Intelligent analysis method and system for energy consumption based on time series and load characteristics
Through the intelligent analysis method of energy consumption based on time series and load characteristics, the problem of traditional energy management systems dealing with complex data relationships and real-time changes is solved, and accurate prediction and optimization of energy demand is achieved, improving the efficiency and cost reduction of energy management are improved. .
Patent Information
- Application Number
- CN202410874562.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-02
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2044-07-02
AI Technical Summary
Traditional energy management systems are difficult to deal with complex data relationships and real-time changing energy needs, cannot process large amounts of data flows in real time, and lack effective tools to analyze the complex relationship between equipment load characteristics and external factors, limiting the improvement of energy efficiency and cost. further reduction.
Using an intelligent energy consumption analysis method based on time series and load characteristics, the predicted feature vector is determined by obtaining the target external factor data and equipment usage patterns, and input it into the factor analysis model to predict the equipment load and energy demand, and integrate real-time data flow , respond to external changes and optimize energy distribution and consumption strategies.
It improves the accuracy of energy forecasts and adaptability of energy management, can identify peak and trough periods of energy consumption, optimize energy allocation, reduce energy costs, and enhance the stability and security of energy supply.
Smart Images

Figure CN118428774B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of energy consumption analysis, and in particular to an intelligent analysis method and system for energy consumption based on time series and load characteristics. Background Art
[0002] Energy management is an important area in modern industrial and commercial operations, especially in the context of rising energy costs and increasingly stringent environmental regulations. Energy management refers to the planning, control and optimization of energy in a scientific and rational way to improve energy efficiency, reduce energy consumption and reduce environmental impact.
[0003] Energy management can optimize the energy use process and adopt efficient equipment and technologies, which can significantly improve energy utilization efficiency, reduce energy waste, and thus reduce energy costs. Energy management can also reduce energy consumption. By diversifying the energy structure and improving energy self-sufficiency and reserve capacity, the stability and security of energy supply can be enhanced and dependence on external energy can be reduced.
[0004] Traditional energy management systems usually rely on basic monitoring and control technologies, which can only provide limited data analysis functions and have difficulty in handling complex data relationships or real-time changing energy demands. In addition, traditional methods have obvious deficiencies in data integration, real-time data processing, and dynamic energy optimization. They cannot process large amounts of data streams in real time, and have difficulty responding quickly to sudden changes in the external environment. They lack effective tools to analyze and utilize the complex relationships between equipment load characteristics and multiple external factors, which limits the improvement of energy efficiency and further reduction of costs. Summary of the invention
[0005] The present application provides a method and system for intelligent analysis of energy consumption based on time series and load characteristics, so as to solve the problem that the existing technology is difficult to handle complex data relationships or real-time changing energy demands.
[0006] In a first aspect, the present application provides an intelligent analysis method for energy consumption based on time series and load characteristics, comprising:
[0007] Acquire at least one target external factor data affecting the energy consumption of the device within the time period to be predicted, wherein the target external factor is determined according to the correlation between the energy consumption of the device and the external factor;
[0008] Determining a prediction feature vector corresponding to a time period to be predicted according to the at least one target external factor data and a device usage pattern of the key device;
[0009] The prediction feature vector is input into a factor analysis model to obtain the equipment prediction load in the prediction period and the energy demand corresponding to the equipment prediction load. The factor analysis model is determined based on the first historical energy consumption data and historical equipment load characteristics.
[0010] Optionally, before obtaining at least one target external factor data affecting the energy consumption of the device within the time period to be predicted, the method further includes:
[0011] Acquire first historical energy consumption data of the key equipment, load characteristic data corresponding to the first historical energy consumption data, and a plurality of historical external factor data;
[0012] Inputting the first historical energy consumption data into a time cycle model to obtain an energy consumption pattern output by the time cycle model;
[0013] The general analysis model is trained according to the energy consumption pattern, the load characteristic data and a plurality of historical external factor data to obtain the trained factor analysis model, which is used to indicate the correlation between equipment energy consumption and external factors.
[0014] Optionally, before inputting the first historical energy consumption data into a time cycle model to obtain an energy consumption pattern output by the time cycle model, the method further includes:
[0015] Screening the first historical energy consumption data to obtain second historical energy consumption data;
[0016] Analyzing and processing the second historical energy consumption data to obtain seasonal characteristics and nonlinear characteristics of the second historical energy consumption data;
[0017] Determine the range of the number of neurons, the range of the number of layers, and the time window of the initial time cycle model according to the seasonal characteristics and the nonlinear characteristics, wherein the time window is the length of time covering a complete seasonal cycle;
[0018] Determine a plurality of network configurations according to the range of the number of neurons and the range of the number of layers, wherein one number of neurons and one number of layers constitute one network configuration;
[0019] Performing training and verification processing on the multiple network configurations to obtain a target network configuration with the smallest verification error;
[0020] The initial time cycle model is trained and optimized according to the time window and the target network configuration to determine the time cycle model.
[0021] Optionally, the filtering and processing the first historical energy consumption data to obtain the second historical energy consumption data includes:
[0022] Calculating and processing the first historical energy consumption data, and determining a standard score for each first historical energy consumption data;
[0023] Determining whether the absolute value of the standard score is greater than a preset value;
[0024] If the absolute value of the standard score is greater than a preset value, the first historical energy consumption data corresponding to the standard score is determined to be an abnormal point, and the abnormal point is eliminated;
[0025] If the absolute value of the standard score is not greater than a preset value, the first historical energy consumption data corresponding to the standard score is determined to be the second historical energy consumption data.
[0026] Optionally, the general analysis model is trained according to the energy consumption pattern, the load characteristic data and a plurality of historical external factor data to obtain the trained factor analysis model, including:
[0027] Initializing the general analysis model to obtain a configured general analysis model;
[0028] Normalizing the load characteristic data and the plurality of historical external factor data to obtain characteristic data;
[0029] The configured general analysis model is trained according to the energy consumption pattern and the characteristic data to obtain a trained factor analysis model.
[0030] Optionally, determining the prediction feature vector corresponding to the time period to be predicted according to the at least one target external factor data and the equipment usage mode of the key equipment includes:
[0031] preprocessing the at least one target external factor data to obtain processed target external factor data;
[0032] A prediction feature vector corresponding to a time period to be predicted is determined according to the processed target external factor data and the device usage pattern.
[0033] Optionally, the method further includes:
[0034] Use feedback collection system to monitor and obtain actual energy consumption data in real time;
[0035] Determine the energy deviation value based on the energy demand corresponding to the actual energy consumption data and the predicted load of the equipment;
[0036] According to the energy deviation value, the model parameters of the factor analysis model are updated, and the energy allocation strategy and consumption strategy are adjusted.
[0037] In a second aspect, the present application provides an energy consumption intelligent analysis device based on time series and load characteristics, comprising:
[0038] An acquisition module, used to acquire at least one target external factor data affecting the energy consumption of the device within the time period to be predicted, wherein the target external factor is determined according to the correlation between the energy consumption of the device and the external factor;
[0039] A determination module, configured to determine a prediction feature vector corresponding to a time period to be predicted based on the at least one target external factor data and a device usage pattern of a key device;
[0040] An input module is used to input the prediction feature vector into a factor analysis model to obtain the equipment prediction load for the prediction period and the energy demand corresponding to the equipment prediction load. The factor analysis model is determined based on the first historical energy consumption data and historical equipment load characteristics.
[0041] Optionally, the device further comprises: a training module;
[0042] The acquisition module is further used to acquire the first historical energy consumption data of the key equipment, the load characteristic data corresponding to the first historical energy consumption data, and a plurality of historical external factor data;
[0043] The input module is further used to input the first historical energy consumption data into the time cycle model to obtain the energy consumption pattern output by the time cycle model;
[0044] The training module is used to train the general analysis model according to the energy consumption pattern, the load characteristic data and multiple historical external factor data to obtain the trained factor analysis model, and the factor analysis model is used to indicate the correlation between equipment energy consumption and external factors.
[0045] Optionally, the device further comprises: a processing module;
[0046] The processing module is used to filter and process the first historical energy consumption data to obtain second historical energy consumption data;
[0047] The processing module is further used to analyze and process the second historical energy consumption data to obtain seasonal characteristics and nonlinear characteristics of the second historical energy consumption data;
[0048] The determination module is further used to determine the range of the number of neurons, the range of the number of layers and the time window of the initial time cycle model according to the seasonal characteristics and the nonlinear characteristics, wherein the time window is the length of time covering a complete seasonal cycle;
[0049] The determination module is further used to determine a plurality of network configurations according to the range of the number of neurons and the range of the number of layers, wherein one number of neurons and one number of layers constitute one network configuration;
[0050] The processing module is further used to perform training and verification processing on the multiple network configurations to obtain a target network configuration with the minimum verification error;
[0051] The determination module is further used to train and optimize the initial time cycle model according to the time window and the target network configuration to determine the time cycle model.
[0052] Optionally, the device further includes: a judgment module;
[0053] The processing module is further used to perform calculation processing on the first historical energy consumption data and determine a standard score for each first historical energy consumption data;
[0054] The judging module is used to judge whether the absolute value of the standard score is greater than a preset value;
[0055] The determination module is further configured to determine that the first historical energy consumption data corresponding to the standard score is an abnormal point if the absolute value of the standard score is greater than a preset value, and remove the abnormal point;
[0056] The determination module is further configured to determine that the first historical energy consumption data corresponding to the standard score is the second historical energy consumption data if the absolute value of the standard score is not greater than a preset value.
[0057] Optionally, the processing module is further used to initialize the general analysis model to obtain a configured general analysis model;
[0058] The processing module is further used to perform normalization processing on the load characteristic data and the plurality of historical external factor data to obtain characteristic data;
[0059] The processing module is further used to train the configured general analysis model according to the energy consumption pattern and the characteristic data to obtain a trained factor analysis model.
[0060] Optionally, the processing module is further used to pre-process the at least one target external factor data to obtain processed target external factor data;
[0061] The determination module is further used to determine a prediction feature vector corresponding to a time period to be predicted based on the processed target external factor data and the device usage pattern.
[0062] Optionally, the acquisition module is further used to monitor and acquire actual energy consumption data in real time using a feedback collection system;
[0063] The determination module is further used to determine the energy deviation value according to the actual energy consumption data and the energy demand corresponding to the predicted load of the equipment;
[0064] The processing module is further used to update the model parameters of the factor analysis model according to the energy deviation value, and adjust the energy allocation strategy and consumption strategy.
[0065] In a third aspect, the present application provides an energy consumption intelligent analysis device based on time series and load characteristics, the device comprising:
[0066] Memory;
[0067] processor;
[0068] Wherein, the memory stores computer-executable instructions;
[0069] The processor executes the computer-executable instructions stored in the memory to implement the intelligent analysis method for energy consumption based on time series and load characteristics as described in the first aspect and various possible implementation methods of the first aspect.
[0070] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the intelligent analysis method of energy consumption based on time series and load characteristics as described in the above-mentioned first aspect and various possible implementation methods of the first aspect.
[0071] The energy consumption intelligent analysis method and system based on time series and load characteristics provided by the present application obtains at least one target external factor data that affects the energy consumption of the equipment in the time period to be predicted. According to the at least one target external factor data and the equipment usage pattern of the key equipment, the prediction feature vector corresponding to the time period to be predicted is determined. The prediction feature vector is input into the factor analysis model to obtain the equipment prediction load of the time period to be predicted and the energy demand corresponding to the equipment prediction load. The method can integrate real-time data streams and respond to external changes in a timely manner, and use external influencing factors and equipment usage patterns to jointly predict future energy demand, thereby improving the accuracy of the prediction. BRIEF DESCRIPTION OF THE DRAWINGS
[0072] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.
[0073] Figure 1 A schematic diagram of the framework of the energy consumption intelligent analysis system provided in the embodiment of the present application;
[0074] Figure 2 Schematic diagram of the process of the energy consumption intelligent analysis method based on time series and load characteristics provided in the embodiment of the present application Figure 1 ;
[0075] Figure 3 Schematic diagram of the process of the energy consumption intelligent analysis method based on time series and load characteristics provided in the embodiment of the present application Figure 2 ;
[0076] Figure 4 Schematic diagram of the process of the energy consumption intelligent analysis method based on time series and load characteristics provided in the embodiment of the present application Figure 3 ;
[0077] Figure 5 Schematic diagram of the process of the energy consumption intelligent analysis method based on time series and load characteristics provided in the embodiment of the present application Figure 4 ;
[0078] Figure 6 A schematic diagram of the structure of the energy consumption intelligent analysis device based on time series and load characteristics provided in this application;
[0079] Figure 7 A schematic diagram of the structure of the energy consumption intelligent analysis device based on time series and load characteristics provided in this application.
[0080] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION
[0081] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation methods described in the following exemplary embodiments do not represent all implementation methods consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims, rather than all embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0082] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein, for example. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, system, product or device 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 units that are not clearly listed or inherent to these processes, products or devices.
[0083] In the embodiments of the present application, the words "exemplary" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.
[0084] Energy management is an important area in modern industrial and commercial operations, especially in the context of rising energy costs and increasingly stringent environmental regulations. Energy management refers to the planning, control and optimization of energy in a scientific and rational way to improve energy efficiency, reduce energy consumption and reduce environmental impact.
[0085] Energy management can optimize the energy use process and adopt efficient equipment and technologies, which can significantly improve energy utilization efficiency, reduce energy waste, and thus reduce energy costs. Energy management can also reduce energy consumption. By diversifying the energy structure and improving energy self-sufficiency and reserve capacity, the stability and security of energy supply can be enhanced and dependence on external energy can be reduced.
[0086] Traditional energy management systems usually rely on basic monitoring and control technologies, which can only provide limited data analysis functions and have difficulty in handling complex data relationships or real-time changing energy demands. In addition, traditional methods have obvious deficiencies in data integration, real-time data processing, and dynamic energy optimization, which limits the improvement of energy efficiency and further reduction of costs. The existing technology lacks effective tools to analyze and utilize the complex relationship between equipment load characteristics and various external factors. Traditional energy management systems are usually unable to process large amounts of data streams in real time and have difficulty responding quickly to sudden changes in the external environment. Existing energy management solutions usually use a fixed model, lack flexibility, and cannot automatically adjust when faced with changes in data characteristics. Traditional methods adopt static or manual decision-making methods in energy allocation and consumption strategy formulation, which are inefficient and not accurate enough.
[0087] In response to the above problems, this application proposes an intelligent analysis method for energy consumption based on time series and load characteristics. This method can effectively identify peak and valley periods of energy consumption by real-time monitoring and forecasting of energy demand. Using this information, it can optimize energy distribution and adjust equipment operation plans, and enhance the adaptability and responsiveness of energy management. This application provides an intelligent analysis method and system for energy consumption based on time series and load characteristics with real-time data processing and response, adaptive forecasting and model optimization, and intelligent energy distribution strategy. This method can accurately capture and analyze these complex data interactions, realize intelligent analysis of energy consumption, and optimize energy use.
[0088] Figure 1 This is a schematic diagram of the framework of the energy consumption intelligent analysis system provided in the embodiment of the present application. Figure 1 As shown, the system includes: a data collection module, a data processing and analysis module, a model optimization and adaptive adjustment module, a real-time monitoring and response module, and an energy strategy optimization module.
[0089] The data collection module includes: Multi-function energy meter interface: responsible for real-time monitoring and recording of energy usage of various key equipment, and ensuring data accuracy and time consistency; Sensor network: deployed on key equipment to collect equipment operating status data, including start / stop time, operating frequency, temperature, pressure, etc.; Data transmission system: synchronizes the collected data to the central data processing center through a secure network connection.
[0090] The data processing and analysis modules include: pre-processing sub-module: cleaning, normalizing and outlier detection of collected data, using methods such as Z-Score to process abnormal data points; time series analysis tools: using techniques such as autoregression and seasonal decomposition to identify basic patterns and cyclical changes in energy consumption; factor analysis machine model: establishing a prediction model based on the complex relationship between equipment load characteristics and external factors, capturing and analyzing the interaction between features.
[0091] The model optimization and adaptive adjustment modules include: Dynamic network adjustment tool: dynamically adjust the number of layers and neurons of the LSTM network according to the seasonal and nonlinear characteristics of the data; Parameter optimization algorithm: use grid search and cross-validation methods to optimize model parameters to improve prediction accuracy and model generalization ability.
[0092] The real-time monitoring and response module includes: data stream processing platform: integrating real-time data stream processing technology, receiving and analyzing new data in real time, and quickly responding to external changes and emergencies; intelligent scheduling system: based on model prediction results, dynamically adjusting energy distribution and equipment operation strategies, such as adjusting loads or switching to backup energy systems during predicted high energy consumption periods.
[0093] The energy strategy optimization module includes: intelligent energy allocation strategy: optimize energy allocation according to forecast results, implement differential pricing strategies and demand response measures to reduce energy costs and improve efficiency; strategy adjustment feedback mechanism: continuously adjust and optimize forecast models and energy use strategies based on actual energy consumption and user feedback.
[0094] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.
[0095] Figure 2 Schematic diagram of the process of the energy consumption intelligent analysis method based on time series and load characteristics provided in the embodiment of the present application Figure 1 The execution subject of this embodiment is the energy consumption intelligent analysis system. Figure 2 As shown, the method includes:
[0096] S101: Acquire at least one target external factor data affecting device energy consumption within a time period to be predicted, wherein the target external factor is determined based on a correlation between device energy consumption and the external factor.
[0097] The target external factor data refers to the external factor data that affects the energy consumption of the equipment. For example, the external factor data may be weather forecast data for the time period to be predicted, or working day / non-working day information for the time period to be predicted.
[0098] It is understandable that external factors will have different effects on equipment energy consumption. For example, on weekdays, energy demand is higher, equipment operation time is longer than non-working days, and equipment energy consumption is also higher. Therefore, external factors are also an important consideration when predicting energy demand. After obtaining the external factor data, the data is first processed to make it suitable for model input. For numerical data, it is normalized to unify its range to [0, 1] or [-1, 1]. For categorical data, it is converted to a numerical value through unique hot encoding. For example, numerical data includes: temperature and humidity, categorical data includes time representation such as hours, days, months, working days, non-working days and holidays.
[0099] S102: Determine a prediction feature vector corresponding to a time period to be predicted according to the at least one target external factor data and a device usage pattern of a key device.
[0100] Optionally, the at least one target external factor data is preprocessed to obtain processed target external factor data, and a prediction feature vector corresponding to the time period to be predicted is determined based on the processed target external factor data and the device usage pattern.
[0101] It can be understood that based on the collected external factor data and the known equipment usage patterns of key equipment, feature vectors of future time points are generated, which will be used as inputs to the factor analysis model to predict future energy demand and equipment load. The future time point is also the period to be predicted.
[0102] S103: Input the prediction feature vector into a factor analysis model to obtain the predicted equipment load for the prediction period and the energy demand corresponding to the predicted equipment load, wherein the factor analysis model is determined based on the first historical energy consumption data and historical equipment load characteristics.
[0103] It is understandable that the factor analysis model is used to predict the feature vector at future moments and output the predicted energy demand. The factor analysis model will comprehensively consider the influence of each feature and its interaction effect, predict the equipment load at different time points, analyze the prediction results, pay special attention to the time period with abnormally high predicted energy consumption, and the influence of important external factors. Applying the prediction results to the energy management system can optimize energy distribution and equipment operation strategies.
[0104] Optionally, after obtaining the predicted load of the equipment in the time period to be predicted and the energy demand corresponding to the predicted load of the equipment, the energy allocation and consumption strategy can be optimized according to the prediction result to improve the energy utilization efficiency.
[0105] Analyze and process the prediction results to identify the high energy consumption periods and the key factors affecting the peak energy consumption. Analyze the model prediction results, pay attention to the time periods predicted to have abnormally high energy consumption, whether they are caused by specific external conditions (extreme weather) or specific times (peak hours on weekdays), and determine the main external factors affecting the peak energy consumption. For example, external factors can be changes in temperature, humidity or special events.
[0106] Optimize energy allocation based on the results of the analysis. For periods of high energy consumption, adjust the operating time of high-energy-consuming equipment, such as adjusting some non-urgent tasks or operations to off-peak periods. If the peak energy consumption forecast is accurate, reduce energy consumption during these periods through demand response strategies, such as adjusting temperature control settings, limiting certain operations, or starting backup energy systems.
[0107] At the same time, we optimize the strategy of energy consumption and implement time-differentiated pricing strategies based on the forecast results of energy demand to encourage users to use more energy during off-peak hours and reduce energy use during peak hours. Based on energy demand forecasts, we make energy purchase decisions in advance and use forecast information to select the best purchase time and contract to reduce energy costs.
[0108] The intelligent analysis method of energy consumption based on time series and load characteristics provided in this embodiment obtains at least one target external factor data that affects the energy consumption of the equipment in the time period to be predicted, wherein the target external factor is determined based on the correlation between the energy consumption of the equipment and the external factor. According to at least one target external factor data and the equipment usage pattern of the key equipment, the prediction feature vector corresponding to the time period to be predicted is determined. The prediction feature vector is input into the factor analysis model to obtain the equipment prediction load of the time period to be predicted and the energy demand corresponding to the equipment prediction load. The factor analysis model is determined based on the first historical energy consumption data and the historical equipment load characteristics. This method can integrate real-time data streams and respond to external changes in a timely manner, and use external influencing factors and equipment usage patterns to jointly predict future energy demand, thereby improving the accuracy of the prediction.
[0109] Figure 3 Schematic diagram of the process of the energy consumption intelligent analysis method based on time series and load characteristics provided in the embodiment of the present application Figure 2 This embodiment is in Figure 2 Based on the embodiment, a detailed description of the time cycle model training process in the energy consumption intelligent analysis method based on time series and load characteristics is provided. Figure 3 As shown, the method includes:
[0110] S201: Acquire first historical energy consumption data of the key equipment, load characteristic data corresponding to the first historical energy consumption data, and multiple historical external factor data.
[0111] It is understandable that key equipment is integrated with multifunctional energy meters to continuously record the energy usage of the equipment, including electricity, water and gas. Among them, the energy meter has data recording and time stamping functions to ensure the accuracy of data and time consistency. The equipment operation status data is collected through sensors installed on the equipment, and the equipment operation status data includes the start / stop time of the equipment, operating frequency, temperature, pressure and other operating parameters.
[0112] The real-time data collection system automatically collects data from energy meters and sensors according to the data collection time interval, and transmits the collected data as a data set to the central data processing center through a secure network connection. The real-time data collection system runs synchronously with the energy meters and sensors, and sets the data collection time interval to ensure that the data is updated according to the preset time period (every hour, every day), so as to meet the time series requirements of the analysis. At the same time, in order to ensure the high reliability and traceability of the data set, encryption and security measures are used to protect the security and integrity of the data during transmission.
[0113] S202: Calculate and process the first historical energy consumption data, and determine a standard score for each first historical energy consumption data.
[0114] It is understandable that before analyzing and processing the data set, statistical outlier detection is introduced to identify and process anomalies in energy consumption data. In addition to time features, environmental and socioeconomic variables are added as input features, including holidays, weekdays and weekends. In the processing of energy consumption data, it is crucial to process and identify outliers because they represent data collection errors, equipment failures or unconventional usage patterns, which will distort the results of subsequent analysis.
[0115] Outliers can be effectively identified using statistical outlier detection methods, which are also known as Z-Score methods, and are applicable to normally distributed data sets. Z-Score is a method for measuring the degree of deviation of a single data point from the mean. In this method, the Z-Score value of each point represents the number of standard deviations from the mean of the data set. The calculation formula for Z-Score is:
[0116]
[0117] in, X is the observed value, is the sample mean, is the sample standard deviation. According to the above formula, the mean and standard deviation of the data set are calculated. For each observation in the data set, , and get its standard score Z-Score value.
[0118] S203: Determine whether the absolute value of the standard score is greater than a preset value, if so, execute step S204, if not, execute step S205.
[0119] Understandably, outliers are identified based on the standard score and a preset value is selected , the preset value is compared with the standard score to filter out the outliers in the data set. For example, the preset value can be set to 2.5, 3 or 3.5. The setting of the preset value depends on the strictness of the desired outlier identification.
[0120] S204: If the absolute value of the standard score is greater than a preset value, determining that the first historical energy consumption data corresponding to the standard score is an abnormal point, and eliminating the abnormal point.
[0121] Understandably, if ,but It is considered to be an outlier, or abnormal point, which will affect the training of the prediction model, so the data value is directly removed.
[0122] S205: If the absolute value of the standard score is not greater than a preset value, determining that the first historical energy consumption data corresponding to the standard score is the second historical energy consumption data.
[0123] Understandably, if ,but It is considered to be a normal point, and the data value is used as the second historical energy consumption data. After all the second historical energy consumption data are determined, the second historical energy consumption data are analyzed and processed, and the model is trained using the characteristics of the second historical energy consumption data.
[0124] S206: Analyze and process the second historical energy consumption data to obtain seasonal characteristics and nonlinear characteristics of the second historical energy consumption data.
[0125] It can be understood that the seasonal characteristics of the data are analyzed using autocorrelation and partial autocorrelation functions (ACF and PACF). ACF shows an obvious periodic pattern, indicating that the data has strong seasonality. The nonlinearity of the data can be evaluated by visualizing the distribution of the data or using nonlinear tests. For example, the BDS test can be used to evaluate the nonlinearity of the data.
[0126] S207: Determine the range of the number of neurons, the range of the number of layers, and the time window of the initial time cycle model according to the seasonal characteristics and the nonlinear characteristics, wherein the time window is the length of time covering a complete seasonal cycle.
[0127] Understandably, based on the seasonality and nonlinearity of the data, a range can be initially set. If the data shows strong seasonality, more layers are needed to capture these patterns; if the data is highly nonlinear, increasing the number of neurons in each layer helps the model capture more complex relationships. The time cycle model is also a model composed of a long short-term memory network, referred to as the LSTM model.
[0128] The selection of the time window is based on seasonal characteristics. The time window should cover at least one complete seasonal cycle. If the data has daily seasonality, the window is set to 24 hours. Based on the autocorrelation function, observe the ACF plot to determine the farthest time step that affects current consumption. If the ACF is still significant after 48 hours, this indicates that at least a 48-hour window is needed. Determine the significant seasonal lengths in the data, such as 24, 48 or more hours, and use these as candidate windows for experiments. Train the model for each window length and evaluate its performance through cross-validation. Select the window length that performs better than others.
[0129] S208: Determine a plurality of network configurations according to the range of the number of neurons and the range of the number of layers, wherein one number of neurons and one number of layers constitute one network configuration.
[0130] It is understandable that a performance-based approach is used to test different network configurations. For example, a grid search method can be used to define a range of the number of neurons and a range of the number of layers, and then train and validate each configuration to select the model with the smallest validation error. For example, the number of neurons can range from 50 to 200, and the number of layers can range from 1 to 3, where one network configuration can be 50 neurons and 1 layer.
[0131] S209: Perform training and verification processing on the multiple network configurations to obtain a target network configuration with the minimum verification error.
[0132] It is understandable that the number of layers and neurons of the LSTM network are dynamically adjusted according to the seasonal and nonlinear characteristics of the data. For example, a deeper network structure is used during the peak energy consumption period in winter and summer to improve the learning and prediction capabilities of the model. Based on the adaptive selection, the final number of LSTM layers and neurons is determined. Train and verify multiple network configurations to determine the verification error. After multiple configurations, the network configuration corresponding to the model with the smallest verification error is selected as the target network configuration.
[0133] Optionally, a grid search algorithm is used to select the model parameters with the best output performance:
[0134] Define the parameters to be tested and their possible values. For the LSTM model, you can test different numbers of layers and the number of neurons in each layer. List of parameters to be tested: Example of number of layers: 1, 2, 3; Example of number of neurons: 50, 100, 150, 200. Create a model instance for each parameter combination. For each combination of number of layers and number of neurons, build a corresponding LSTM network. Set other parameters of the model, such as activation function, optimizer, etc. Train and evaluate the model for each parameter combination, and use cross-validation to evaluate the performance of each combination to ensure the stability and reliability of the results. Based on the evaluation results, for example, the evaluation result can be the minimum mean square error, and the parameter combination with excellent performance is selected. Record and report the configuration and performance indicators of the best model. Retrain the model on the entire dataset using the selected best parameter combination. Verify the final performance of the model to ensure that the model performs similarly on future data as during cross-validation.
[0135] Cross-validation specifically means: dividing the data set into multiple parts (can be divided into 3 parts), using one part as the test set and the others as the training set in turn, so as to perform multiple tests on each combination to ensure the accuracy of the evaluation.
[0136] S209: Training and optimizing the initial time cycle model according to the time window and the target network configuration to determine the time cycle model.
[0137] It can be understood that after determining the time window and network configuration, the length of the input sequence is defined, that is, the size of the historical data time window that the model should consider when making predictions, matching the periodic characteristics of energy consumption, setting the dropout rate to prevent overfitting, selecting ReLU or tanh as the activation function, using the RMSprop optimizer, and using the mean square error (MSE) as the loss function, thus forming a preliminary LSTM model, and using the time series cross-validation method to evaluate the performance of the LSTM model to simulate various situations in actual operation, and dynamically adjust the learning rate according to the performance of the model during the training process. Combine the output of LSTM with the application of statistical methods to enhance the smoothing of energy consumption trends and the accuracy of prediction.
[0138] Applied statistical methods include: Moving average: Calculate the moving average of LSTM output and select the average of the most recent N time steps as the final output, which is suitable for eliminating fluctuations within a cycle; Exponential smoothing: If the data has more complex trends and seasonality, use exponential smoothing to give higher weights to recent data and adapt more flexibly to trend changes; Parameter optimization: Adjust the parameters of the statistical method (such as the smoothing coefficient) based on historical data to ensure the best smoothing effect, integrate the LSTM model into the real-time data stream processing platform, and respond quickly to emergencies. The data stream processing platform includes Apache Kafka or Apache Storm, which enables the model to instantly receive and analyze the latest data and respond to emergencies or significant consumption changes in a timely manner.
[0139] Optionally, to ensure accurate data transmission, set up a data stream interface and use the data stream management system (Apache Kafka) as the middle layer for data transmission to collect and distribute energy consumption data in real time. Configure the data stream to access the model service to ensure that the data can be delivered to the model processing unit in real time.
[0140] Optionally, deploy the trained LSTM model as a service, using TensorFlow Serving to create an API service that can receive real-time data and return predictions. Ensure that the model can quickly load and process real-time data, taking into account latency and processing power requirements. Process input data in real time, including standardization and format conversion to fit the model's input requirements, and implement necessary outlier detection and correction to ensure input data quality. Establish a feedback loop between the model output and the energy management system, adjust energy usage strategies based on model predictions, and adjust and optimize the model itself based on actual energy usage.
[0141] The intelligent analysis method of energy consumption based on time series and load characteristics provided in this embodiment obtains the seasonal characteristics and nonlinear characteristics of the second historical energy consumption data by screening and analyzing the first historical energy consumption data. According to the seasonal characteristics and nonlinear characteristics, the range of the number of neurons, the range of the number of layers and the time window of the initial time cycle model are determined. According to the range of the number of neurons and the range of the number of layers, multiple network configurations are determined. The multiple network configurations are trained and verified to obtain the target network configuration with the minimum verification error. The initial time cycle model is trained and optimized according to the time window and the target network configuration to determine the time cycle model. The model can instantly receive and analyze the latest data and respond to emergencies or significant changes in consumption in a timely manner.
[0142] Figure 4 Schematic diagram of the process of the energy consumption intelligent analysis method based on time series and load characteristics provided in the embodiment of the present application Figure 3 This embodiment is in Figure 2Based on the embodiment, a detailed description of the factor analysis model training process in the energy consumption intelligent analysis method based on time series and load characteristics is provided. Figure 4 As shown, the method includes:
[0143] S301: Acquire first historical energy consumption data of the key equipment, load characteristic data corresponding to the first historical energy consumption data, and multiple historical external factor data.
[0144] Among them, step S301 is similar to step S201 and will not be repeated here.
[0145] S302: Input the first historical energy consumption data into a time cycle model to obtain an energy consumption pattern output by the time cycle model.
[0146] It is understandable that the historical energy consumption data is input into the trained time cycle model. The time cycle model is set with an early warning system. When the model predicts an abnormal consumption pattern, it can immediately notify relevant personnel or automatically take measures. The time cycle model can not only more accurately identify and predict energy consumption patterns, but also effectively respond to unconventional consumption behaviors and extreme events. In the actual application process, when abnormal energy consumption is detected, the abnormal energy consumption pattern is input into the general analysis model to obtain the external influencing factors of abnormal energy consumption, so as to facilitate the relevant personnel to adjust the consumption strategy.
[0147] S303: Initialize the general analysis model to obtain a configured general analysis model.
[0148] As you can understand, first configure the general analysis model, initialize the factor analysis machine model, and configure the core parameters, including the dimension of the factor, which determines the complexity of the interaction term. The factor dimension is generally set to a lower dimension, such as 10~50. At the same time, set hyperparameters such as learning rate and number of iterations, use an appropriate loss function, and complete the configuration of the general analysis model.
[0149] S304: normalizing the load characteristic data and the plurality of historical external factor data to obtain characteristic data.
[0150] It can be understood that the load characteristic data and historical external factor data are divided into two categories: numerical data and categorical data. They are processed separately according to the data category. The numerical data is normalized and the categorical data is converted into a one-hot encoding. The processed data set forms feature data. The classification method and processing method are similar to step S101 and will not be repeated here.
[0151] S305: Training the configured general analysis model according to the energy consumption pattern and the characteristic data to obtain a trained factor analysis model, where the factor analysis model is used to indicate the correlation between device energy consumption and external factors.
[0152] It is understandable that the energy consumption pattern identified by time series analysis and feature data output by the time cycle model are used as feature sets to train the configured general analysis model. During the training process, the model will learn not only the impact of a single feature on energy consumption, but also the impact of interactions between features. Cross-validation techniques are used to evaluate the performance of the model and prevent overfitting. The factor analysis machine can handle the interaction between features, use model parameters and outputs to analyze which factors and their interactions have a significant impact on energy consumption data, and use the trained factor analysis model to predict energy consumption and analyze the impact of external factors, providing decision support for energy management and optimization.
[0153] The trained factor analysis model includes the correlation analysis results between equipment energy consumption and external factors, and the dot product of the factor vector is used to capture the interaction between features. The model parameters (including bias terms, weights of single features, and weights of interactions between features, which have been trained using historical data) can significantly represent the contribution of each factor to energy consumption.
[0154] Optionally, the factor analysis machine is a machine learning algorithm that effectively processes high-dimensional sparse data sets and captures all interactions between variables. The core idea is to use factorization technology to approximate the interaction weights between each pair of features, making the factor analysis machine particularly suitable for scenarios containing a large number of category features. In the correlation analysis between equipment load characteristics and external factors in the embodiment of the present application, the prediction model of the factor analysis machine is expressed as the following formula:
[0155]
[0156] in, is the predicted output of the model, x is a eigenvector, where is the value of the ith feature, is the global bias, is the weight of the ith feature, and is the factor vector of the i-th and j-th features, with dimension k, indicating the number of latent factors, yes and The dot product of is used to model the interaction between feature i and feature j.
[0157] Bias and linear terms: and It is a component of the traditional linear regression model and is responsible for capturing the independent contribution of features;
[0158] Interaction terms: It is the core of FM and can simulate the interaction effect between any two features. Each feature i and j is represented by its own factor vector and To represent the dot product of these factor vectors Provides a numerical value that indicates how two features jointly affect the target variable.
[0159] Compared to the model that needs to learn a weight for each pair of feature combinations, the factor analysis machine uses a low-rank matrix v The method greatly reduces the number of parameters of the model while maintaining the ability to capture complex interactions. The factor analysis machine can handle any real-valued features and naturally expands to multidimensional and multimodal data. When applying the factor analysis machine to model the relationship between equipment load characteristics and external factors, the focus is on selecting appropriate factor dimensions. k and a suitable optimization algorithm (stochastic gradient descent) to train the model. In addition, the best hyperparameters ( k , learning rate) are crucial to building accurate and reliable prediction models.
[0160] The intelligent analysis method for energy consumption based on time series and load characteristics provided in this embodiment obtains the first historical energy consumption data of the key equipment, the load characteristic data corresponding to the first historical energy consumption data, and multiple historical external factor data. The first historical energy consumption data is input into the time cycle model to obtain the energy consumption pattern output by the time cycle model. The general analysis model is initialized to obtain a configured general analysis model. The load characteristic data and the multiple historical external factor data are normalized to obtain characteristic data. The configured general analysis model is trained according to the energy consumption pattern and the characteristic data to obtain a trained factor analysis model. The model effectively handles the complex interactions between equipment energy consumption and multiple external factors, allowing adaptation to different types of data inputs and changing environmental conditions, thereby providing continuous and dynamic energy management optimization support.
[0161] Figure 5 Schematic diagram of the process of the energy consumption intelligent analysis method based on time series and load characteristics provided in the embodiment of the present application Figure 4 This embodiment is in Figure 2 Based on the embodiment, a detailed process description of the implementation of the feedback collection system in the energy consumption intelligent analysis method based on time series and load characteristics is provided. Figure 5 As shown, the method includes:
[0162] S401: Use a feedback collection system to monitor and obtain actual energy consumption data in real time.
[0163] It is understandable that through the feedback mechanism, the prediction model and optimization strategy can be adjusted according to the actual energy consumption to achieve continuous energy management optimization. First, establish a feedback collection system, implement automation systems and sensor technology to continuously monitor and record actual energy consumption data, including equipment usage, peak energy consumption hours, and total energy consumption, and regularly collect user feedback on energy use, including any reports on equipment performance or energy shortages.
[0164] S402: Determine an energy deviation value according to the actual energy consumption data and the energy demand corresponding to the predicted load of the equipment.
[0165] It can be understood that the performance of the prediction model can be evaluated through actual energy consumption data, the prediction results can be compared with the actual energy consumption data regularly, the prediction accuracy and deviation can be analyzed, and common error types and patterns in model predictions can be identified, including systematic deviations or random errors, and the association of these errors with specific external conditions or operating modes.
[0166] S403: According to the energy deviation value, the model parameters of the factor analysis model are updated, and the energy allocation strategy and consumption strategy are adjusted.
[0167] It is understandable that the model parameters, including the weights and factor vectors in the factor analysis machine, are adjusted based on the error analysis results (energy deviation values), and the model is adjusted and optimized to better reflect the actual energy usage. Based on the results of the model adjustment and the new energy usage data, the existing energy allocation strategy and consumption strategy are re-evaluated, and the energy management strategy is dynamically adjusted based on the latest model predictions and market or environmental changes, including modifying the peak operation strategy or adjusting the energy purchase plan.
[0168] The energy consumption intelligent analysis method based on time series and load characteristics provided in this embodiment monitors and obtains actual energy consumption data in real time by adopting a feedback collection system. The energy deviation value is determined based on the energy demand corresponding to the actual energy consumption data and the predicted load of the equipment. According to the energy deviation value, the model parameters of the factor analysis model are updated, and the energy allocation strategy and consumption strategy are adjusted. The implementation of this method enables the energy management system to continuously learn and adapt, and to cope with changes in environmental and operating conditions by continuously adjusting the prediction model and optimization strategy, thereby maintaining long-term energy management optimization.
[0169] Figure 6 This is a schematic diagram of the structure of the energy consumption intelligent analysis device based on time series and load characteristics provided in this application. Figure 6As shown, the energy consumption intelligent analysis device 500 based on time series and load characteristics provided by the present application includes:
[0170] An acquisition module 501 is used to acquire at least one target external factor data affecting the energy consumption of the device within the period to be predicted, wherein the target external factor is determined according to the correlation between the energy consumption of the device and the external factor;
[0171] A determination module 502 is used to determine a prediction feature vector corresponding to a time period to be predicted according to the at least one target external factor data and a device usage pattern of a key device;
[0172] The input module 503 is used to input the prediction feature vector into the factor analysis model to obtain the equipment prediction load for the prediction period and the energy demand corresponding to the equipment prediction load. The factor analysis model is determined based on the first historical energy consumption data and historical equipment load characteristics.
[0173] Optionally, the device further includes: a training module 504;
[0174] The acquisition module 501 is further used to acquire the first historical energy consumption data of the key equipment, the load characteristic data corresponding to the first historical energy consumption data, and a plurality of historical external factor data;
[0175] The input module 503 is further used to input the first historical energy consumption data into the time cycle model to obtain the energy consumption pattern output by the time cycle model;
[0176] The training module 504 is used to train the general analysis model according to the energy consumption pattern, the load characteristic data and multiple historical external factor data to obtain the trained factor analysis model, and the factor analysis model is used to indicate the correlation between equipment energy consumption and external factors.
[0177] Optionally, the device further includes: a processing module 505;
[0178] The processing module 505 is used to filter and process the first historical energy consumption data to obtain second historical energy consumption data;
[0179] The processing module 505 is further used to analyze and process the second historical energy consumption data to obtain seasonal characteristics and nonlinear characteristics of the second historical energy consumption data;
[0180] The determination module 502 is further used to determine the range of the number of neurons, the range of the number of layers and the time window of the initial time cycle model according to the seasonal characteristics and the nonlinear characteristics, wherein the time window is the length of time covering a complete seasonal cycle;
[0181] The determination module 502 is further used to determine a plurality of network configurations according to the range of the number of neurons and the range of the number of layers, wherein one number of neurons and one number of layers constitute one network configuration;
[0182] The processing module 505 is further used to perform training and verification processing on the multiple network configurations to obtain a target network configuration with the minimum verification error;
[0183] The determination module 502 is further configured to train and optimize the initial time cycle model according to the time window and the target network configuration to determine the time cycle model.
[0184] Optionally, the device further includes: a judgment module 506;
[0185] The processing module 505 is further used to calculate and process the first historical energy consumption data and determine the standard score of each first historical energy consumption data;
[0186] The judging module 506 is used to judge whether the absolute value of the standard score is greater than a preset value;
[0187] The determination module 502 is further configured to determine that the first historical energy consumption data corresponding to the standard score is an abnormal point if the absolute value of the standard score is greater than a preset value, and remove the abnormal point;
[0188] The determination module 502 is further configured to determine that the first historical energy consumption data corresponding to the standard score is the second historical energy consumption data if the absolute value of the standard score is not greater than a preset value.
[0189] Optionally, the processing module 505 is further used to initialize the general analysis model to obtain a configured general analysis model;
[0190] The processing module 505 is further used to perform normalization processing on the load characteristic data and the plurality of historical external factor data to obtain characteristic data;
[0191] The processing module 505 is further used to train the configured general analysis model according to the energy consumption pattern and the characteristic data to obtain a trained factor analysis model.
[0192] Optionally, the processing module 505 is further used to pre-process the at least one target external factor data to obtain processed target external factor data;
[0193] The determination module 502 is further configured to determine a prediction feature vector corresponding to a time period to be predicted according to the processed target external factor data and the device usage pattern.
[0194] Optionally, the acquisition module 501 is further used to monitor and acquire actual energy consumption data in real time using a feedback collection system;
[0195] The determination module 502 is further used to determine the energy deviation value according to the actual energy consumption data and the energy demand corresponding to the predicted load of the equipment;
[0196] The processing module 505 is further used to update the model parameters of the factor analysis model according to the energy deviation value, and adjust the energy allocation strategy and consumption strategy.
[0197] Figure 7 This is a schematic diagram of the structure of the energy consumption intelligent analysis device based on time series and load characteristics provided in this application. Figure 7 As shown, the present application provides an energy consumption intelligent analysis device based on time series and load characteristics. The energy consumption intelligent analysis device based on time series and load characteristics 600 includes: a receiver 601, a transmitter 602, a processor 603 and a memory 604.
[0198] Receiver 601, used for receiving instructions and data;
[0199] A transmitter 602, used for sending instructions and data;
[0200] Memory 604, used to store computer-executable instructions;
[0201] The processor 603 is used to execute the computer-executable instructions stored in the memory 604 to implement the various steps performed by the energy consumption intelligent analysis method based on time series and load characteristics in the above embodiment. For details, please refer to the relevant description in the above embodiment of the energy consumption intelligent analysis method based on time series and load characteristics.
[0202] Optionally, the memory 604 may be independent or integrated with the processor 603 .
[0203] When the memory 604 is independently provided, the electronic device further includes a bus for connecting the memory 604 and the processor 603 .
[0204] The present application also provides a computer-readable storage medium, which stores computer execution instructions. When a processor executes the computer execution instructions, it implements the energy consumption intelligent analysis method based on time series and load characteristics as executed by the energy consumption intelligent analysis device based on time series and load characteristics as mentioned above.
[0205] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In hardware implementations, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or transient medium). As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.
[0206] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the following claims.
[0207] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.
Claims
1. An intelligent energy consumption analysis method based on time series and load characteristics, characterized in that: The method comprises: Acquire first historical energy consumption data of key equipment, load characteristic data corresponding to the first historical energy consumption data, and multiple historical external factor data; Inputting the first historical energy consumption data into a time cycle model to obtain an energy consumption pattern output by the time cycle model; According to the energy consumption pattern, the load characteristic data and a plurality of historical external factor data, a general analysis model is trained to obtain a trained factor analysis model, wherein the factor analysis model is used to indicate the correlation between the energy consumption of the equipment and the external factors; Acquire at least one target external factor data affecting the energy consumption of the device within the time period to be predicted, wherein the target external factor is determined according to the correlation between the energy consumption of the device and the external factor; Determining a prediction feature vector corresponding to a time period to be predicted according to the at least one target external factor data and a device usage pattern of the key device; The prediction feature vector is input into a factor analysis model to obtain the equipment prediction load in the prediction period and the energy demand corresponding to the equipment prediction load. The factor analysis model is determined based on the first historical energy consumption data and historical equipment load characteristics.
2. The method according to claim 1, characterized in that Before inputting the first historical energy consumption data into the time cycle model to obtain the energy consumption pattern output by the time cycle model, the method further includes: Screening the first historical energy consumption data to obtain second historical energy consumption data; Analyzing and processing the second historical energy consumption data to obtain seasonal characteristics and nonlinear characteristics of the second historical energy consumption data; Determine the range of the number of neurons, the range of the number of layers, and the time window of the initial time cycle model according to the seasonal characteristics and the nonlinear characteristics, wherein the time window is the length of time covering a complete seasonal cycle; Determine a plurality of network configurations according to the range of the number of neurons and the range of the number of layers, wherein one number of neurons and one number of layers constitute one network configuration; Performing training and verification processing on the multiple network configurations to obtain a target network configuration with the smallest verification error; The initial time cycle model is trained and optimized according to the time window and the target network configuration to determine the time cycle model.
3. The method according to claim 1, characterized in that The filtering and processing of the first historical energy consumption data to obtain the second historical energy consumption data includes: Calculating and processing the first historical energy consumption data, and determining a standard score for each first historical energy consumption data; Determining whether the absolute value of the standard score is greater than a preset value; If the absolute value of the standard score is greater than a preset value, the first historical energy consumption data corresponding to the standard score is determined to be an abnormal point, and the abnormal point is eliminated; If the absolute value of the standard score is not greater than a preset value, the first historical energy consumption data corresponding to the standard score is determined to be the second historical energy consumption data.
4. The method according to claim 1, characterized in that: The general analysis model is trained according to the energy consumption pattern, the load characteristic data and a plurality of historical external factor data to obtain the trained factor analysis model, including: Initializing the general analysis model to obtain a configured general analysis model; Normalizing the load characteristic data and the plurality of historical external factor data to obtain characteristic data; The configured general analysis model is trained according to the energy consumption pattern and the characteristic data to obtain a trained factor analysis model.
5. The method according to claim 1, characterized in that The step of determining a prediction feature vector corresponding to a time period to be predicted according to the at least one target external factor data and the equipment usage mode of the key equipment includes: preprocessing the at least one target external factor data to obtain processed target external factor data; A prediction feature vector corresponding to a time period to be predicted is determined according to the processed target external factor data and the device usage pattern.
6. The method according to claim 1, characterized in that The method further comprises: Use feedback collection system to monitor and obtain actual energy consumption data in real time; Determine the energy deviation value based on the energy demand corresponding to the actual energy consumption data and the predicted load of the equipment; According to the energy deviation value, the model parameters of the factor analysis model are updated, and the energy allocation strategy and consumption strategy are adjusted.
7. An intelligent energy consumption analysis device based on time series and load characteristics, characterized in that: The device comprises: An acquisition module, used to acquire first historical energy consumption data of key equipment, load characteristic data corresponding to the first historical energy consumption data, and multiple historical external factor data; An input module, used for inputting the first historical energy consumption data into a time cycle model to obtain an energy consumption pattern output by the time cycle model; A training module, used for training a general analysis model according to the energy consumption pattern, the load characteristic data and a plurality of historical external factor data, to obtain a trained factor analysis model, wherein the factor analysis model is used for indicating the correlation between the energy consumption of the equipment and the external factors; The acquisition module is further used to acquire at least one target external factor data affecting the energy consumption of the device within the time period to be predicted, wherein the target external factor is determined according to the correlation between the energy consumption of the device and the external factor; A determination module, configured to determine a prediction feature vector corresponding to a time period to be predicted based on the at least one target external factor data and a device usage pattern of a key device; The input module is also used to input the prediction feature vector into a factor analysis model to obtain the equipment prediction load for the prediction period and the energy demand corresponding to the equipment prediction load. The factor analysis model is determined based on the first historical energy consumption data and historical equipment load characteristics.
8. An intelligent energy consumption analysis device based on time series and load characteristics, characterized in that: include: Memory; processor; Wherein, the memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the energy consumption intelligent analysis method based on time series and load characteristics as described in any one of claims 1-6.
9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the energy consumption intelligent analysis method based on time series and load characteristics as described in any one of claims 1 to 6.
Citation Information
Patent Citations
Energy load prediction method, device and equipment
CN113947007A
Energy big data processing method and device, computer equipment and storage medium
CN114492998A
Regional energy network key factor extraction method based on factor analysis method
CN115526535A
Power consumption evaluation method and system for plc escalator
CN117370760A
Multi-element load prediction method for integrated energy system
CN117439045A
