Electric energy meter adaptive adjustment method under time period power consumption data analysis

Abstract

The invention discloses an electric energy meter adaptive adjustment method under time period power consumption data analysis, and belongs to the technical field of power management, and the method comprises the steps: collecting and obtaining a time period power consumption data flow of a target user; defining electricity consumption data dimension information and electricity consumption index information, and classifying and identifying the time period electricity consumption standard data flow according to the electricity consumption data dimension information and the electricity consumption index information; constructing a power utilization data integrated analysis network; time period division is carried out based on the time period electricity utilization characteristic parameters, and an electric energy meter self-adaptive adjustment strategy is constructed; and obtaining time period real-time data of the electric energy meter, and performing metering adaptive adjustment on the time period real-time data of the electric energy meter based on the adaptive adjustment strategy of the electric energy meter. The technical problems of insufficient metering precision and low power consumption management efficiency caused by the fact that an electric energy meter cannot be adaptively adjusted according to diversified power consumption characteristics in the prior art are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of power management, and in particular to a method for self-adapting an electric energy meter under time period power consumption data analysis. Background Art

[0002] With the acceleration of global energy transformation, the power industry is gradually moving towards digitalization, intelligence and refined management. As an important infrastructure for modern energy management, smart grid can achieve efficient distribution and intelligent regulation of electric energy through the deep integration of information technology and power system. In the smart grid system, the electric energy meter is not only a basic electricity metering device, but also a core node of intelligent management. Therefore, the study of the adaptive adjustment method of electric energy meter based on data analysis is of great significance to improving the operating efficiency of the power system and the level of user electricity management.

[0003] At present, most smart energy meters are still based on fixed time periods and standard metering rules, lacking in-depth analysis of users' personalized electricity consumption characteristics and diversified electricity consumption behaviors. Especially in scenarios where there are large differences in peak and valley electricity consumption and where electricity consumption behaviors change frequently and dynamically, the lag in the adjustment strategy of traditional energy meters can easily lead to metering errors or waste of resources. In addition, the existing technology still has deficiencies in data processing and cannot effectively solve the problems of abnormal data processing, real-time adjustment response, and multi-dimensional analysis. In order to adapt to the complex electricity consumption environment, an adaptive adjustment method for energy meters under time period electricity consumption data analysis has emerged to solve the technical problem that the energy meters in the existing technology cannot be adaptively adjusted according to the diversified electricity consumption characteristics, resulting in insufficient metering accuracy and low electricity management efficiency. Summary of the invention

[0004] The present application provides an adaptive adjustment method for electric energy meters based on time period electricity consumption data analysis, aiming to solve the technical problem in the prior art that electric energy meters cannot be adaptively adjusted according to diversified electricity consumption characteristics, resulting in insufficient metering accuracy and low electricity consumption management efficiency.

[0005] In view of the above problems, the present application provides an adaptive adjustment method for an electric energy meter based on time period electricity consumption data analysis.

[0006] The present application provides a method for adaptively adjusting an electric energy meter under period electricity consumption data analysis, the method comprising collecting and acquiring a period electricity consumption data stream of a target user, performing standardized preprocessing on the period electricity consumption data stream through a data preprocessing program, and obtaining a period electricity consumption standard data stream; defining electricity consumption data dimension information and electricity consumption index information, classifying and identifying the period electricity consumption standard data stream according to the electricity consumption data dimension information and electricity consumption index information, and obtaining a multi-dimensional period electricity consumption index data set; constructing an electricity consumption data integration analysis network, performing electricity consumption characteristic analysis on the multi-dimensional period electricity consumption index data set based on the electricity consumption data integration analysis network, and determining period electricity consumption characteristic parameters; performing period division based on the period electricity consumption characteristic parameters, obtaining electricity consumption characteristic period information, and constructing an electric energy meter adaptive adjustment strategy based on the electricity consumption characteristic period information; obtaining the electric energy meter period real-time data, and performing metering adaptive adjustment on the electric energy meter period real-time data based on the electric energy meter adaptive adjustment strategy.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: By adopting standardized preprocessing of time-period electricity consumption data streams, multi-dimensional analysis, integrated training of deep learning models, and adaptive adjustment strategies based on time-period electricity consumption characteristic parameters, this technical solution solves the technical problem that the electric energy meter in the prior art cannot be adaptively adjusted for diversified electricity consumption characteristics, resulting in insufficient metering accuracy and low electricity management efficiency. Through this technical solution, the electric energy meter can dynamically adjust the metering method in real time according to different electricity consumption characteristics and time periods, realize high-precision electricity consumption metering and more intelligent electricity consumption management, and improve the management efficiency of the power system and the user's electricity consumption experience.

[0008] 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. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 A flow chart of a method for adaptively adjusting an electric energy meter based on time period electricity consumption data analysis is provided for an embodiment of the present application.

[0010] Figure 2 A flow chart of constructing an integrated analysis network for electricity consumption data in a method for adaptively adjusting an electric energy meter under analysis of electricity consumption data during a time period is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0011] The overall idea of ​​the technical solution provided by this application is as follows: The embodiment of the present application provides an adaptive adjustment method for an electric energy meter under the analysis of electricity consumption data during a certain period. By collecting the electricity consumption data of target users during a certain period, standardized preprocessing and multi-dimensional analysis are performed to identify the electricity consumption characteristics and build an adaptive adjustment strategy. The strategy makes real-time adaptive adjustments to the electric energy meter based on the electricity consumption characteristic parameters, thereby improving the accuracy of electricity metering and the efficiency of electricity management. The entire process combines data preprocessing, characteristic analysis, deep learning models and adaptive adjustment strategies to ensure that the electric energy meter can flexibly respond to different electricity consumption demands and changes, and optimize electricity management and metering accuracy.

[0012] After introducing the basic principles of the present application, various non-limiting implementation methods of the present application will be specifically described below in conjunction with the drawings in the specification.

[0013] Examples, such as Figure 1 As shown, the embodiment of the present application provides an electric energy meter adaptive adjustment method under time period electricity consumption data analysis, the method comprising: Step S100: acquiring the target user's time period electricity consumption data stream, and performing standardization preprocessing on the time period electricity consumption data stream through a data preprocessing program to obtain a time period electricity consumption standard data stream.

[0014] Specifically, the target user's time period electricity consumption data stream refers to the electricity consumption data collected from a specific user and arranged in chronological order, such as hourly electricity consumption records. The source of the data stream can be a smart meter, electricity terminal equipment or other monitoring equipment. The data preprocessor refers to the algorithm or process for cleaning, filtering and formatting the collected raw data to improve the data quality and ensure the accuracy of subsequent analysis. The time period electricity consumption standard data stream refers to the high-quality, standardized data stream after preprocessing, which meets the format and quality requirements required for analysis and can be used for further feature analysis or modeling.

[0015] First, collect the original time period electricity consumption data stream from the target user's smart meter or other monitoring equipment. These data include timestamps, electricity consumption, equipment status, etc., but the original data has abnormal values ​​(such as extreme electricity consumption), missing values ​​(such as unrecorded time periods) or inconsistent formats. In the preprocessing stage, use the following methods and tools to gradually process the data, such as abnormal data cleaning, missing value filling, time series alignment, etc. Finally, after completing all preprocessing operations, a standardized time period electricity consumption data stream is generated, in which the data is free of abnormalities, missing values, and time points are aligned to meet the needs of subsequent analysis.

[0016] This step improves the integrity and standardization of the data. After preprocessing, the obtained standard data stream is complete and consistent, providing a reliable data basis for the adaptive adjustment strategy of the smart energy meter.

[0017] Step S200: define power consumption data dimension information and power consumption index information, classify and identify the power consumption standard data stream of the time period according to the power consumption data dimension information and power consumption index information, and obtain a multi-dimensional time period power consumption index data set.

[0018] Specifically, the dimensional information of electricity consumption data refers to the features in electricity consumption data that can be classified and analyzed from different angles. Common dimensions include: time dimension: divided by time units such as hours, days, and months; location dimension, divided according to geographical location or user location; device dimension is divided based on device category. Electricity consumption indicator information refers to key data or performance indicators related to electricity consumption, which are used to evaluate electricity consumption behavior or electricity use efficiency. Common indicators include: electricity consumption, electricity cost, load rate, etc. Multi-dimensional time period electricity consumption indicator data set refers to a data set that is classified and identified based on multiple dimensions and indicators. These data sets are organized according to different dimensions and indicators and can be used for further analysis and decision support.

[0019] First, define the dimension information and index information of electricity consumption data. You can select multiple dimensions and indicators according to the needs of data analysis. For example, select the time dimension (such as hour, day, week), location dimension (such as region, floor) and equipment dimension (such as air conditioner, lighting fixture, etc.). At the same time, select commonly used electricity consumption indicators such as electricity consumption (unit: kWh), electricity cost (unit: yuan) and load rate (unit: %). These dimensions and indicators will serve as the basis for subsequent data classification and identification.

[0020] Next, based on the defined dimension information and indicator information, the standard data stream of electricity consumption in each period is classified and identified. The specific method is to first identify each electricity consumption data according to dimensions such as time, location, and device. Then, the data is grouped according to these dimensions, and the electricity consumption indicators of each group are calculated separately, such as electricity consumption per hour, per region, or per device. Python can be used for data processing. Group by dimensions such as time, location, or device, and then calculate the relevant indicators. Finally, the generated multi-dimensional period electricity consumption indicator dataset will contain electricity consumption data under each dimension.

[0021] This step can analyze the power consumption data of each period from multiple angles by defining dimension information and indicator information and classifying and identifying them, thus improving the level of data processing refinement. In practical applications, this method can make power management decisions more accurate.

[0022] Step S300: constructing an integrated analysis network for electricity consumption data, and performing an electricity consumption characteristic analysis on the multi-dimensional time period electricity consumption index data set based on the integrated analysis network for electricity consumption data to determine the time period electricity consumption characteristic parameters.

[0023] Specifically, the electricity consumption data integrated analysis network refers to a network system that combines multiple data sources, dimensions and indicators for analysis. In-depth analysis is performed by integrating multi-dimensional data. The goal of the analysis network is to mine the potential laws and characteristics in electricity consumption data through learning and pattern recognition. Electricity consumption characteristic analysis refers to the analysis of the time characteristics in electricity consumption data, which usually includes: fluctuations in electricity consumption, peak and trough periods, electricity load characteristics, and time period electricity consumption characteristic parameters.

[0024] First, build an electricity data integration and analysis network. The core task of this network is to integrate data from different dimensions to form a multi-dimensional data set. Data integration includes the integration of multiple indicators such as time period electricity consumption, electricity cost, load rate, etc. In this process, deep learning or machine learning methods are usually used to extract data features. In specific implementation, neural networks (such as LSTM, long short-term memory network) can be used to capture the time series characteristics and fluctuations in electricity consumption data. By training a large amount of time period data, neural networks can identify periodic changes, fluctuation patterns, and peak and trough periods of electricity consumption. For example, LSTM can be used to learn daily and monthly electricity consumption patterns, and identify peak and trough periods of electricity consumption through its time series analysis capabilities.

[0025] Next, use data processing tools to preprocess and train the data. Standardize and normalize the electricity consumption data to ensure data quality and model training results. During the analysis process, use cluster analysis (such as K-means) or classification algorithms (such as decision trees and support vector machines) to identify peak hours, valley hours, and other important features in electricity consumption data.

[0026] By building an integrated analysis network for electricity consumption data, it is possible to accurately identify and predict electricity consumption characteristics in different time periods. This analysis helps to clarify electricity demand during peak and off-peak hours, and provides data support to optimize the allocation of power resources and energy management.

[0027] Step S400: Divide the time periods based on the power consumption characteristic parameters of the time periods, obtain power consumption characteristic time period information, and construct an adaptive adjustment strategy for the electric energy meter according to the power consumption characteristic time period information.

[0028] Specifically, the time period power consumption characteristic parameters refer to a series of characteristic data obtained after analyzing the power consumption behavior of the target users in different time periods. These parameters include but are not limited to the fluctuation of power consumption, peak and trough periods of power consumption, and the stability of power consumption. It reflects the distribution of power demand in a day or a week. Time period division refers to dividing a day or other time period into multiple different time periods based on the time period power consumption characteristic parameters. These time periods represent different power consumption characteristics, such as peak, trough and normal periods. The power consumption characteristic time period information is the detailed information of each time period obtained on the basis of time period division, including the power consumption characteristics of the time period (for example: peak period, trough period, stable period, etc.), and the specific power consumption mode related to the time period. The power meter adaptive adjustment strategy is a strategy for adjusting the measurement and metering method of the power meter according to the power consumption characteristic information of the time period. Through the adaptive adjustment strategy, the power meter can automatically adjust its measurement parameters in different power consumption periods to better reflect the actual power consumption.

[0029] First, by analyzing the power consumption characteristic parameters of the time period, the power consumption of the target user is divided into multiple time periods according to time. These time periods usually reflect the different characteristics of the user's power consumption. For example, suppose that by analyzing the power consumption data of a residential area, the following power consumption characteristic parameters are obtained: peak time period (for example, from 9 am to 11 am, the power consumption is large and the fluctuation is large); valley time period (for example, from 1 am to 3 am, the power consumption is low and the fluctuation is small); stable time period (for example, from 2 pm to 4 pm, the power consumption is stable and the fluctuation is moderate). Based on these characteristic parameters, these different time periods can be divided and the corresponding power consumption characteristic time period information can be obtained. Next, according to the different power consumption characteristic time period information, the adaptive adjustment strategy of the electric energy meter is constructed. For example, for peak time periods, the electric energy meter can increase the sampling frequency to more accurately capture the highly volatile power consumption data; while for valley time periods, the electric energy meter can reduce the sampling frequency to reduce the burden of data collection and save battery or storage space. For stable time periods, the electric energy meter can use standard sampling frequency and accuracy. The design of these strategies can not only improve the efficiency of data collection, but also maintain high data accuracy during high-load periods, thereby achieving more accurate measurement and optimization.

[0030] This adaptive adjustment strategy can significantly improve the measurement accuracy and data collection efficiency of the energy meter. During peak hours, the system can more accurately reflect the user's electricity demand and avoid errors; during off-peak hours, by reducing the sampling frequency and accuracy, battery and storage resources are saved and the service life of the energy meter is extended.

[0031] Step S500: acquiring the real-time data of the electric energy meter during a period of time, and performing metering adaptive adjustment on the real-time data of the electric energy meter during a period of time based on the electric energy meter adaptive adjustment strategy.

[0032] Specifically, the real-time data of the energy meter refers to the real-time electricity consumption data collected by the energy meter during a specific period of time. It usually includes electricity consumption information such as power consumption, voltage, current, and power. Metering adaptive adjustment means that in different electricity consumption periods or electricity demand situations, the energy meter will make real-time adjustments to the data collection process according to the set adaptive adjustment strategy, thereby optimizing measurement accuracy and data processing efficiency.

[0033] First, the electric energy meter will collect data in real time according to the preset sampling period (for example, every minute, every hour, etc.). The electric energy meter monitors and records the user's electricity consumption behavior through built-in sensors, and obtains data including instantaneous power and cumulative power consumption. When obtaining real-time data for a period of time, the electric energy meter can collect data according to the period of time or a specific time interval. These data can be transmitted to the data processing center through a wireless network or a wired communication interface for further analysis. Next, the electric energy meter will adjust the acquired real-time data for the period of time based on the pre-built adaptive adjustment strategy. The sampling frequency adjustment uses a dynamic adjustment algorithm to adjust the sampling frequency of the electric energy meter according to the signal strength and the power consumption characteristics of the period of time. For example, a control algorithm (such as PID control) can be used to adjust the sampling frequency according to the fluctuation of power consumption.

[0034] Through this step, the energy meter can provide the best data collection effect at different power consumption periods. In addition, based on the adaptive adjustment strategy, the energy meter not only improves the measurement accuracy, but also reduces unnecessary power consumption and storage requirements through dynamic adjustment, optimizing the efficiency and reliability of the overall system.

[0035] Furthermore, the method of obtaining a standard data stream for electricity consumption in a time period includes: identifying abnormal data on the electricity consumption data stream in the time period to obtain an abnormal electricity consumption data set, wherein the abnormal electricity consumption data set includes outliers, missing values, and over-threshold data; determining an abnormal value processing program and a data alignment program according to the data preprocessing program; performing processing program matching and abnormal value cleaning on the abnormal electricity consumption data set based on the abnormal value processing program to determine an available data stream for electricity consumption in the time period; and using the data alignment program to perform time series alignment on the available electricity consumption data stream in the time period to obtain the standard data stream for electricity consumption in the time period.

[0036] Specifically, abnormal data identification refers to the detection of unexpected parts of data through specific algorithms or rules, including outliers, missing values, and over-threshold data. Among them, outliers refer to electricity consumption data that deviates significantly from the normal range. For example, the time period electricity consumption data of a certain user is usually between 0.5-2kWh, but a certain period is recorded as 50kWh, which is an outlier. Missing values ​​refer to the missing electricity consumption data of certain time periods due to equipment failure or collection problems. For example, the data at 8:00 is empty. Over-threshold data refers to data that exceeds the reasonable threshold set by the system. For example, the single-hour electricity consumption of an industrial equipment exceeds the set upper limit of 100kWh. The outlier handler is a set of methods for identifying and processing abnormal data, including cleaning, interpolation, substitution, etc. The data alignment program is to adjust the data collected in different time periods into a unified time series format to ensure the accuracy of the analysis. The available data stream of time period electricity consumption refers to the high-quality data stream after the abnormal data is cleaned and the abnormal and invalid data is removed.

[0037] First, identify abnormal data in the target user's electricity consumption data stream during the period to obtain an abnormal electricity consumption data set. Use statistical methods (such as box plot method or standard deviation threshold method) to identify outliers, use data inspection tools (such as pandas's isnull() method) to detect missing values, and mark over-threshold data based on the set electricity safety or equipment restriction threshold.

[0038] Next, according to the data preprocessing procedure, determine the applicable outlier processing procedure and data alignment procedure. For outliers, you can choose to remove or replace. For missing values, the mean interpolation method can be used to fill, such as the electricity consumption at 10:00 is replaced by the mean of 9:00 and 11:00 (1.8kWh). The over-threshold data can be adjusted to the maximum allowable value (such as 100kWh) according to the set ratio. After completing the abnormal data cleaning, the available data stream of electricity consumption in the time period is obtained. Subsequently, the data alignment program is used to align the time series. Finally, the preprocessed standard data stream contains the normalized electricity consumption records for each hour.

[0039] Through this step, abnormal data (such as outliers, missing values, and over-threshold data) are effectively removed, and consistent processing of time series is achieved, which significantly improves the integrity and standardization of data. This lays a solid foundation for subsequent power consumption characteristics analysis and adaptive adjustment of smart energy meters, and improves the prediction accuracy of data-driven models and the efficiency of power consumption management.

[0040] Further, such as Figure 2As shown, the construction of the integrated analysis network for electricity consumption data includes: collecting and acquiring multi-source electricity consumption data streams from multi-terminal users, performing standardized preprocessing on the multi-source electricity consumption data streams through a data preprocessing program to obtain multi-source electricity consumption standard data streams; acquiring electricity consumption data analysis targets, performing characteristic index decomposition on the electricity consumption data analysis targets, and obtaining a set of electricity consumption data characteristic analysis indicators; performing distributed training on the multi-source electricity consumption standard data streams according to the set of electricity consumption data characteristic analysis indicators to obtain a multi-terminal electricity consumption data analysis network set; extracting a model parameter set of the multi-terminal electricity consumption data analysis network set, integrating the training and fusion of the model parameter set, and constructing the integrated analysis network for electricity consumption data.

[0041] Specifically, the multi-source electricity consumption data stream of multi-terminal users refers to the electricity consumption data stream from multiple different users or multiple different power sources. These data streams may include electricity consumption information in different regions, different devices or different time periods. The multi-source electricity consumption standard data stream refers to the electricity consumption data stream after standardization preprocessing. The standardization process adjusts the data from different sources to a unified format and range so that they can be compared and analyzed in the same data set. The electricity consumption data analysis target refers to the specific goal or requirement of this data analysis. The electricity consumption data characteristic analysis indicator set refers to the extraction of a series of related indicators by decomposing the analysis target, including: electricity consumption, electricity consumption period, load change, peak-valley difference, etc., which are used to describe and quantify electricity consumption characteristics. Distributed training refers to the process of simultaneously training data on multiple computing nodes. The multi-terminal electricity consumption data analysis network set refers to the network of multiple analysis model sets, each model focusing on the electricity consumption analysis of a certain user group or a certain period of time. The model parameter set refers to the process in which the model continuously adjusts its internal parameters (such as weights and biases) to adapt to the data during the training process. The extracted parameter set contains the optimal parameters learned during the training process, which are the core part of the analysis network. Ensemble training fusion refers to fusing the parameters of multiple different training models to improve the overall performance and accuracy of the model.

[0042] First, electricity consumption data is collected from multiple sources to form a "multi-source electricity consumption data stream for multiple users". These data come from energy metering devices in homes, commercial buildings or industrial equipment, and each user's electricity consumption data has different formats and time granularity. In order to ensure that data from different sources can be analyzed uniformly, a data preprocessing program is required, which includes data cleaning, denoising, missing value filling, outlier processing and other processes. At this point, the pandas library in Python can be used for data cleaning, and the StandardScaler of scikit-learn can be used for data standardization to ensure that data from different users have the same scale and unit.

[0043] Then, the analysis objectives of power consumption data are clarified, such as analyzing the peak power consumption period in a certain area, or identifying devices with large power load fluctuations. These objectives are disassembled through the power consumption data characteristic analysis indicator set to obtain indicators related to the analysis objectives (such as power consumption peak, load fluctuation, etc.). For example, for the peak power consumption analysis objectives, indicators such as "maximum power consumption" and "power load change rate" can be extracted.

[0044] Next, we analyze the indicator set based on the characteristics of the electricity consumption data and conduct distributed training on the standardized data. Distributed training can utilize multiple computing resources to increase the speed of model training.

[0045] The multi-terminal power consumption data analysis network set refers to a collection of multiple training models. Each model is trained independently and learns for different user terminals and different power consumption scenarios. After the training of each model is completed, a set of model parameter sets will be obtained. These parameters are the results of the model optimization during the training process and represent the abstraction of power consumption characteristics.

[0046] Ultimately, all model parameters are combined through ensemble training to improve the overall performance of the model. Ensemble methods such as model averaging or weighted voting can combine the advantages of different models, reduce bias and improve adaptability to multi-end data. For example, data analysis results from different users can be combined to achieve more accurate time period electricity consumption forecasts.

[0047] This step can achieve accurate analysis of electricity consumption characteristics of different user groups, different time periods, and different devices through integrated analysis of electricity consumption data from multiple terminals and sources. Through this integrated analysis network, key characteristics such as peaks, valleys, and fluctuations in electricity consumption can be efficiently identified, thereby providing strong support for the adaptive adjustment of electricity meters. In addition, the results of the integrated analysis can be applied to systems such as dynamic electricity pricing and demand response to further optimize the allocation of power resources.

[0048] Furthermore, the method of obtaining a multi-terminal electricity consumption data analysis network set includes: performing data association extraction on the multi-source electricity consumption standard data stream according to the electricity consumption data characteristic analysis indicator set to obtain a multi-source electricity consumption indicator data set; performing characteristic analysis on the multi-source electricity consumption indicator data set to determine electricity consumption data characteristic information, and selecting to obtain a deep learning combination network based on the electricity consumption data characteristic information; using the deep learning combination network to perform distributed training on the multi-source electricity consumption indicator data set to obtain a multi-terminal electricity consumption indicator analysis network set; and integrating the multi-terminal electricity consumption indicator analysis network set in parallel according to the multi-terminal users to obtain the multi-terminal electricity consumption data analysis network set.

[0049] Specifically, data association extraction refers to extracting relevant features or associated data from multi-source electricity consumption data. Through this extraction, the mutual connection or similarity between multiple data can be identified. The multi-source electricity consumption index data set is a data set obtained by association extraction of electricity consumption data from multiple sources. It contains the collated data of key indicators for each user and each time period after characteristic analysis, usually in a multi-dimensional structure. The characteristic information of electricity consumption data is the characteristic information about electricity consumption behavior obtained after analyzing the data. It includes but is not limited to the periodic changes in electricity consumption, the characteristics of electricity consumption during peak and trough periods, abnormal fluctuations and other information. The deep learning combination network is a technology that combines multiple deep learning models. It can improve the accuracy and robustness of prediction by training different neural network models in parallel and combining their results. The multi-terminal electricity consumption index analysis network set is a collection of multiple deep learning network models, each of which is used to analyze the indicator data of different users or different electricity consumption environments. By training and optimizing each network, it can better adapt to different electricity consumption needs and changes. Parallel connection integration is to connect multiple trained network models in parallel according to user groups (or other dimensions) and integrate them into an integrated analysis network to improve the adaptability and analysis capabilities of the system.

[0050] First, data association extraction is performed on the multi-source electricity consumption data stream. The purpose of this process is to extract valuable and relevant features from the electricity consumption data from various sources. For example, by analyzing the electricity consumption of different users in a specific period of time, the relationship between factors such as weather, season, and device type and electricity consumption can be found. These associated data are used to construct a multi-source electricity consumption indicator dataset.

[0051] Next, by analyzing the characteristics of the multi-source electricity consumption indicator dataset, key data features for training are extracted. These features include the peak load period of each user, electricity consumption fluctuations, etc. These features can help establish the user's electricity consumption behavior model. After that, a deep learning combination network is selected to train the extracted multi-source electricity consumption indicator dataset. This step involves using deep learning algorithms such as convolutional neural networks (CNN) or long short-term memory networks (LSTM) to model the temporal characteristics of electricity consumption data. When selecting a deep learning model, choose a suitable network architecture based on the characteristics of different users and electricity consumption environments. For example, LSTM can be used to process electricity consumption data with time dependencies, and CNN is suitable for processing data related to local features.

[0052] During the training process, distributed training technology is used to accelerate the learning process of the model. The training tasks of the model will be assigned to multiple computing nodes (such as multiple GPUs or servers), each node processes a part of the data, and finally the results of all nodes are integrated. Use deep learning frameworks such as TensorFlow or PyTorch to implement distributed training.

[0053] The results of multiple models obtained through training are connected and integrated in parallel to form a comprehensive multi-terminal power consumption data analysis network set. This analysis network set includes the power consumption characteristic analysis capabilities of each end (user or region), and can flexibly respond to the needs of different power consumption scenarios. For example, for different regions or different power consumption environments, the network will select the most suitable analysis model for real-time data prediction and adjustment based on the pre-set strategy.

[0054] This step builds and trains a deep learning combined network to comprehensively and accurately analyze complex electricity consumption data from multiple users. Distributed training and feature analysis not only improves the efficiency of analysis, but also enhances the model's adaptability to different electricity consumption environments.

[0055] Furthermore, the method of obtaining a multi-source electricity consumption indicator data set includes: performing a criticality assessment on each analysis indicator in the electricity consumption data characteristic analysis indicator set to obtain a set of analysis indicator criticality coefficients; determining a data association depth factor based on the set of analysis indicator criticality coefficients; performing data association extraction on the multi-source electricity consumption standard data stream according to the electricity consumption data characteristic analysis indicator set to obtain an initial indicator association data set; taking the initial indicator association data set as the association center, performing data collection and expansion on the multi-source electricity consumption standard data stream based on the data association depth factor to obtain the multi-source electricity consumption indicator data set.

[0056] Specifically, the analysis indicator critical coefficient set is a set of coefficients used to measure the influence of different analysis indicators on the final analysis results. By evaluating the importance of each indicator, it is possible to determine which indicators have higher predictive value in the analysis of electricity consumption data. For example, the time period (daytime and nighttime) has a greater impact on electricity consumption than seasonal changes, so its coefficient is higher. The data association depth factor is a factor used to strengthen the correlation between data. The depth factor usually refers to the complex relationship between factors such as energy, time, and environment calculated by certain mathematical or statistical methods in multidimensional data analysis, which helps to improve the generalization ability of the model in different data dimensions. The initial indicator association data set is a data set generated according to the characteristic analysis indicator set during the data association extraction process, which contains preliminary correlation information between multiple indicators. Through this data set, the basic connection between different indicators can be understood. Data acquisition expansion refers to further expanding the data based on the initial data set using certain technical means and algorithms in order to explore deeper associations and laws. Through acquisition expansion, more sample data can be added or new features can be synthesized, thereby enriching the diversity and representativeness of the data set.

[0057] First, a criticality assessment is performed on each analysis indicator in the power consumption data characteristic analysis indicator set. The goal of the criticality assessment is to quantify the importance of each indicator to the overall analysis model. For example, in power consumption analysis, "use time period" may have a greater impact on power consumption than "outdoor temperature", so the criticality coefficient of "use time period" will be higher. Through algorithms such as information gain or Shannon entropy, a criticality coefficient can be assigned to each indicator to reflect its relative importance in power consumption data analysis.

[0058] Then, based on these evaluated critical coefficient sets, the data association depth factor is calculated. For example, if the power demand in a certain period of time shows significant changes within a certain temperature range, these temperature ranges can be used as depth factors to enhance the model's understanding of the changes in power consumption. The calculation of data association depth factors is usually achieved through statistical regression analysis or machine learning algorithms (such as random forests and support vector machines) to help establish efficient associations between data of different dimensions.

[0059] Next, based on the power consumption data characteristic analysis indicator set, data association extraction is performed on the multi-source power consumption standard data stream to obtain the initial indicator association data set. The purpose of data association extraction is to discover the potential relationship between data. For example, in different seasons, the use of air conditioning and heating will affect the change of power consumption data. Therefore, the correlation between season and power consumption can be found by analyzing historical data to form an initial association data set.

[0060] After obtaining the initial indicator association data set, the multi-source electricity consumption standard data stream is expanded based on the data association depth factor. The specific method is to introduce the correlation depth factor between the data into the data stream expansion process. For example, if a high electricity consumption peak is strongly correlated with a high temperature in a specific time period, the temperature peak period and the high electricity consumption scene can be synthesized by expanding the data set to provide more diverse sample data for the model. Ultimately, the multi-source electricity consumption indicator data set obtained through the above steps is a more comprehensive and rich data set that can better support model training and electricity consumption behavior analysis.

[0061] Through this step, comprehensive analysis and processing of multi-source electricity consumption data can be achieved, thereby increasing the diversity and complexity of the data. By introducing key evaluation and data association depth factors, deeper laws in electricity consumption data can be mined to improve the accuracy of model prediction. Data expansion and enhancement enable the model to adapt to more diverse scenarios and avoid overfitting problems caused by insufficient data.

[0062] Furthermore, the construction of the electricity consumption data integrated analysis network includes: integrating the model parameter set for training and fusion to obtain an initial data integrated analysis network, and evaluating and verifying the network performance parameters of the initial data integrated analysis network; performing an optimization strategy analysis on the initial data integrated analysis network based on the network performance parameters to determine the integrated analysis network optimization strategy; iteratively updating the parameters of the initial data integrated analysis network based on the integrated analysis network optimization strategy until the preset performance parameters are met to obtain the electricity consumption data integrated analysis network.

[0063] Specifically, the initial data integration analysis network refers to the preliminary network model obtained after fusing models or data sets from multiple sources. By integrating different data processing models, a network that can analyze data from multiple angles and in all directions is formed. Network performance parameters are used to evaluate the performance of the integrated analysis network, usually including accuracy, recall, precision, F1 value, training loss, test error, etc. Optimization strategy analysis refers to the use of various optimization algorithms to improve the performance of the integrated analysis network. Optimization strategies include adjusting the network architecture, modifying the learning rate, applying regularization techniques, using more efficient loss functions, etc. Parameter iterative update refers to the update of model parameters (such as weights and biases) through a continuous iterative process during network training. After each update, the model will make a prediction based on the new parameters, calculate the loss and perform back propagation to adjust the parameters until the expected performance indicators are achieved. Preset performance parameters refer to the network performance requirements that have been set before network optimization. Usually include the minimum error range or the maximum tolerable test error, etc. Through these preset parameters, developers can tune the network within certain performance standards.

[0064] First, the obtained model parameter set is integrated and trained. The process of integrated training and fusion is to merge the parameter sets of multiple independently trained models and obtain a more generalized preliminary model through appropriate algorithms (such as weighted average, voting mechanism, etc.). This fusion method can improve the robustness and accuracy of the overall model. Next, the network performance parameters of the initial data integration analysis network are evaluated. The network performance parameters are generally determined by evaluation on the test set. Cross-validation methods (for example, K-fold cross-validation) can be used to obtain the performance of the network on different data sets to evaluate the generalization ability of the model. Evaluation indicators include prediction accuracy, recall rate, F1 value, etc. According to the evaluation results, optimization strategy analysis is performed. If the model performance is not ideal, several optimization strategies will be adopted, such as: adjusting the learning rate (gradually reducing the learning rate through the learning rate scheduler to avoid overfitting); modifying the network structure (for example, increasing or decreasing the number of neural network layers); using regularization techniques (such as L2 regularization or Dropout) to reduce overfitting; using more efficient optimization algorithms (such as Adam, RMSProp, etc.) to accelerate the training process. Through these optimization strategies, the performance of the network can be effectively improved, errors can be reduced, and it can be made more in line with actual needs.

[0065] Finally, according to the determined optimization strategy, the parameters are iteratively updated. In this stage, the parameters of the model are gradually adjusted through the back propagation algorithm to achieve the optimization goal. After each update, the prediction ability of the network will be improved to a certain extent until the preset performance parameters are met. Usually a stopping criterion is set, such as reaching a certain performance indicator (such as the test set error is lower than a certain value) or stopping after a fixed number of iterations.

[0066] Through the above steps, the electricity consumption data integration and analysis network obtained will have higher prediction accuracy and stronger generalization ability. By integrating the parameters of multiple models, the system can adapt to the electricity consumption patterns of different users and continuously improve the prediction effect through optimization strategies. In addition, by iteratively updating parameters, it can ensure that the network maintains efficient and stable performance in actual applications, thereby supporting the adaptive adjustment of electricity meters and realizing more intelligent electricity consumption management.

[0067] Furthermore, the adaptive adjustment of the real-time data of the electric energy meter time period based on the electric energy meter adaptive adjustment strategy includes: obtaining the current power consumption characteristic time period of the real-time data of the electric energy meter time period; determining the current time period adjustment strategy based on matching the current power consumption characteristic time period with the electric energy meter adaptive adjustment strategy; obtaining the electric energy meter time period metering adjustment parameters according to the current time period adjustment strategy; and adaptively adjusting the real-time data of the electric energy meter time period based on the electric energy meter time period metering adjustment parameters.

[0068] Specifically, the current power consumption characteristic period refers to the current power consumption period according to the preset period division standard during the operation of the power meter. For example, a day is divided into different time periods (such as daytime, evening, late night, etc.), and the power consumption characteristics of each period are different, such as the degree and frequency of power consumption fluctuations. The time period adjustment strategy refers to the adjustment rules of the power meter formulated for specific time periods (such as peak period, trough period, etc.). Different power consumption periods will determine the corresponding adjustment strategies according to the fluctuation characteristics of power consumption, such as increasing the sampling frequency and improving the measurement accuracy. The power meter period measurement adjustment parameters refer to the parameters obtained according to the adaptive adjustment strategy of the power meter within a specific time period, which are used to adjust the working mode of the power meter. These parameters include sampling frequency, accuracy setting, etc.

[0069] The electric energy meter first obtains the power consumption data of the current period through real-time monitoring, and determines the current power consumption period according to the predetermined period division rules. Next, the system will find and match the corresponding adaptive adjustment strategy according to the current power consumption characteristic period. For example, if the system recognizes that the current period is a peak period, the system will select an adjustment strategy with a higher sampling frequency and higher precision; if the current period is a valley period, it will select a strategy with lower sampling frequency and precision to save battery and storage space. According to the matched period adjustment strategy, the electric energy meter will calculate the required metering adjustment parameters. These parameters include: Sampling frequency: During peak periods, the sampling frequency needs to be increased to capture changes in power consumption, while during valley periods, the sampling frequency can be reduced. Measurement accuracy: During periods of heavy load, the electric energy meter needs to improve measurement accuracy to ensure accurate measurement; during periods of low load, the accuracy can be moderately reduced to save computing resources.

[0070] Finally, the electric energy meter makes adaptive adjustments to the real-time data of the time period based on the obtained adjustment parameters. For example: Increase the sampling frequency: During peak hours, the electric energy meter increases the sampling frequency (such as increasing the sampling from once every 10 seconds to once every second) to ensure the accuracy and real-time nature of the data. Reduce the sampling frequency: During off-peak hours, the electric energy meter can reduce the frequency of data collection, thereby reducing the amount of data and avoiding excessive collection.

[0071] Through this step, the energy meter can automatically adjust its working mode according to the real-time power consumption characteristics, achieving accurate real-time data collection and flexible adaptive adjustment. It can provide high-precision measurements during peak hours to ensure data accuracy, while saving energy and storage resources during off-peak hours. Overall, this intelligent adjustment strategy not only improves the quality and efficiency of data, but also extends the service life of equipment, reduces unnecessary energy consumption, and improves the economy and reliability of the power metering system.

[0072] Furthermore, the method includes: monitoring and obtaining the adjustment response rate of the electric energy meter, determining the adjustment delay coefficient of the electric energy meter according to the adjustment response rate of the electric energy meter; performing a metering loss analysis on the adjustment delay coefficient of the electric energy meter to obtain the metering loss factor of the electric energy meter, and determining the metering compensation parameter according to the metering loss factor of the electric energy meter; and performing compensation optimization and correction on the metering adjustment parameters of the electric energy meter time period based on the metering compensation parameter.

[0073] Specifically, the energy meter adjustment response rate refers to the speed at which the energy meter responds to the instruction after receiving the adjustment instruction. This rate measures the time required for the energy meter to start the adjustment operation from receiving the adjustment command. The energy meter adjustment delay coefficient is a coefficient generated by the time difference between the actual response and the theoretical response when the energy meter is performing adaptive adjustment. This coefficient represents the degree of delay in the energy meter response adjustment strategy, which is usually related to factors such as the hardware performance of the device and network delay. The energy meter measurement loss factor refers to the decrease in the accuracy of the measurement data or the loss of data in certain time periods caused by factors such as adjustment delay during the energy meter adjustment process. The measurement loss factor is a coefficient used to describe these losses, reflecting the measurement error caused by the adjustment delay. The measurement compensation parameter refers to the parameter used to correct the measurement loss of the energy meter caused by adjustment delay or other factors. The compensation parameter is usually determined by analyzing the loss factor, with the aim of making the metering data of the energy meter more accurate. The time period measurement adjustment parameter refers to the measurement parameter adjusted according to the power consumption characteristics and time period requirements during the energy meter adaptive adjustment process, including sampling frequency, accuracy, etc.

[0074] When the energy meter is adaptively adjusted, the system will first monitor the energy meter's adjustment response rate, that is, the time from when the energy meter receives the adjustment command to when it starts to implement the adjustment. For example, if the energy meter needs to adjust the sampling frequency according to a certain power consumption characteristic period, the system will record the time from when the adjustment command is issued to when the sampling frequency actually changes. Based on the monitored energy meter adjustment response rate, the system calculates the adjustment delay coefficient. The adjustment delay coefficient is used to reflect the degree of delay in the energy meter adjustment process. Assuming that the adjustment response of the energy meter takes 1 second, the system will set a coefficient to represent the difference between the adjustment response and the expectation. For example, a delay coefficient of 0.95 means that the actual effect of the adjustment response is delayed by 5%.

[0075] Next, the system performs a metering loss analysis based on the meter adjustment delay coefficient. Since adjustment delays can cause distortion or errors in data measurement, the system calculates the metering loss caused by the adjustment delay. For example, if the delay causes a decrease in sampling accuracy, the system can analyze the difference between the actual measurement value and the expected value based on historical data, and then derive the metering loss factor.

[0076] Based on the above analysis, the metering loss factor of the energy meter will be calculated. For example, if the energy meter has measurement errors during peak hours due to delayed adjustment, the loss factor is 0.8, which means that the actual measurement value is 80% of the expected value. The system will determine the metering compensation parameters that need to be applied based on this factor, such as increasing the measurement accuracy of certain time periods or adjusting the sampling frequency.

[0077] Finally, the system will optimize and correct the period metering adjustment parameters based on the metering compensation parameters. For example, if the electric energy meter has a 5% error due to delayed adjustment in a certain period, the system will increase the sampling frequency or measurement accuracy by 5% during the next period adjustment to compensate for the previous metering loss and ensure that the final measurement result is more accurate.

[0078] Through this step, the energy meter can effectively identify the delay and metering loss in the adjustment response process, and take appropriate compensation measures to optimize the accuracy of the data. This not only improves the accuracy of the measurement, but also ensures that the energy meter can continue to provide accurate metering data in different power consumption characteristic periods, avoiding data deviations caused by delays and other factors, thereby improving the reliability and accuracy of the entire power metering system.

[0079] In summary, the method for adaptively adjusting an electric energy meter under time period electricity consumption data analysis provided by the embodiment of the present application has the following technical effects: 1. Through standardized preprocessing of time period electricity consumption data and identification of abnormal data, abnormal data can be effectively eliminated and data quality can be improved, thereby providing accurate basic data for the adaptive adjustment of the electricity meter and ensuring the accuracy and reliability of electricity metering.

[0080] 2. By building an integrated analysis network for electricity consumption data and conducting distributed training, we can deeply analyze multi-source electricity consumption data and extract multi-dimensional electricity consumption characteristics, thereby improving the prediction accuracy and adaptive adjustment capability of user electricity consumption behavior and optimizing the adjustment strategy of the electricity meter.

[0081] 3. Through the evaluation of key analysis indicators and the use of data correlation depth factors, relevant information of electricity consumption data can be extracted and expanded more accurately, which improves the coverage and depth of electricity consumption data sets and facilitates more detailed user demand analysis.

[0082] 4. By compensating and optimizing the delay coefficient of the electric energy meter, the metering error caused by hardware or network delay can be reduced, and more accurate electric energy metering can be achieved, thereby improving the overall efficiency and user satisfaction of the power management system.

[0083] Any step of the method described above can be stored as a computer instruction or program in an unlimited computer memory, and can be called and recognized by an unlimited computer processor to implement any method in the embodiments of the present application, without any unnecessary restrictions.

[0084] Furthermore, the first or second mentioned above not only represents an order relationship, but also represents a specific concept, and / or refers to the selection of multiple elements individually or in whole. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application intends to include these modifications and variations.

Claims

1. The method for adaptively adjusting an electric energy meter based on the analysis of electric energy consumption data during a period of time is characterized in that: The method comprises: Collect and obtain the target user's time period electricity consumption data stream, perform standardization preprocessing on the time period electricity consumption data stream through a data preprocessing program, and obtain a time period electricity consumption standard data stream; Defining electricity consumption data dimension information and electricity consumption index information, classifying and identifying the electricity consumption standard data stream for the time period according to the electricity consumption data dimension information and electricity consumption index information, and obtaining a multi-dimensional time period electricity consumption index data set; Constructing an integrated analysis network for electricity consumption data, and performing an electricity consumption characteristic analysis on the multi-dimensional time period electricity consumption index data set based on the integrated analysis network for electricity consumption data to determine the time period electricity consumption characteristic parameters; Divide the time periods based on the power consumption characteristic parameters of the time periods, obtain power consumption characteristic time period information, and construct an adaptive adjustment strategy for the electric energy meter according to the power consumption characteristic time period information; Acquire the real-time data of the electric energy meter during the time period, and perform metering adaptive adjustment on the real-time data of the electric energy meter during the time period based on the electric energy meter adaptive adjustment strategy.

2. The method for adaptively adjusting an electric energy meter under time period power consumption data analysis as claimed in claim 1, characterized in that: The step of obtaining the standard data stream of electricity consumption during a time period includes: Performing abnormal data identification on the electricity consumption data stream during the time period to obtain an abnormal electricity consumption data set, wherein the abnormal electricity consumption data set includes outliers, missing values, and over-threshold data; According to the data preprocessing procedure, determining an outlier processing procedure and a data alignment procedure; Based on the outlier processing program, the power consumption abnormality data set is matched with a processing program and outlier cleaning is performed to determine the available data flow of power consumption in a time period; The data alignment program is used to perform time series alignment on the available data streams of electricity consumption in the time period to obtain the standard data streams of electricity consumption in the time period.

3. The method for adaptively adjusting an electric energy meter under time period power consumption data analysis as claimed in claim 1, characterized in that: The construction of the electricity consumption data integration analysis network includes: Collect and obtain multi-source electricity consumption data streams of multiple users, and perform standardization preprocessing on the multi-source electricity consumption data streams through a data preprocessing program to obtain multi-source electricity consumption standard data streams; Acquire a power consumption data analysis target, decompose characteristic indicators of the power consumption data analysis target, and obtain a power consumption data characteristic analysis indicator set; Performing distributed training on the multi-source electricity consumption standard data stream according to the electricity consumption data characteristic analysis indicator set to obtain a multi-terminal electricity consumption data analysis network set; The model parameter set of the multi-terminal power consumption data analysis network set is extracted, the model parameter set is integrated and trained, and the power consumption data integrated analysis network is constructed.

4. The method for adaptively adjusting an electric energy meter under time period power consumption data analysis as claimed in claim 3, characterized in that: The obtaining of a multi-terminal power consumption data analysis network set includes: Performing data association extraction on the multi-source electricity consumption standard data stream according to the electricity consumption data characteristic analysis indicator set to obtain a multi-source electricity consumption indicator data set; Performing characteristic analysis on the multi-source electricity consumption index data set to determine characteristic information of the electricity consumption data, and selecting to obtain a deep learning combination network based on the characteristic information of the electricity consumption data; Using the deep learning combination network to perform distributed training on the multi-source electricity consumption index data set to obtain a multi-terminal electricity consumption index analysis network set; The multi-terminal electricity consumption index analysis network set is connected and integrated in parallel according to the multi-terminal users to obtain the multi-terminal electricity consumption data analysis network set.

5. The method for adaptively adjusting an electric energy meter under time period power consumption data analysis as claimed in claim 4, characterized in that: The obtaining of a multi-source electricity consumption index data set includes: Performing a criticality assessment on each analysis indicator in the power consumption data characteristic analysis indicator set to obtain a criticality coefficient set of the analysis indicator; Determining a data association depth factor according to the analysis indicator criticality coefficient set; Performing data association extraction on the multi-source electricity consumption standard data stream according to the electricity consumption data characteristic analysis indicator set to obtain an initial indicator association data set; Taking the initial indicator association data set as the association center, data collection and expansion are performed on the multi-source electricity consumption standard data stream based on the data association depth factor to obtain the multi-source electricity consumption indicator data set.

6. The method for adaptively adjusting an electric energy meter under time period power consumption data analysis as claimed in claim 3, characterized in that: The step of constructing the electricity consumption data integration and analysis network comprises: Performing integrated training and fusion on the model parameter set to obtain an initial data integrated analysis network, and evaluating and verifying the network performance parameters of the initial data integrated analysis network; Performing an optimization strategy analysis on the initial data integration analysis network based on the network performance parameters to determine an optimization strategy for the integration analysis network; The parameters of the initial data integrated analysis network are iteratively updated based on the integrated analysis network optimization strategy until preset performance parameters are met, thereby obtaining the power consumption data integrated analysis network.

7. The method for adaptively adjusting an electric energy meter under time period power consumption data analysis as claimed in claim 1, characterized in that: The adaptively adjusting the real-time data of the electric energy meter during a period based on the electric energy meter adaptive adjustment strategy includes: Obtain the current power consumption characteristic period of the real-time data of the electric energy meter; Based on the matching of the current power consumption characteristic period and the self-adaptive adjustment strategy of the electric energy meter, the current period adjustment strategy is determined; According to the current time period adjustment strategy, obtaining the time period measurement adjustment parameters of the electric energy meter; Based on the electric energy meter period measurement adjustment parameters, the electric energy meter period real-time data is adaptively adjusted.

8. The method for adaptively adjusting an electric energy meter under time period power consumption data analysis as claimed in claim 7, characterized in that: The method comprises: Monitoring and obtaining an adjustment response rate of an electric energy meter, and determining an adjustment delay coefficient of the electric energy meter according to the adjustment response rate of the electric energy meter; Performing a metering loss analysis on the electric energy meter adjustment delay coefficient to obtain a metering loss factor of the electric energy meter, and determining a metering compensation parameter according to the metering loss factor of the electric energy meter; The metering adjustment parameters of the electric energy meter during the period are compensated, optimized and corrected based on the metering compensation parameters.

Images