wherein, The statistical characteristics include at least one of the coefficient of variation and the characteristic mean.

8. A computer-readable storage medium storing instructions that, when executed, implement the load identification method as described in any one of claims 1 to 7.

9. An apparatus comprising a processor coupled to a communication interface, the processor being configured to execute a computer program or instructions to implement the load identification method as described in any one of claims 1 to 7.

Images