Non-intrusive power data acquisition method, platform, equipment and medium

Abstract

The invention discloses a non-intrusive power data acquisition method, platform, equipment and medium, and relates to the related field of power grid technology, and the method comprises the steps: carrying out the efficacy feature recognition of electric appliances, and carrying out the classification of the electric appliances according to the efficacy features of the electric appliances; acquiring electric signal data, and performing time sequence characteristic analysis on the electric signal data according to a preset stage segmentation relation; based on the efficacy classification label, performing type clustering on the electric appliance electric signal time sequence feature list, performing feature alignment on a clustering result, and identifying differentiated data and time sequence features; similarity evaluation is carried out, an electric signal amplifier is configured, and efficacy identification features are determined; searching efficacy recognition features, electric signal recognition features and amplified electric signal features, performing feature aggregation, and generating a data acquisition identity tag; and carrying out non-intrusive electric power data acquisition. The technical problem that existing non-intrusive electric power data acquisition is insufficient in accuracy and reliability is solved, and the technical effect of improving the accuracy and reliability of data acquisition is achieved.

Description

Technical Field

[0001] The present application relates to the field of power grid technology, and in particular to non-intrusive power data acquisition methods, platforms, equipment and media. Background Art

[0002] Power data collection is used to monitor all aspects of the power system in real time, accurately and efficiently, and provide data support for the stable operation and optimized management of the power system. In the field of power data collection, there are mainly two technical methods: invasive and non-invasive. Invasive power data collection requires the installation of sensors or measuring devices on electrical equipment or power lines to directly obtain power data. Although the data is highly accurate, the installation and maintenance costs are high, and it will have a certain impact on the normal operation of the power system. Non-invasive power data collection does not require physical changes to electrical equipment or power lines. By monitoring the total current, voltage and other parameters of the power line, the power consumption of each electrical equipment can be estimated. However, due to the large number of types and quantities of electrical equipment and their different power consumption characteristics, it becomes difficult to accurately identify the consumption of each electrical equipment in the total power data, and data confusion and misidentification are prone to occur, affecting the accuracy and reliability of data collection.

[0003] Among the current related technologies, non-invasive power data collection has technical problems such as insufficient accuracy and reliability. Summary of the invention

[0004] The present application provides a non-invasive power data collection method, platform, equipment and medium, identifies and classifies the functional characteristics of electrical appliances in the data collection area, obtains functional classification labels, obtains the electrical signal data of these electrical appliances, and performs timing feature analysis to construct a time series feature list of electrical signals of the electrical appliances. Based on the functional classification labels, the time series feature list is clustered and aligned by type, differentiated data and time series features are identified, and similarity is further evaluated based on the differentiated data and time series features. An electrical signal amplifier is configured to amplify the electrical signals that meet the confusion conditions. Using the regional electrical appliance label as an index, the functional identification feature, the electrical signal identification feature and the amplified electrical signal feature are aggregated to generate a data collection identity label, and a target identification is provided for power data collection. Technical means such as these achieve the technical effect of improving the accuracy and reliability of data collection.

[0005] The present application provides a non-intrusive power data collection method, including: The electrical appliances in the data collection area are identified for their functional characteristics, and the electrical appliances are classified according to their functional characteristics to obtain functional classification labels; the electrical signal data of the regional electrical appliances are obtained, and the timing characteristics of the electrical signal data are analyzed according to the preset stage segmentation relationship to construct a list of electrical signal timing characteristics of the electrical appliances; based on the functional classification labels, the electrical signal timing characteristics list of the electrical appliances is clustered by type, and the clustering results are feature aligned to identify differentiated data and timing characteristics; similarity evaluation is performed based on the differentiated data and timing characteristics, and an electrical signal amplifier is configured based on the similarity evaluation results to determine the functional identification characteristics, and the electrical signal amplifier is used to amplify the electrical signals whose similarity evaluation results meet the confusion conditions; the functional identification characteristics, electrical signal identification characteristics and amplified electrical signal characteristics are searched with the labels of the regional electrical appliances as indexes, and the functional identification characteristics, the electrical signal identification characteristics and the amplified electrical signal characteristics are feature aggregated to generate a data collection identity label; non-invasive power data collection of regional electrical appliances is performed based on the data collection identity label.

[0006] In a possible implementation, the electrical appliances in the data collection area are identified for their functional characteristics, classified according to their functional characteristics, and functional classification labels are obtained, and the following processing is performed: Collect the load characteristics of regional electrical appliances, configure the levels of the load characteristics, and obtain the load levels; based on the load levels, identify the starting current peak, steady-state power, power curve, and harmonic characteristics of the electrical appliance signals to obtain the electrical appliance efficacy characteristics, and establish a mapping association between the load level and the electrical appliance efficacy characteristics; configure data tags for the electrical appliance type, the load level, and the electrical appliance efficacy characteristics, integrate the data tags with a mapping association relationship, identify the efficacy characteristics of the regional electrical appliances, and obtain the efficacy classification tag.

[0007] In a possible implementation, the electrical signal data is subjected to a time series feature analysis according to a preset stage segmentation relationship, and the following processing is performed: Establish a stage template signal according to the preset stage, the preset stage includes a startup stage, a stable operation stage, and a shutdown stage; use the electrical signal data and the stage template signal to perform timing alignment, and calculate the distance measurement result between the timing signals; based on the distance measurement result, determine the key timing nodes of the stage segmentation interval to obtain the preset stage segmentation relationship; perform timing fluctuation feature analysis of the electrical signal data according to the preset stage segmentation relationship to obtain the electrical signal timing characteristics, including the fluctuation amplitude, fluctuation frequency, and periodic fluctuation of the current, voltage or power.

[0008] In a possible implementation, the electrical signal data is used to perform timing alignment with the phase template signal, a distance measurement result between the timing signals is calculated, and the following processing is performed: The electrical signal data is time-aligned with the stage template signal to obtain a timing data pair; the Euclidean distance of the timing data pair is calculated; based on the Euclidean distance, a minimum distance is found through a dynamic programming algorithm to obtain an accumulated path, and the Euclidean distances in the path are superimposed according to the accumulated path to obtain the distance measurement result.

[0009] In a possible implementation, based on the distance measurement result, the key timing nodes of the stage segmentation interval are determined, and the following processing is performed: According to the change of the accumulated distance value in the accumulation path, the first distance change attenuation node of the starting timing is identified, and the key timing node of the startup phase is obtained, and the key timing node of the startup phase is the inflection point where the distance change rate decreases; the key timing node of the startup phase is used as the starting observation point, and the change rate of the accumulated distance value begins to increase, and the inflection point where the distance change rate increases is determined to obtain the key timing node of the closing phase; the part between the key timing node of the startup phase and the key timing node of the closing phase is the timing interval of the stable operation phase, and the timing interval of the stable operation phase is the area where the distance changes slowly.

[0010] In a possible implementation, the electrical signal timing characteristics are obtained and the following processing is performed: Calculate the mean of the electrical signal sequence: , where N is the length of the electrical signal sequence, is the electrical signal data at time t; according to the time sequence relationship of the electrical signal sequence, the time lag is calculated, and the time lag is the alignment amount of the signal itself and the signal after r time units of the post-time sequence; the autocorrelation value of the time lag is calculated: ,in is the mean of the electrical signal sequence, is the autocorrelation value of the time lag r; according to the Identify periodic fluctuation characteristics.

[0011] In a possible implementation, the efficacy identification feature, the electrical signal identification feature, and the amplified electrical signal feature are feature aggregated to generate a data collection identity tag, and the following processing is performed: The efficacy identification feature, the electrical signal identification feature and the amplified electrical signal feature are respectively represented as feature vectors; multi-scale analysis is performed on each of the efficacy identification feature, the electrical signal identification feature and the amplified electrical signal feature to obtain multi-scale features; based on the multi-scale features, the attention weight of each feature is calculated through a self-attention mechanism; the attention weights are used to calculate the attention scores of the efficacy identification feature, the electrical signal identification feature and the amplified electrical signal feature, and then weighted summation is performed to obtain a comprehensive feature; the comprehensive feature is mapped to a data collection identity label for regional electrical appliances.

[0012] This application also provides a non-intrusive power data acquisition platform, including: An electrical appliance function classification module, the electrical appliance function classification module is used to identify the function characteristics of the electrical appliances in the data collection area, classify the electrical appliances according to the electrical appliance function characteristics, and obtain the function classification label; an electrical signal timing feature analysis module, the electrical signal timing feature analysis module is used to obtain the electrical signal data of the regional electrical appliances, perform timing feature analysis on the electrical signal data according to the preset stage segmentation relationship, and construct an electrical appliance electrical signal timing feature list; a differential data identification module, the differential data identification module is used to type cluster the electrical appliance electrical signal timing feature list based on the function classification label, and perform feature alignment on the clustering results to identify the differential data and timing features; an electrical signal amplification module, the electrical signal amplification module is used to obtain the electrical signal data of the regional electrical appliances, perform timing feature analysis on the electrical signal data according to the preset stage segmentation relationship, and construct an electrical appliance electrical signal timing feature list; a differential data identification module, the differential data identification module is used to type cluster the electrical appliance electrical signal timing feature list based on the function classification label, and perform feature alignment on the clustering results to identify the differential data and timing features; a The differentiated data and the time series features are evaluated for similarity, an electric signal amplifier is configured according to the similarity evaluation result, and the efficacy identification feature is determined, and the electric signal amplifier is used to amplify the electric signal whose similarity evaluation result meets the confusion condition; a data acquisition identity label generation module, the data acquisition identity label generation module is used to search for efficacy identification features, electric signal identification features and amplified electric signal features with the labels of regional electrical appliances as indexes, and feature aggregate the efficacy identification features, the electric signal identification features and the amplified electric signal features to generate a data acquisition identity label; a non-invasive power data acquisition module, the non-invasive power data acquisition module is used to perform non-invasive power data acquisition of regional electrical appliances based on the data acquisition identity label.

[0013] The present application also provides an electronic device, including: The memory is used to store executable instructions; the processor is used to implement a non-intrusive power data acquisition method when executing the executable instructions stored in the memory.

[0014] The present application also provides a computer-readable storage medium, comprising: A computer program is stored thereon, and when the program is executed by a processor, a non-intrusive power data collection method is implemented.

[0015] The non-intrusive power data collection method, platform, device and medium proposed in this application are first used to identify the functional characteristics of electrical appliances in the data collection area, classify the electrical appliances according to the functional characteristics of the electrical appliances, and obtain functional classification labels. Then, the electrical signal data of the regional electrical appliances are obtained, and the time series characteristics of the electrical signal data are analyzed according to the preset stage segmentation relationship to construct a list of electrical signal time series characteristics of the electrical appliances. Then, based on the functional classification labels, the electrical signal time series characteristics list of the electrical appliances is clustered by type, and the clustering results are feature aligned to identify differentiated data and time series characteristics. Then, similarity evaluation is performed based on the differentiated data and time series characteristics, and an electrical signal amplifier is configured based on the similarity evaluation results to determine the functional identification characteristics, wherein the electrical signal amplifier is used to amplify the electrical signal whose similarity evaluation results meet the confusion condition, and then the functional identification characteristics, electrical signal identification characteristics and amplified electrical signal characteristics are searched with the labels of the regional electrical appliances as indexes, and the functional identification characteristics, electrical signal identification characteristics and amplified electrical signal characteristics are feature aggregated to generate data collection identity labels. Finally, non-intrusive power data collection of regional electrical appliances is performed based on the data collection identity labels, thereby achieving the technical effect of improving the accuracy and reliability of data collection. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the accompanying drawings of the embodiment of the present invention are briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the platform according to the embodiment of the present application. It should be understood that the previous or following operations are not necessarily performed precisely in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0017] Figure 1 A schematic diagram of a flow chart of a non-intrusive power data collection method provided in an embodiment of the present application.

[0018] Figure 2 A schematic diagram of the structure of the non-intrusive power data acquisition platform provided in an embodiment of the present application.

[0019] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0020] Explanation of the accompanying drawings: electrical appliance function classification module 10, electrical signal timing feature analysis module 20, differential data recognition module 30, electrical signal amplification module 40, data acquisition identity tag generation module 50, non-invasive power data acquisition module 60, input device 301, processor 302, memory 303, output device 304. DETAILED DESCRIPTION

[0021] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0022] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0023] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions, for example, a process, method, platform, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0024] The present application provides a non-intrusive power data collection method, such as Figure 1 As shown, the method includes: Step S100, identify the function characteristics of the electrical appliances in the data collection area, classify the electrical appliances according to the function characteristics of the electrical appliances, and obtain function classification labels. Specifically, collect the load characteristics of all electrical appliances in the area, including the rated power and working current of the electrical appliances, and identify the starting current peak, steady-state power, power curve, harmonic characteristics, etc. of the electrical appliance signals according to the load characteristics of the electrical appliances. These characteristics together constitute the function characteristics of the electrical appliances. Configure a type label for each electrical appliance according to the function characteristics of the electrical appliances to obtain function classification labels.

[0025] In a possible implementation, the electrical appliances in the data collection area are identified by the efficacy characteristics, and the electrical appliances are classified according to the electrical appliance efficacy characteristics to obtain efficacy classification labels. Step S100 further includes step S110, collecting the load characteristics of the electrical appliances in the area, configuring the load characteristics by level, and obtaining the load level. Specifically, the load characteristic data of each electrical appliance in the area are collected through sensors or electric meters and other equipment, including parameters such as current, voltage, and power, which reflect the energy consumption of the electrical appliances under different working conditions. The collected load characteristic data are analyzed to identify different working modes or load states of the electrical appliances. For example, an air conditioner has multiple modes such as cooling, heating, and dehumidification, and the load characteristics in each mode are different. According to the analysis results, the load characteristics of the electrical appliances are divided into different levels based on factors such as power range and energy consumption mode. Step S120, based on the load level, the starting current peak, steady-state power, power curve, and harmonic characteristics of the electrical appliance signal are respectively identified to obtain the electrical appliance efficacy characteristics, and the load level is associated with the electrical appliance efficacy characteristics mapping. Specifically, for each load level, a series of feature recognition operations are performed, including the starting current peak (the maximum current value at the moment when the appliance starts, reflecting the starting characteristics of the appliance and the degree of impact on the power grid), steady-state power (the power value of the appliance in a stable working state, reflecting the energy consumption level and working efficiency of the appliance), power curve (the change of power of the appliance over time during operation, reflecting the working mode and energy consumption characteristics of the appliance) and harmonic characteristics (the harmonic components in the current or voltage waveform generated by the appliance, reflecting the electromagnetic compatibility of the appliance and the impact on the power grid). Through feature recognition, the efficacy characteristics of the appliance are extracted, and these characteristics comprehensively reflect the energy consumption characteristics and working state of the appliance. Each load level is mapped and associated with its corresponding efficacy characteristics so that the corresponding feature information can be quickly found when classifying and identifying the appliance. Step S130, configure the data tags of the appliance type, the load level, and the efficacy characteristics of the appliance, integrate the data tags with the mapping association relationship, identify the efficacy characteristics of the regional appliances, and obtain the efficacy classification tags. Specifically, according to the type, load level, and efficacy characteristics of the appliance, a corresponding data tag is configured for each appliance. These tags are used to uniquely identify each appliance and contain its important energy consumption characteristics and working status information. The data tags of all appliances are integrated to form a complete database or data table. This database or data table is used for query and retrieval. Based on the integrated data tags, the efficacy characteristics of each appliance are identified, that is, a unique identifier is configured for it based on the energy consumption characteristics and working status information of the appliance.This implementation method classifies electrical appliances in a refined manner by configuring levels and identifying the characteristics of their loads. Whether they are appliances with large power variations, such as air conditioners and refrigerators, or appliances with small power variations, such as light bulbs, they can all be accurately identified, thereby improving the accuracy and reliability of non-invasive power data collection.

[0026] Step S200, obtain the electrical signal data of the electrical appliances in the area, perform a timing feature analysis on the electrical signal data according to the preset stage segmentation relationship, and construct a timing feature list of the electrical signals of the electrical appliances. Specifically, the electrical signal data of the electrical appliances in the area are obtained through sensors and other equipment. According to the typical working stages of the electrical appliances (such as startup, stable operation, and shutdown), a stage template signal is established, the electrical signal data is time-aligned with the stage template signal, and the distance measurement result between the two is calculated. According to the distance measurement result, the key timing nodes of the stage segmentation interval are determined, such as the key timing nodes of the startup stage and the shutdown stage. According to the preset stage segmentation relationship, the timing fluctuation feature analysis is performed on the electrical signal data to obtain the timing features such as the fluctuation amplitude, fluctuation frequency, and periodic fluctuation of the current, voltage or power.

[0027] In a possible implementation, the electrical signal data is subjected to a time series feature analysis according to a preset stage segmentation relationship, and step S200 further includes step S210, establishing a stage template signal according to the preset stage, wherein the preset stage includes a startup stage, a stable operation stage, and a shutdown stage. Specifically, the template signal represents a signal sequence of typical electrical signal characteristics of a certain stage, and is used for comparison with the actual signal. Specifically, by collecting the electrical signal data when the device is started for multiple times, preprocessing these data, such as denoising, filtering, etc., calculating the average value or typical mode of these data (representative signal mode appearing in multiple observations), and generating a time series template signal for the startup stage. By collecting the electrical signal data of the device in a stable state, extracting the typical steady-state characteristic curve of this stage (the characteristic curve of the electrical signal when the device is in stable operation, which is manifested as a relatively stable waveform), a time series template signal for the stable operation stage is generated. The electrical signal data when the monitoring device is shut down is extracted, and the attenuation characteristics of the electrical signal during the shutdown process are extracted, including fitting the attenuation curve or calculating the attenuation rate, and generating a time series template signal for the shutdown stage. Step S220, using the electrical signal data and the stage template signal to perform time alignment, and calculating the distance measurement result between the time series signals. Specifically, using the dynamic time warping (DTW) algorithm, the actual electrical signal data is aligned with each stage template signal, and the time distance between them is calculated. Step S230, based on the distance measurement result, the key timing nodes of the stage segmentation interval are determined to obtain the preset stage segmentation relationship. Specifically, according to the distance measurement result calculated by the DTW algorithm, the time points at which the actual signal and the template signal have the highest matching degree are found, and these time points are used as the key timing nodes of the stage segmentation interval. Step S240, according to the preset stage segmentation relationship, the timing fluctuation feature analysis of the electrical signal data is performed to obtain the timing characteristics of the electrical signal, including the fluctuation amplitude, fluctuation frequency, and periodic fluctuation of the current, voltage or power. Specifically, in each stage, the characteristics such as the fluctuation amplitude, fluctuation frequency, and periodic fluctuation of the current, voltage or power are calculated using statistical methods or signal processing technology. This implementation method improves the accuracy of identifying each stage of the equipment operation by constructing a template signal and comparing it with the actual signal, thereby improving the accuracy of the timing fluctuation feature extraction. The segmentation results are dynamically adjusted through the DTW algorithm to adapt to signal fluctuations and phase changes, thereby enhancing the robustness of the system.

[0028] In a possible implementation, the electrical signal data is used to perform time alignment with the stage template signal, and the distance measurement result between the time series signals is calculated. Step S220 further includes step S221, performing time alignment with the stage template signal to obtain a time series data pair. Specifically, the electrical signal data sequence and the stage template signal sequence are read, and the electrical signal data sequence and the stage template signal sequence are aligned on the time axis using the preliminary step of the dynamic time warping (DTW) algorithm. During the alignment process, a series of time series data pairs are generated, each data pair containing the signal values ​​of the electrical signal data sequence and the stage template signal sequence at a certain time point. Step S222, calculating the Euclidean distance of the time series data pair. Specifically, for each time series data pair, its Euclidean distance is calculated, that is, the square root of the sum of the squares of the differences between the signal values ​​of the electrical signal data sequence and the stage template signal sequence at a certain time point. The Euclidean distances of all time series data pairs are stored in a matrix, and the rows and columns of the matrix correspond to the time points of the electrical signal data sequence and the stage template signal sequence, respectively. Step S223, according to the Euclidean distance, find the minimum distance through the dynamic programming algorithm, obtain the cumulative path, and superimpose the Euclidean distance in the path according to the cumulative path to obtain the distance measurement result. Specifically, using the dynamic programming algorithm, starting from the upper left corner of the matrix, gradually moving to the lower right corner, find the minimum path sum (i.e., the minimum distance cumulative sum) from the starting point to the end point. In each step, select the minimum distance in the position to the right, below, or in the lower right corner of the current position, and add it to the distance of the current position. Repeat the above steps until the lower right corner of the matrix is ​​reached, and the cumulative sum obtained at this time is the distance measurement result. According to the cumulative path, backtrack to obtain the best matching path between the electrical signal data sequence and the stage template signal sequence. This implementation method ensures that the electrical signal data sequence and the stage template signal sequence are compared on the same time scale through timing alignment, and quantifies the difference between the electrical signal data sequence and the stage template signal sequence by calculating the Euclidean distance of the timing data pair. Through the dynamic programming algorithm, the complexity of directly calculating all possible paths is avoided, the calculation efficiency of the minimum distance accumulation sum between the electrical signal data sequence and the stage template signal sequence is improved, and an accurate measurement of the degree of matching between the electrical signal data and the stage template signal is achieved.

[0029] In a possible implementation, based on the distance measurement result, the key timing nodes of the stage segmentation interval are determined, and step S230 further includes step S231, according to the cumulative distance value change in the cumulative path, the first distance change attenuation node of the start timing is identified, and the key timing node of the startup phase is obtained. The key timing node of the startup phase is the inflection point where the distance change rate decreases. Specifically, starting from the starting point of the cumulative path, the change of the cumulative distance value is gradually analyzed along the path, and the difference between adjacent cumulative distance values ​​is calculated to obtain the distance change amount. The position where the distance change amount first decreases significantly is identified, that is, the inflection point where the distance change rate decreases. This point marks the critical moment when the electrical signal changes from an initial unstable state (such as the initial stage of device startup) to a relatively stable state. This inflection point is used as the key timing node of the startup phase, which represents the end of the device startup phase.

[0030] Step S232, taking the key timing node of the startup phase as the starting observation point, identifying the point where the rate of change of the cumulative distance value begins to increase, determining the inflection point where the rate of change of the distance increases, and obtaining the key timing node of the shutdown phase. Specifically, taking the key timing node of the startup phase as the starting point, continue to analyze the change of the cumulative distance value along the accumulation path. Calculate the difference between adjacent cumulative distance values ​​to obtain a new distance change, identify the position where the distance change begins to increase significantly, that is, the inflection point where the rate of change of the distance increases, this point marks the key moment when the electrical signal changes from a stable operating state to an unstable state (such as device shutdown). Take this inflection point as the key timing node of the shutdown phase, which represents the end of the stable operating phase of the device and the beginning of the shutdown phase. Step S233, the part between the key timing node of the startup phase and the key timing node of the shutdown phase is the timing interval of the stable operating phase, and the timing interval of the stable operating phase is the area where the distance changes slowly. Specifically, according to the key timing nodes of the startup phase and the shutdown phase, determine the time range between them. The electrical signal data within this time range corresponds to the stable operating phase of the device, that is, the time period when the device operates under normal working conditions. During this time period, the cumulative distance value changes relatively smoothly, without obvious fluctuations or mutations. This implementation method identifies the key timing nodes of the startup phase and the shutdown phase, determines that the period between the key timing nodes of the startup phase and the shutdown phase is the stable operation phase, accurately divides the different phases of the equipment, provides a clear time range for the operation of the equipment under normal working conditions, and improves the accuracy of determining the preset phase segmentation relationship.

[0031] In a possible implementation, the electrical signal timing feature is obtained, and step S240 further includes step S241, calculating the mean of the electrical signal sequence: , where N is the length of the electrical signal sequence, The electrical signal data at time t. Specifically, an electrical signal sequence is obtained, which contains continuous electrical signal data from the start time to the current time. The mean of the electrical signal sequence is calculated, that is, the sum of all electrical signal data divided by the length N of the electrical signal sequence. The mean reflects the overall average level of the electrical signal sequence. Step S242, according to the timing relationship of the electrical signal sequence, calculate the time lag, and the time lag is the alignment of the signal itself with the signal after r time units of the post-time sequence. Specifically, the time lag r represents the number of time units in which the signal itself is aligned with the post-time sequence. For each time point t in the electrical signal sequence, find the signal data at time point t+r. Calculate the alignment of the signal itself with the signal data after r time units of the post-time sequence, that is, the time lag. This alignment can be the difference, ratio or other related measurement of the signal, which is used to analyze the trend of the signal over time. Step S243, calculate the autocorrelation value of the time lag: ,in is the mean of the electrical signal sequence, is the autocorrelation value of the time lag r. Specifically, the time lag calculated in step S242 is centered (i.e., the mean value is subtracted) using the mean value of the electrical signal sequence calculated in step S241. For each time lag r, its autocorrelation value is calculated. The autocorrelation value is a measure of the similarity between the time lag and itself at different time points, and is used to analyze the periodic characteristics of the signal. Step S244, according to the Identify periodic fluctuation characteristics. Specifically, analyze the autocorrelation values ​​calculated in step S243, especially the changes in the autocorrelation values ​​when r takes different values. Find the r value at which the autocorrelation value reaches the maximum or close to the maximum, and this r value corresponds to the period length of the signal. According to the periodic changes in the autocorrelation values, determine the periodic fluctuation characteristics of the signal, including the period length, fluctuation amplitude, etc. This implementation method reveals the hidden periodic information in the signal by calculating the mean, time lag and autocorrelation value of the electrical signal sequence, and accurately identifies the periodic fluctuation characteristics of the signal according to the changes in the autocorrelation values.

[0032] Step S300, based on the efficacy classification label, the electrical appliance electrical signal timing feature list is clustered by type, and the clustering results are feature aligned to identify differentiated data and timing features. Specifically, according to the efficacy classification label, the electrical appliance electrical signal timing feature list is clustered by type, and appliances with similar features are grouped into one category. The clustering results are feature aligned to ensure that the features of appliances in the same category are consistent and comparable. In the clustering process, data with large differences in features from other appliances in the same category are identified, and these difference data represent different appliances. Further timing feature recognition is performed on the differentiated data to extract its unique timing features.

[0033] Step S400, perform similarity evaluation based on the differentiated data and time series features, configure an electrical signal amplifier based on the similarity evaluation results, determine the efficacy identification features, and the electrical signal amplifier is used to amplify the electrical signal whose similarity evaluation results meet the confusion condition. Specifically, perform similarity evaluation on the differentiated data and its time series features to evaluate the similarity between the data. Based on the similarity evaluation results, configure an electrical signal amplifier (a device for amplifying electrical signals) to amplify the electrical signal whose similarity evaluation results meet the confusion condition (high electrical appliance feature similarity) to improve signal strength and recognition accuracy. Combine the amplified electrical signal and the original data to determine the efficacy identification features of the electrical appliance.

[0034] Step S500, using the label of the regional electrical appliance as an index to search for the function identification feature, the electric signal identification feature, and the amplified electric signal feature, and performing feature aggregation on the function identification feature, the electric signal identification feature, and the amplified electric signal feature to generate a data collection identity label. Specifically, using the label of the regional electrical appliance as an index, search for the corresponding function identification feature, the electric signal identification feature, and the amplified electric signal feature. Aggregate the searched features to form a comprehensive feature set. Map the comprehensive feature set to a data collection identity label of the regional electrical appliance (a unique label used to identify and recognize the regional electrical appliance) for data collection and identification.

[0035] In a possible implementation, the efficacy identification feature, the electric signal identification feature and the amplified electric signal feature are feature aggregated to generate a data collection identity tag, and step S500 further includes step S510, wherein the efficacy identification feature, the electric signal identification feature and the amplified electric signal feature are respectively represented as feature vectors. Specifically, the efficacy identification feature, the electric signal identification feature and the amplified electric signal feature are converted into feature vectors, and the feature vectors are used to describe multiple features of an object. These feature vectors can be numerical or obtained through some transformation (such as unique hot encoding, embedding, etc.). Step S520, multi-scale analysis is performed on each feature of the efficacy identification feature, the electric signal identification feature and the amplified electric signal feature to obtain multi-scale features. Specifically, multi-scale analysis is performed on each feature vector, that is, a filter (such as wavelet transform) is used to extract features at different time scales or spatial scales. Capture features and information at different levels. Step S530, based on the multi-scale features, the attention weight of each feature is calculated through a self-attention mechanism. Specifically, based on multi-scale features, the self-attention mechanism assigns attention weights to each feature by calculating the correlation between features, and the attention weights represent the relative importance between features. Step S540, using the attention weights to calculate the attention scores of the efficacy identification feature, the electrical signal identification feature, and the amplified electrical signal feature, and then performing weighted summation to obtain a comprehensive feature. Specifically, the dot product between the features is calculated, and the weights are normalized using the softmax function to obtain the attention score of each feature. The attention scores of the efficacy identification feature, the electrical signal identification feature, and the amplified electrical signal feature are weighted summed to obtain a comprehensive feature. Step S550, mapping the comprehensive feature to a data collection identity tag for regional electrical appliances. Specifically, a classifier (such as a support vector machine) is used to map the comprehensive feature to a data collection identity tag for regional electrical appliances, which is used to uniquely identify regional electrical appliances. This implementation adopts the self-attention mechanism to integrate features from different sources (functional recognition features, electrical signal recognition features, and amplified electrical signal features). The self-attention mechanism can dynamically assign weights according to the correlation between features, thereby integrating information more effectively, avoiding the problem of unreasonable weight allocation that may occur in traditional weighted average or simple feature splicing methods, and improving the accuracy and robustness of identity tag generation. Step S600, non-invasive power data collection of regional electrical appliances is performed based on the data collection identity tag. Specifically, based on the generated data collection identity tag, non-invasive power data collection is performed on regional electrical appliances, including collecting real-time energy consumption data, working status and other information of the electrical appliances, and these data are used for energy consumption analysis, energy management and other applications. The embodiment of the present application performs functional feature recognition and classification on the electrical appliances in the data collection area, obtains functional classification tags, obtains the electrical signal data of these electrical appliances, and performs timing feature analysis, constructs a time series feature list of electrical signals of electrical appliances, performs type clustering and feature alignment on the time series feature list based on the functional classification tags, identifies differentiated data and time series features, further performs similarity evaluation based on differentiated data and time series features, configures an electrical signal amplifier to amplify the electrical signal that meets the confusion condition, uses the regional electrical appliance tag as an index, aggregates functional identification features, electrical signal identification features and amplified electrical signal features, generates data collection identity tags, and provides target identification and other technical means for power data collection, thereby achieving the technical effect of improving the accuracy and reliability of data collection.

[0036] In the above, refer to Figure 1 The non-intrusive power data acquisition method according to an embodiment of the present invention is described in detail. Figure 2 A non-intrusive power data acquisition platform according to an embodiment of the present invention is described.

[0037] The non-intrusive power data acquisition platform according to the embodiment of the present invention is used to solve the technical problem of insufficient accuracy and reliability of existing non-intrusive power data acquisition, and achieve the technical effect of improving the accuracy and reliability of data acquisition. The non-intrusive power data acquisition platform includes: an electrical function classification module 10, an electrical signal timing feature analysis module 20, a differential data recognition module 30, an electrical signal amplification module 40, a data acquisition identity tag generation module 50, and a non-intrusive power data acquisition module 60.

[0038] The electrical appliance function classification module 10 is used to identify the function characteristics of the electrical appliances in the data collection area, classify the electrical appliances according to the electrical appliance function characteristics, and obtain the function classification label; the electrical signal timing feature analysis module 20 is used to obtain the electrical signal data of the regional electrical appliances, perform timing feature analysis on the electrical signal data according to the preset stage segmentation relationship, and construct an electrical appliance electrical signal timing feature list; the differential data identification module 30 is used to perform type clustering on the electrical appliance electrical signal timing feature list based on the function classification label, and perform feature alignment on the clustering results to identify the differential data and timing features; the electrical signal amplification module 40 is used to obtain the electrical signal data of the regional electrical appliances, perform timing feature analysis on the electrical signal data according to the preset stage segmentation relationship, and construct an electrical appliance electrical signal timing feature list; the differential data identification module 30 is used to perform type clustering on the electrical appliance electrical signal timing feature list based on the function classification label, and perform feature alignment on the clustering results to identify the differential data and timing features; the electrical signal amplification module 40 is used to perform The timing characteristics are evaluated for similarity, and an electric signal amplifier is configured according to the similarity evaluation result to determine the efficacy identification feature, and the electric signal amplifier is used to amplify the electric signal whose similarity evaluation result meets the confusion condition; the data acquisition identity label generation module 50 is used to search for efficacy identification features, electric signal identification features and amplified electric signal features with the labels of regional electrical appliances as indexes, and feature aggregate the efficacy identification features, the electric signal identification features and the amplified electric signal features to generate a data acquisition identity label; the non-invasive power data acquisition module 60 is used to perform non-invasive power data acquisition of regional electrical appliances based on the data acquisition identity label.

[0039] The specific configuration of the electrical appliance efficacy classification module 10 will be described in detail below. As described above, the electrical appliances in the data collection area are identified for efficacy characteristics, and the electrical appliances are classified according to the electrical appliance efficacy characteristics to obtain efficacy classification labels. The electrical appliance efficacy classification module 10 may further include: a load level configuration unit for collecting the load characteristics of the electrical appliances in the area, performing level configuration on the load characteristics, and obtaining the load level; an electrical appliance efficacy characteristic identification unit for respectively identifying the starting current peak value, steady-state power, power curve, and harmonic characteristics of the electrical appliance signal based on the load level to obtain the electrical appliance efficacy characteristics, and establishing a mapping association between the load level and the electrical appliance efficacy characteristics; a data label integration unit for configuring the data labels of the electrical appliance type, the load level, and the electrical appliance efficacy characteristics, integrating the data labels with the mapping association relationship, identifying the efficacy characteristics of the electrical appliances in the area, and obtaining the efficacy classification label.

[0040] The specific configuration of the electrical signal timing feature analysis module 20 will be described in detail below. As described above, the electrical signal data is subjected to timing feature analysis according to a preset stage segmentation relationship, and the electrical signal timing feature analysis module 20 may further include: a stage template signal establishment unit for establishing a stage template signal according to a preset stage, wherein the preset stage includes a startup stage, a stable operation stage, and a shutdown stage; a timing alignment unit for performing timing alignment using the electrical signal data and the stage template signal, and calculating a distance measurement result between timing signals; a key timing node determination unit for determining the key timing nodes of the stage segmentation interval based on the distance measurement result, and obtaining the preset stage segmentation relationship; a timing fluctuation feature analysis unit for performing timing fluctuation feature analysis of the electrical signal data according to the preset stage segmentation relationship, and obtaining electrical signal timing features, including fluctuation amplitude, fluctuation frequency, and periodic fluctuation of current, voltage, or power.

[0041] Among them, the electrical signal data is used to perform timing alignment with the stage template signal to calculate the distance measurement result between the timing signals. The timing alignment unit may further include: a timing data pair acquisition subunit for performing timing alignment with the stage template signal to obtain a timing data pair; a Euclidean distance calculation subunit for calculating the Euclidean distance of the timing data pair; a Euclidean distance superposition subunit for finding the minimum distance according to the Euclidean distance through a dynamic programming algorithm to obtain an accumulated path, and superimposing the Euclidean distance in the path according to the accumulated path to obtain the distance measurement result.

[0042] Among them, based on the distance measurement result, the key timing nodes of the stage division interval are determined, and the key timing node determination unit may further include: a startup phase key timing node acquisition subunit is used to identify the first distance change attenuation node of the starting timing according to the cumulative distance value change in the accumulated path, and obtain the key timing node of the startup phase, and the key timing node of the startup phase is the inflection point where the distance change rate decreases; the shutdown phase key timing node acquisition subunit is used to take the key timing node of the startup phase as the starting observation point, identify the increase in the rate of change of the cumulative distance value, determine the inflection point where the distance change rate increases, and obtain the key timing node of the shutdown phase; the stable operation phase timing interval determination subunit is used to define the part between the key timing node of the startup phase and the key timing node of the shutdown phase as the timing interval of the stable operation phase, and the timing interval of the stable operation phase is an area where the distance changes slowly.

[0043] Wherein, the electrical signal time series feature is obtained, and the time series fluctuation feature analysis unit may further include: an electrical signal sequence mean value calculation subunit for calculating the mean value of the electrical signal sequence: , where N is the length of the electrical signal sequence, is the electrical signal data at time t; the time lag calculation subunit is used to calculate the time lag according to the timing relationship of the electrical signal sequence, and the time lag is the alignment amount of the signal itself and the signal after r time units of the post-time sequence; the autocorrelation value calculation subunit is used to calculate the autocorrelation value of the time lag: ,in is the mean of the electrical signal sequence, is the autocorrelation value of the time lag r; the periodic fluctuation feature identification subunit is used to identify the periodic fluctuation feature according to the Identify periodic fluctuation characteristics.

[0044] The specific configuration of the data acquisition identity label generation module 50 will be described in detail below. As described above, the efficacy identification feature, the electric signal identification feature and the amplified electric signal feature are feature aggregated to generate a data acquisition identity label. The data acquisition identity label generation module 50 may further include: a feature vector generation unit for representing the efficacy identification feature, the electric signal identification feature and the amplified electric signal feature as feature vectors, respectively; a multi-scale analysis unit for performing multi-scale analysis on each feature of the efficacy identification feature, the electric signal identification feature and the amplified electric signal feature to obtain multi-scale features; an attention weight calculation unit for calculating the attention weight of each feature through a self-attention mechanism based on the multi-scale features; a comprehensive feature calculation unit for calculating the attention scores of the efficacy identification feature, the electric signal identification feature and the amplified electric signal feature using the attention weight, and then performing weighted summation to obtain a comprehensive feature; and a data acquisition identity label mapping unit for mapping the comprehensive feature into a data acquisition identity label for a regional electrical appliance.

[0045] The non-intrusive power data collection platform provided in the embodiment of the present invention can execute the non-intrusive power data collection method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0046] Although the present application makes various references to certain modules in the platform according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be implemented; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0047] Based on the foregoing embodiments, the embodiments of the present application further provide an electronic device and a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor of the electronic device, it can implement the method described in any of the foregoing embodiments.

[0048] Figure 3 1 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention, showing a block diagram of an exemplary electronic device suitable for implementing an implementation of the present invention. Figure 3 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention. The electronic device is in the form of a general computing device, and its components may include but are not limited to an input device 301, a processor 302, a memory 303, and an output device 304. Among them, the processor 302 may be one or more; the memory 303 may include a computer-readable medium and at least one program product, and the program product has a set (at least one) of program modules, which are configured to perform the functions of each embodiment of the present application.

[0049] The memory 303 shown in the embodiment of the present invention may adopt any combination of one or more computer-readable media; the computer-readable storage medium may be, but is not limited to, an infrared, semiconductor system, device or device, or any combination of the above, for storing software programs, computer executable programs and modules, such as the program instructions / modules corresponding to the non-intrusive power data acquisition method in the embodiment of the present invention. The processor 302 executes various functional applications and data processing of the computer device by running the software programs, instructions and modules stored in the memory 303, that is, realizing the above-mentioned non-intrusive power data acquisition method.

[0050] The above specific implementation manner does not constitute a limitation to the protection scope of the present application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application. In some cases, the actions or steps recorded in the present application can be performed in an order different from that in the embodiment and can still achieve the desired results. In addition, the process depicted in the accompanying drawings does not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A non-intrusive power data collection method, characterized in that: The non-intrusive power data collection method comprises: Identify the functional characteristics of electrical appliances in the data collection area, classify the electrical appliances according to the functional characteristics, and obtain functional classification labels; Acquire electrical signal data of regional electrical appliances, perform time series feature analysis on the electrical signal data according to a preset phase segmentation relationship, and construct a time series feature list of electrical signal of electrical appliances; Based on the efficacy classification labels, the electrical signal timing feature list of the electrical appliance is clustered by type, and the clustering results are feature aligned to identify differentiated data and timing features; Perform similarity evaluation based on the differentiated data and the time series features, configure an electrical signal amplifier based on the similarity evaluation results, and determine the efficacy identification features, wherein the electrical signal amplifier is used to amplify the electrical signal whose similarity evaluation results meet the confusion condition; Using the labels of regional electrical appliances as indexes, searching for function identification features, electrical signal identification features, and amplified electrical signal features, aggregating the function identification features, the electrical signal identification features, and the amplified electrical signal features, and generating a data collection identity label; Non-intrusive power data collection of regional electrical appliances is performed based on the data collection identity tag.

2. The non-intrusive power data collection method according to claim 1, characterized in that: The step of identifying the function characteristics of the electrical appliances in the data collection area, classifying the electrical appliances according to the function characteristics, and obtaining function classification labels includes: Collecting load characteristics of regional electrical appliances, performing level configuration on the load characteristics, and obtaining load levels; Based on the load level, the starting current peak value, steady-state power, power curve, and harmonic characteristics of the electrical appliance signal are respectively identified to obtain the electrical appliance efficacy characteristics, and a mapping association between the load level and the electrical appliance efficacy characteristics is established; The data tags of the electrical appliance type, the load level, and the electrical appliance function characteristics are configured, the data tags are integrated by mapping the association relationship, the function characteristics of the electrical appliances in the area are marked, and the function classification tags are obtained.

3. The non-intrusive power data collection method according to claim 1, characterized in that: Performing time series feature analysis on the electrical signal data according to a preset stage segmentation relationship includes: Establishing a phase template signal according to a preset phase, wherein the preset phase includes a startup phase, a stable operation phase, and a shutdown phase; Performing time sequence alignment with the phase template signal using the electrical signal data, and calculating a distance measurement result between the time sequence signals; Based on the distance measurement result, determine the key timing nodes of the stage segmentation interval to obtain the preset stage segmentation relationship; The timing fluctuation characteristics of the electrical signal data are analyzed according to the preset stage segmentation relationship to obtain the timing characteristics of the electrical signal, including the fluctuation amplitude, fluctuation frequency, and periodic fluctuation of the current, voltage or power.

4. The non-intrusive power data collection method according to claim 3, characterized in that: Performing time sequence alignment using the electrical signal data and the phase template signal, and calculating a distance measurement result between the time sequence signals, including: Performing time sequence alignment on the electrical signal data and the phase template signal to obtain a time sequence data pair; Calculating the Euclidean distance of the time series data pair; According to the Euclidean distance, a minimum distance is found through a dynamic programming algorithm to obtain an accumulated path, and the Euclidean distances in the path are superimposed according to the accumulated path to obtain the distance measurement result.

5. The non-intrusive power data collection method according to claim 4, characterized in that: Based on the distance measurement result, determining the key timing nodes of the stage segmentation interval includes: According to the cumulative distance value change in the accumulation path, the first distance change attenuation node of the start timing is identified to obtain the key timing node of the startup phase, wherein the key timing node of the startup phase is the inflection point where the distance change rate decreases; Taking the key timing node of the startup phase as the starting observation point, identifying that the rate of change of the cumulative distance value begins to increase, determining the inflection point where the rate of change of the distance increases, and obtaining the key timing node of the shutdown phase; The portion between the key timing node of the startup phase and the key timing node of the shutdown phase is the timing interval of the stable operation phase, and the timing interval of the stable operation phase is a region where the distance changes smoothly.

6. The non-intrusive power data collection method according to claim 3, characterized in that: Obtain the timing characteristics of electrical signals, including: Calculate the mean of the electrical signal sequence: , where N is the length of the electrical signal sequence, is the electrical signal data at time t; According to the timing relationship of the electrical signal sequence, a time lag is calculated, wherein the time lag is the alignment amount between the signal itself and the signal after r time units of the post-time sequence; Compute the autocorrelation values ​​for time lags: ,in is the mean of the electrical signal sequence, is the autocorrelation value of the time lag r; According to the Identify periodic fluctuation characteristics.

7. The non-intrusive power data collection method according to claim 1, characterized in that: The step of aggregating the efficacy identification feature, the electrical signal identification feature, and the amplified electrical signal feature to generate a data collection identity tag includes: Representing the efficacy identification feature, the electrical signal identification feature, and the amplified electrical signal feature as feature vectors respectively; Performing multi-scale analysis on each of the efficacy identification feature, the electrical signal identification feature, and the amplified electrical signal feature to obtain a multi-scale feature; Based on the multi-scale features, calculating the attention weight of each feature through a self-attention mechanism; Calculating the attention scores of the efficacy identification feature, the electrical signal identification feature and the amplified electrical signal feature using the attention weight, and then performing weighted summation to obtain a comprehensive feature; The comprehensive features are mapped into data collection identity tags of regional electrical appliances.

8. Non-intrusive power data acquisition platform, characterized by: The platform is used to implement the non-intrusive power data collection method according to any one of claims 1 to 7, and the platform includes: An electrical appliance function classification module, which is used to identify the function characteristics of electrical appliances in the data collection area, classify the electrical appliances according to the function characteristics, and obtain function classification labels; An electrical signal timing feature analysis module, which is used to obtain electrical signal data of regional electrical appliances, perform timing feature analysis on the electrical signal data according to a preset stage segmentation relationship, and construct a timing feature list of electrical appliance electrical signals; A differentiated data identification module, the differentiated data identification module is used to perform type clustering on the electrical signal timing feature list of the electrical appliance based on the efficacy classification label, and perform feature alignment on the clustering result to identify differentiated data and timing features; An electric signal amplification module, the electric signal amplification module is used to perform similarity evaluation according to the differentiated data and the time series characteristics, configure an electric signal amplifier according to the similarity evaluation result, determine the efficacy identification feature, and the electric signal amplifier is used to amplify the electric signal whose similarity evaluation result meets the confusion condition; A data collection identity tag generation module, the data collection identity tag generation module is used to search for function identification features, electric signal identification features and amplified electric signal features using the labels of regional electrical appliances as indexes, perform feature aggregation on the function identification features, the electric signal identification features and the amplified electric signal features, and generate a data collection identity tag; A non-intrusive power data acquisition module is used to perform non-intrusive power data acquisition of regional electrical appliances based on the data acquisition identity tag.

9. An electronic device, characterized in that: The electronic device comprises: A memory for storing executable instructions; The processor is used to implement the non-intrusive power data collection method according to any one of claims 1 to 7 when executing the executable instructions stored in the memory.

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

Images