Non-intrusive load monitoring analysis method, device, equipment and storage medium

Abstract

The invention provides a non-intrusive load monitoring analysis method, device and equipment and a storage medium, and the method comprises the steps: respectively building a current template and a voltage template for a plurality of electric appliances in a monitoring target during normal operation; collecting power utilization data of a main switch circuit of the monitoring target; the power consumption data comprises current data and voltage data; inputting the power utilization data into a pre-trained graph neural network for monitoring every set time, and determining the electric appliance attribute of each electric appliance; and extracting the feature vectors of the electric appliances, performing feature propagation, capturing the association between the electric appliances, and determining the working state of each electric appliance. According to the invention, the problem of poor load monitoring accuracy in the prior art can be solved, and more efficient and accurate electric appliance energy consumption monitoring and decomposition can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent load monitoring, and in particular to a non-invasive load monitoring and analysis method, device, equipment and storage medium. Background Art

[0002] With the advent of the energy internet era, clean power generation, efficient distribution, and convenient electricity consumption are revolutionizing smart grids. Non-intrusive load monitoring technology, a cutting-edge technology for monitoring the operating conditions of electrical equipment, eliminates the need for near-load sensor installations. Instead, it monitors and analyzes voltage and current signals at the incoming distribution lines to capture internal load characteristics that characterize the power consumption behavior of different load types.

[0003] However, in scenarios with a large number of electrical appliances and complex load variations, existing non-invasive load detection technologies lack optimal recognition and robustness, making it difficult to extract effective features from complex load data. This is especially challenging when multiple appliances are operating simultaneously. Furthermore, transient characteristics present when appliances are started and shut down can lead to inaccurate data collection.

[0004] There is currently no effective solution to the problem of poor load monitoring accuracy in existing related technologies. Summary of the Invention

[0005] The present invention provides a non-invasive load monitoring and analysis method, device, equipment and storage medium to solve the problem of poor load monitoring accuracy in existing related technologies and achieve more efficient and accurate electrical energy consumption monitoring and decomposition.

[0006] In a first aspect, the present invention provides a non-intrusive load monitoring and analysis method, comprising: Establishing current templates and voltage templates for multiple electrical appliances within the monitoring target during normal operation; Collecting power consumption data of the main switch circuit of the monitoring target; the power consumption data includes current data and voltage data; At set intervals, the electricity usage data is input into a pre-trained graph neural network for monitoring to determine the electrical properties of each of the electrical appliances; Extracting feature vectors of the electrical appliances and performing feature propagation to capture the associations between each of the electrical appliances and determine the working status of each of the electrical appliances; Based on the working status of the electrical appliance identified at each time interval, a time period health assessment is performed on the electrical appliance to generate a health status result of the electrical appliance.

[0007] According to a non-invasive load monitoring and analysis method provided by the present invention, a current template and a voltage template are established for each of multiple electrical appliances within a monitoring target during normal operation, including: Connecting each of the electrical appliances to a power source individually to collect current data and voltage data of the electrical appliances under different working states; the working states include a transient state during intervention and a steady state during normal operation; performing normalization processing on the current data and the voltage data; extracting steady-state characteristics and transient characteristics of the current data and the voltage data, and analyzing frequency domain characteristics of the current data and the voltage data; The steady-state characteristics and transient characteristics of the current data and the voltage data are combined into a multi-dimensional feature vector to obtain a current template and a voltage template when the electrical appliance is operating normally.

[0008] According to a non-intrusive load monitoring and analysis method provided by the present invention, the steady-state characteristics include the average value, standard deviation and power of the current data and the voltage data; The transient characteristics include the rise time, fall time and peak value of the current data and the voltage data; The frequency domain characteristics include harmonic content and frequency distribution.

[0009] According to a non-invasive load monitoring and analysis method provided by the present invention, the current data and the voltage data are normalized, including: filtering the current data and the voltage data to normalize the current data and the voltage data to a unified scale.

[0010] According to a non-invasive load monitoring and analysis method provided by the present invention, the power usage data is input into a pre-trained graph neural network for monitoring at set intervals to determine the electrical properties of each electrical appliance, including: detecting transient events in changes of the current data and the voltage data through a sliding window; The electrical appliance attributes of the electrical appliance are determined by combining the transient event detection result and the time sequence as well as the current template and the voltage template of each electrical appliance; the electrical appliance attributes include the type of the electrical appliance, the start time point and the shutdown time point.

[0011] According to a non-intrusive load monitoring and analysis method provided by the present invention, the feature vectors of the electrical appliances are extracted and feature propagation is performed to capture the association between each of the electrical appliances and determine the working status of each of the electrical appliances, including: Assigning an initial feature vector to each node corresponding to the electrical appliance; Feature propagation is performed through the graph neural network, and feature updates are performed based on neighboring nodes of the node corresponding to each of the electrical appliances to determine the working status of each of the electrical appliances.

[0012] According to a non-intrusive load monitoring and analysis method provided by the present invention, an initial feature vector is assigned to each node corresponding to the electrical appliance, including: For each of the electrical appliances, feature extraction is performed on the current template and voltage template of the electrical appliance to obtain template features; An initial feature vector is determined based on the template feature of the electrical appliance, and the initial feature vector is allocated to a node corresponding to the electrical appliance.

[0013] According to a non-intrusive load monitoring and analysis method provided by the present invention, based on the working status of the electrical appliance identified at each time interval, a time period health assessment is performed on the electrical appliance to generate a health status result of the electrical appliance, including: Extracting features of the working state of the electrical appliance to obtain working state features; combining the working status characteristics of the appliance with predefined evaluation rules to score the health status of the appliance; A health status result of the electrical appliances is determined based on the health status score of each of the electrical appliances.

[0014] In a second aspect, the present invention further provides a non-intrusive load monitoring and analysis device, comprising: A construction module is used to establish a current template and a voltage template for each of the multiple electrical appliances in the monitoring target during normal operation; An acquisition module, configured to acquire power consumption data of the main switch circuit of the monitoring target; the power consumption data includes current data and voltage data; A monitoring module, configured to input the electricity usage data into a pre-trained graph neural network for monitoring at set intervals to determine the electrical properties of each of the electrical appliances; An updating module, configured to extract feature vectors of the electrical appliances and perform feature propagation, capture the associations between each of the electrical appliances, and determine the working status of each of the electrical appliances; The evaluation module is configured to perform a time period health evaluation on the electrical appliance based on the working status of the electrical appliance identified at each time interval, and generate a health status result of the electrical appliance.

[0015] In a third aspect, the present invention further provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the non-invasive load monitoring and analysis method as described in the first aspect above is implemented.

[0016] In a fourth aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the non-intrusive load monitoring and analysis method as described in the first aspect above.

[0017] In a fifth aspect, the present invention further provides a computer program product, comprising a computer program, which, when executed by a processor, implements the non-intrusive load monitoring and analysis method as described in the first aspect above.

[0018] Compared with the prior art, the present invention has the following beneficial effects: The non-invasive load monitoring and analysis method provided by the present invention addresses the problem of data collection errors caused by transient characteristics when electrical appliances are started and shut down. By integrating the characteristics of graph neural networks, it is possible to more accurately capture key features in load data, enhance the processing capabilities of transient signals, and thus improve the accuracy of data collection. In addition, by introducing graph neural networks to model the complex relationships and interactive characteristics between electrical appliances, the recognition rate of load decomposition can be improved in scenarios with diverse electrical appliance types and complex load changes, and it has stronger robustness and adaptability, and can cope with complex changes in actual scenarios. By utilizing the advantages of graph neural networks in processing high-dimensional and unstructured data, more effective electrical appliance features can be extracted from the composite data of complex electrical appliances. When multiple electrical appliances are running at the same time, the energy consumption characteristics of each device can also be efficiently distinguished to achieve accurate energy consumption decomposition. Through the above process, the problem of poor load monitoring accuracy in existing related technologies can be solved, and more efficient and accurate electrical appliance energy consumption monitoring and decomposition can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0020] Figure 1 is a flow chart of the non-intrusive load monitoring and analysis method provided by the present invention; Figure 2 This is a structural block diagram of the non-intrusive load monitoring and analysis device provided by the present invention; Figure 3 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0021] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0022] The present invention provides a non-invasive load monitoring and analysis method. Figure 1 This is a flow chart of the non-invasive load monitoring and analysis method provided by the present invention, such as Figure 1 As shown, the method includes the following steps: Step S101, establishing current templates and voltage templates for each of the multiple electrical appliances in the monitoring target during normal operation; Step S102: collecting power consumption data of the main switch circuit of the monitoring target; the power consumption data includes current data and voltage data; Step S103: Inputting the electricity usage data into the pre-trained graph neural network for monitoring at set intervals to determine the electrical attributes of each appliance; Step S104: extracting the feature vectors of the appliances and performing feature propagation to capture the associations between each appliance and determine the operating status of each appliance. This includes: assigning an initial feature vector to each appliance node; performing feature propagation through a graph neural network, updating features based on the neighboring nodes of each appliance node, and determining the operating status of each appliance. Step S105 , based on the working status of the electrical appliance identified at each time interval, a time period health assessment is performed on the electrical appliance to generate a health status result of the electrical appliance.

[0023] In this method, the monitoring target can be a household, company, or other organization. First, current and voltage templates are constructed for multiple appliances within the monitoring target during normal operation, providing basic data for subsequent graph neural network (GNN) feature identification. Next, electricity usage data for the target's main switching circuit is collected to generate a time-series electricity data packet. Each appliance is considered a node in the graph, characterized by its static properties (such as appliance type and power level) and dynamic properties (such as transient characteristics). Edges in the graph represent connections between appliances, and edge weights represent the mutual influence between appliances. At set intervals, the collected electricity usage data is input into the GNN to determine the appliance attributes of each appliance. Feature vectors are then extracted and, based on the associations between appliances, the features of the corresponding nodes are updated to determine the operating status of each appliance. Finally, the health status of each appliance is assessed based on the identified operating status at each time interval, resulting in a health status result.

[0024] In this process, by integrating the characteristics of graph neural networks (GNNs), we can more accurately capture key features in load data, enhance the ability to process transient signals, and thus improve data collection accuracy, addressing data collection errors caused by transient characteristics during appliance startup and shutdown. Furthermore, by introducing GNNs to model the complex relationships and interactions between appliances, we can improve the recognition rate of load decomposition in scenarios with diverse appliance types and complex load variations. This also provides greater robustness and adaptability to handle the complex changes in real-world scenarios. By leveraging the advantages of GNNs in processing high-dimensional and unstructured data, we can extract more effective appliance features from complex composite data of appliances. Even when multiple appliances are operating simultaneously, we can efficiently distinguish the energy consumption characteristics of each device and achieve accurate energy consumption decomposition. This process addresses the poor load monitoring accuracy issues inherent in existing technologies, enabling more efficient and accurate appliance energy consumption monitoring and decomposition.

[0025] In some embodiments, step S101 establishes current templates and voltage templates for multiple electrical appliances within the monitoring target during normal operation, including: connecting each electrical appliance to a power source separately, and collecting current data and voltage data of the electrical appliance under different working states; the working state includes a transient state during intervention and a steady state during normal operation; normalizing the current data and voltage data; extracting steady-state characteristics and transient characteristics of the current data and voltage data, and analyzing the frequency domain characteristics of the current data and voltage data; combining the steady-state characteristics and transient characteristics of the current data and voltage data into a multi-dimensional feature vector to obtain the current template and voltage template of the electrical appliance during normal operation.

[0026] In this embodiment, steady-state features include the average value, standard deviation, and power of current and voltage data; transient features include the rise time, fall time, and peak value of current and voltage data; and frequency domain features include harmonic content and frequency distribution.

[0027] Specifically, the normalization processing is performed on the current data and the voltage data, including: filtering the current data and the voltage data, and normalizing the current data and the voltage data to a unified scale.

[0028] For example, each appliance is individually connected to a power source and data is collected. Current and voltage waveforms are collected under different operating conditions, including transient states during intervention and steady-state operation. Digital filters (such as low-pass filters) are applied to remove high-frequency noise, and signal processing techniques are used to remove random noise and nonlinear distortion from the data. The current and voltage data are normalized to a uniform scale for easy comparison and analysis. Steady-state features of the current and voltage data, including mean, standard deviation, and power (active, reactive, and apparent power), are extracted. Transient features during startup and shutdown are extracted, including the rise and fall times and peak values ​​of the current and voltage. The frequency domain features of the current and voltage are analyzed using a Fast Fourier Transform (FFT). The extracted steady-state and transient features of current and voltage are combined into a multidimensional feature vector, which represents the current and voltage templates for the appliance, with each dimension representing a specific feature. Furthermore, the current and voltage templates for each appliance are stored in a database to facilitate subsequent matching and identification. An index is created for each template to facilitate rapid retrieval and comparison.

[0029] By employing this technical solution, each appliance is individually connected to the power supply so that its current and voltage can be accurately measured. Data is collected from each appliance over a long period of time to capture its current and voltage waveforms under different operating conditions. High-sampling-rate sensors are used to ensure that rapid changes in current and voltage can be captured.

[0030] In some embodiments, in step S103, the electricity usage data is input into the pre-trained graph neural network for monitoring at set intervals to determine the electrical properties of each electrical appliance, including: detecting transient events in the changes of current data and voltage data through a sliding window; combining the transient event detection results and the time series as well as the current template and voltage template of each electrical appliance to determine the electrical properties of the electrical appliance; the electrical properties include the type of appliance, the start time point and the shutdown time point.

[0031] For example, electricity data packets are exchanged and parsed every 12 to 24 hours, using a sliding window approach to detect transient events in current and voltage data changes, which may indicate appliance interruption. By combining transient event detection with time series analysis, the start time of the appliance can be determined, thereby confirming the appliance type and its startup and shutdown times.

[0032] In some embodiments, step S104 assigns an initial feature vector to a node corresponding to each electrical appliance, including: for each electrical appliance, extracting features from the current template and voltage template of the electrical appliance to obtain template features; determining an initial feature vector based on the template features of the electrical appliance, and assigning the initial feature vector to the node corresponding to the electrical appliance.

[0033] For example, current data, voltage data, and their features are fed into a graph neural network for recognition. Graph convolutional layers (such as a graph convolutional network (GCN) and a graph attention network (GAT)) process node and edge information to capture interactions between appliances. The output layer of the graph neural network identifies the operating status of each appliance. This output layer can be a multi-classification layer, with each class corresponding to a specific operating status.

[0034] In this example, each appliance is assigned an initial feature vector, which represents the characteristics of the current and voltage templates. These features are then propagated through multiple graph convolutional layers, each of which updates the node's feature vector. Within each graph convolutional layer, node features are updated based on the characteristics of their neighbors. An output layer is applied to the output of the final graph convolutional layer to generate a prediction of the operating status of each appliance. In practical applications, the structure and parameters of the graph neural network can be adjusted to suit specific problems.

[0035] The graph convolution operation is expressed as:

[0036] in, represents the node features of the l+1th layer, represents the activation function, represents the set of neighbor nodes of node v, Representation node and The normalization coefficient between is the bias, represents the learnable weight matrix.

[0037] At the output layer of the graph neural network, a classifier is used to identify the working status of each appliance. The classifier formula is as follows:

[0038] in, Represents the predicted output vector of node v, each element corresponds to a possible working state, represents the weight matrix of the output layer, represents the bias vector of the output layer, represents the feature vector of the last graph convolution layer, The function is used to convert the output vector into a probability distribution.

[0039] In some embodiments, step S105 performs a time period health assessment on the appliance based on the working status of the appliance identified at each time interval to generate a health status result of the appliance, including: extracting features of the working status of the appliance to obtain working status features; combining the working status features of the appliance with predefined evaluation rules to score the health status of the appliance; and determining the health status result of the appliance based on the health status score of each appliance.

[0040] For example, for each electrical appliance, its working status characteristics are extracted. The working status characteristics include the number of switching times, operating time, and energy consumption. An evaluation rule is defined and a score is assigned based on the working status characteristics of the electrical appliance. The evaluation rule formula is as follows:

[0041] in, represents the health status score of appliance i, represents the score of appliance i based on its running time, represents the score of appliance i based on the number of times it is switched on and off, represents the score of appliance i based on energy consumption, 、 、 are weight coefficients, which are adjusted according to the type and importance of appliance i.

[0042] The calculation formula for the score of appliance i based on operating time is as follows:

[0043] in, Indicates the actual running time of appliance i, Indicates the expected normal operating time of appliance i.

[0044] The calculation formula for the score of appliance i based on the number of times it is switched on and off is as follows:

[0045] in, Indicates the actual number of times the appliance i is switched on and off, Indicates the maximum number of switching times allowed for appliance i.

[0046] The calculation formula for the energy consumption score of appliance i is as follows:

[0047] in, represents the actual energy consumption of appliance i, represents the expected energy consumption of appliance i.

[0048] Furthermore, by aggregating the health scores of each appliance and outputting daily appliance performance and health scores, we can quantitatively assess the health of household appliances, helping users understand their operating conditions and take appropriate maintenance measures. It's important to note that the weight coefficients and expected values ​​in the above formulas need to be adjusted and optimized based on the characteristics of the specific appliance.

[0049] Determining the expected uptime and energy consumption of different appliances involves the following steps: collecting historical operating data for the appliances, including operating time and energy consumption; understanding users' daily usage habits, including usage duration and frequency; obtaining appliance specifications, such as power ratings and typical operating cycles; cleaning the data to remove outliers and erroneous data; and standardizing the data to make it suitable for analysis and modeling. For operating time and energy consumption, the average of the historical data can be used as the expected value. When the data distribution is uneven or contains outliers, the median can be used to reduce the impact of outliers.

[0050] The present invention also provides a non-invasive load monitoring and analysis device. The non-invasive load monitoring and analysis device provided by the present invention is described below. The non-invasive load monitoring and analysis device described below and the non-invasive load monitoring and analysis method described above can be referenced to each other. Figure 2 This is a structural block diagram of the non-intrusive load monitoring and analysis device provided by the present invention, such as Figure 2 As shown, the device includes: A construction module 201 is used to establish a current template and a voltage template for each of the multiple electrical appliances in the monitoring target during normal operation; The acquisition module 202 is used to collect the power consumption data of the main switch circuit of the monitoring target; the power consumption data includes current data and voltage data; Monitoring module 203, for inputting electricity usage data into a pre-trained graph neural network for monitoring at set intervals to determine the electrical properties of each appliance; Update module 204, used to extract feature vectors of electrical appliances and perform feature propagation, capture the association between each electrical appliance, and determine the working status of each electrical appliance; The evaluation module 205 is configured to perform a time period health evaluation on the electrical appliance based on the working status of the electrical appliance identified at each time interval, and generate a health status result of the electrical appliance.

[0051] The present invention also provides a non-invasive load monitoring and analysis system for implementing a non-invasive load monitoring and analysis method, comprising an electricity consumption data detection module, a memory and a non-invasive load analysis module. The electricity consumption data detection module comprises a current sensor, a voltage sensor and a communication module. The current sensor and the voltage sensor collect current data and voltage data at the main switch of the monitoring target. The current sensor and the voltage sensor are respectively connected to the memory through the communication module for communication; the non-invasive load analysis module comprises a computer server and a display. The computer server is pre-installed with an analysis program designed using a non-invasive load monitoring and analysis method based on a fusion graph neural network. The current data and voltage data in the memory are input into the analysis program. The analysis program is run to output daily appliance working conditions and health status scoring results, and the results are displayed on the display.

[0052] Figure 3 An example of a physical structure diagram of an electronic device is shown below. Figure 3 As shown, the electronic device may include: a processor 301, a communications interface 302, a memory 303, and a communication bus 304, wherein the processor 301, the communications interface 302, and the memory 303 communicate with each other via the communication bus 304. The processor 301 may call the logic instructions in the memory 303 to execute the non-intrusive load monitoring and analysis method, which includes: Establishing current templates and voltage templates for multiple electrical appliances within the monitoring target during normal operation; Collect the power consumption data of the main switch circuit of the monitoring target; the power consumption data includes current data and voltage data; At set intervals, the electricity usage data is fed into a pre-trained graph neural network for monitoring and determining the electrical properties of each appliance. Extract the feature vectors of the appliances and perform feature propagation to capture the associations between each appliance and determine the working status of each appliance; Based on the working status of the appliance identified at each time interval, a health assessment of the appliance is performed over the time period to generate a health status result of the appliance.

[0053] Furthermore, the logic instructions in the aforementioned memory 303 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0054] In another aspect, the present invention further provides a computer program product, which includes a computer program. The computer program may be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can perform the non-intrusive load monitoring and analysis method provided by the above methods, which includes: Establishing current templates and voltage templates for multiple electrical appliances within the monitoring target during normal operation; Collect the power consumption data of the main switch circuit of the monitoring target; the power consumption data includes current data and voltage data; At set intervals, the electricity usage data is fed into a pre-trained graph neural network for monitoring and determining the electrical properties of each appliance. Extract the feature vectors of the appliances and perform feature propagation to capture the associations between each appliance and determine the working status of each appliance; Based on the working status of the appliance identified at each time interval, a health assessment of the appliance is performed over the time period to generate a health status result of the appliance.

[0055] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the non-intrusive load monitoring and analysis method provided by the above methods is implemented. The method includes: Establishing current templates and voltage templates for multiple electrical appliances within the monitoring target during normal operation; Collect the power consumption data of the main switch circuit of the monitoring target; the power consumption data includes current data and voltage data; At set intervals, the electricity usage data is fed into a pre-trained graph neural network for monitoring and determining the electrical properties of each appliance. Extract the feature vectors of the appliances and perform feature propagation to capture the associations between each appliance and determine the working status of each appliance; Based on the working status of the appliance identified at each time interval, a health assessment of the appliance is performed over the time period to generate a health status result of the appliance.

[0056] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0057] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0058] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A non-intrusive load monitoring and analysis method, characterized in that: include: Establishing current templates and voltage templates for multiple electrical appliances within the monitoring target during normal operation; Collecting power consumption data of the main switch circuit of the monitoring target; the power consumption data includes current data and voltage data; At set intervals, the electricity usage data is input into a pre-trained graph neural network for monitoring to determine the electrical properties of each of the electrical appliances; Extracting feature vectors of the electrical appliances and performing feature propagation to capture associations between the electrical appliances and determine the operating status of each electrical appliance, including: assigning an initial feature vector to a node corresponding to each electrical appliance; performing feature propagation through the graph neural network, updating features based on neighboring nodes of each node corresponding to the electrical appliance, and determining the operating status of each electrical appliance; Based on the working status of the electrical appliance identified at each time interval, a time period health assessment is performed on the electrical appliance to generate a health status result of the electrical appliance.

2. The non-intrusive load monitoring and analysis method according to claim 1, characterized in that: Establish current templates and voltage templates for normal operation of multiple electrical appliances within the monitoring target, including: Connecting each of the electrical appliances to a power source individually to collect current data and voltage data of the electrical appliances under different working states; the working states include a transient state during intervention and a steady state during normal operation; performing normalization processing on the current data and the voltage data; extracting steady-state characteristics and transient characteristics of the current data and the voltage data, and analyzing frequency domain characteristics of the current data and the voltage data; The steady-state characteristics and transient characteristics of the current data and the voltage data are combined into a multi-dimensional feature vector to obtain a current template and a voltage template when the electrical appliance is operating normally.

3. The non-intrusive load monitoring and analysis method according to claim 2, characterized in that: The steady-state characteristics include the average value, standard deviation and power of the current data and the voltage data; The transient characteristics include the rise time, fall time and peak value of the current data and the voltage data; The frequency domain characteristics include harmonic content and frequency distribution.

4. The non-intrusive load monitoring and analysis method according to claim 2, characterized in that: Normalizing the current data and the voltage data includes filtering the current data and the voltage data to normalize the current data and the voltage data to a unified scale.

5. The non-intrusive load monitoring and analysis method according to claim 1, characterized in that: At set intervals, the electricity usage data is input into a pre-trained graph neural network for monitoring to determine the electrical properties of each appliance, including: detecting transient events in changes of the current data and the voltage data through a sliding window; The electrical appliance attributes of the electrical appliance are determined by combining the transient event detection result and the time sequence as well as the current template and the voltage template of each electrical appliance; the electrical appliance attributes include the type of the electrical appliance, the start time point and the shutdown time point.

6. The non-intrusive load monitoring and analysis method according to claim 1, characterized in that: Assign an initial feature vector to each node corresponding to the electrical appliance, including: For each of the electrical appliances, feature extraction is performed on the current template and voltage template of the electrical appliance to obtain template features; An initial feature vector is determined based on the template feature of the electrical appliance, and the initial feature vector is allocated to a node corresponding to the electrical appliance.

7. The non-intrusive load monitoring and analysis method according to claim 1, characterized in that: Based on the working status of the electrical appliance identified at each time interval, a time period health assessment is performed on the electrical appliance to generate a health status result of the electrical appliance, including: Extracting features of the working state of the electrical appliance to obtain working state features; combining the working status characteristics of the appliance with predefined evaluation rules to score the health status of the appliance; A health status result of the electrical appliances is determined based on the health status score of each of the electrical appliances.

8. A non-intrusive load monitoring and analysis device, characterized in that: include: A construction module is used to establish a current template and a voltage template for each of the multiple electrical appliances in the monitoring target during normal operation; An acquisition module, configured to acquire power consumption data of the main switch circuit of the monitoring target; the power consumption data includes current data and voltage data; A monitoring module, configured to input the electricity usage data into a pre-trained graph neural network for monitoring at set intervals to determine the electrical properties of each of the electrical appliances; An updating module, configured to extract feature vectors of the electrical appliances and perform feature propagation, capture the associations between each of the electrical appliances, and determine the working status of each of the electrical appliances; The evaluation module is configured to perform a time period health evaluation on the electrical appliance based on the working status of the electrical appliance identified at each time interval, and generate a health status result of the electrical appliance.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the non-intrusive load monitoring and analysis method according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the non-intrusive load monitoring and analysis method according to any one of claims 1 to 7 is implemented.

Images