Electricity consumption strategy recommendation method and system based on scientific energy consumption and settlement

Abstract

The present invention relates to the field of power management technology, and in particular to a method and system for recommending power consumption strategies based on scientific energy consumption and settlement. The method comprises the following steps: obtaining climate data; extracting the power grid operation status from the announcements, reports and APIs issued by the power trading center, thereby obtaining power grid operation status data; obtaining historical energy price data, and performing a time series-based periodic analysis on the historical energy price data to obtain primary energy price data; selecting and extracting key features based on changes in the power supply and demand relationship based on climate data, power grid operation status data and primary energy price data, thereby obtaining a power supply and demand feature set; using a gradient boosting tree algorithm to predict the power supply and demand feature set for the power supply and demand feature set, and obtaining intraday power spot supply and demand forecast data and its confidence interval. The present invention can quickly adjust strategies according to real-time market conditions through dynamic energy optimization strategy formulation.

Description

Technical Field

[0001] The present invention relates to the technical field of power management, and in particular to a method and system for recommending power consumption strategies based on scientific energy consumption and settlement. Background Art

[0002] The purpose of the power market reform is to establish a sound power market system to realize the marketization of power transactions and thus achieve the optimal allocation of power resources. With the continuous development of the power market reform, market players have put forward higher requirements for free competition and price openness; at the same time, on the power supply side, the grid-connected capacity of new energy sources with volatility and random output characteristics has continued to increase, which has increased the complexity of power grid dispatching and put forward higher requirements for power grid flexibility and reliability; on the power demand side, with the continuous penetration of distributed power sources, energy storage and other resources, load aggregators have continued to emerge, and the demand side of the power grid has gradually shown the characteristics of diversified users, load fluctuations and unpredictability. The above problems have prompted the power supply and demand business to develop in the direction of source-grid-load-storage multi-energy complementation and coordinated work at the macro level, and in the direction of intelligent interaction and scientific energy use at the micro level. In this process, power sales companies, as market players, participate in wholesale and retail transactions, which is an important part of realizing friendly interaction between the power supply side and the demand side. In the power spot market, the power purchase of power sales companies is not limited to the medium and long-term market, such as: annual trading market, monthly trading market, etc., but also includes the spot market. The spot price of electricity fluctuates due to the influence of supply and demand, which in turn affects the wholesale cost of power sales companies. However, the retail end of power sales companies (i.e., the price of electricity sold to users) is usually fixed. In order to control the risk of spot prices being higher than retail contract prices, power sales companies usually issue demand response strategies to achieve cost savings. From the perspective of power grid supply and demand balance, low spot prices represent loose supply and demand, and vice versa, tight supply and demand. How to mobilize demand-side resources through incentive signals of electricity spot prices, such as distributed power sources, electric vehicles, energy storage, flexible and adjustable loads, etc., give full play to the supporting role of various resources for the power grid, and achieve friendly interaction between the demand side and the power grid has become a technical difficulty. From the perspective of power sales company operations, how to design scientific energy optimization guidelines based on spot price signals and comprehensively evaluate the benefits of demand-side response are technical problems that need to be solved urgently. Summary of the invention

[0003] Based on this, it is necessary for the present invention to provide a method and system for recommending electricity consumption strategies based on scientific energy consumption and settlement, so as to solve at least one of the above technical problems.

[0004] To achieve the above purpose, a method for recommending electricity consumption strategy based on scientific energy consumption and settlement is provided, comprising the following steps:

[0005] Step S1: Acquire climate data; extract the grid operation status from the announcements, reports and APIs issued by the power trading center to obtain grid operation status data; obtain historical energy price data, and perform a periodic analysis of the historical energy price data based on time series to obtain primary energy price data;

[0006] Step S2: Select and extract key features based on the change of power supply and demand relationship according to climate data, power grid operation status data and primary energy price data, so as to obtain a power supply and demand feature set; use the gradient boosting tree algorithm to predict the power supply and demand feature set, and obtain the intraday power spot supply and demand forecast data and its confidence interval;

[0007] Step S3: Formulate a dynamic energy optimization strategy based on the electricity spot supply and demand forecast data and its confidence interval, thereby obtaining energy optimization strategy data, wherein the dynamic energy optimization strategy is specifically formulated by subdividing the electricity consumption period into a safe low-price zone, a possible low-price zone, a price fluctuation zone, and a high-price risk zone, and formulating differentiated electricity consumption recommendations for each zone;

[0008] Step S4: Real-time monitoring of the power supply and demand relationship of the energy optimization strategy data, and comparison of the actual electricity price with the power spot supply and demand forecast data and its confidence interval to obtain forecast deviation interval data; triggering an adjustment mechanism for the energy optimization strategy data according to the forecast deviation interval data, thereby obtaining energy guidance update data;

[0009] Step S5: Calculate the benefits of energy-consuming enterprises responding to energy consumption optimization guidelines based on the updated energy consumption guidelines data, and compare the estimated response benefits with the actual response benefits to obtain response benefit difference data; adjust the strategy optimization parameters based on the response benefit difference data to achieve adaptive optimization of the electricity consumption strategy.

[0010] The climate data of the present invention is crucial to the operation of the power system. Accurate climate data can help the system better predict load demand, optimize power generation plans, and improve energy scheduling. Extracting grid operation status data from information released by the power trading center can help the system monitor the real-time status of the power grid, promptly discover and solve potential problems, and improve the stability and reliability of the power grid; the analysis and prediction of historical energy price data can help enterprises make more informed energy procurement decisions and plan future electricity costs. By selecting and extracting key features of changes in power supply and demand, the spot supply and demand of electricity can be more accurately predicted, helping enterprises make more effective power decisions. The algorithm can effectively handle complex nonlinear relationships, predict power supply and demand feature sets, provide more accurate power spot supply and demand forecast data, and help enterprises optimize power plans. Through the dynamic energy optimization strategy formulated based on the spot supply and demand forecast data, enterprises can more flexibly adjust power plans to obtain the best power benefits. Real-time monitoring of power supply and demand relationships helps enterprises adjust power strategies in a timely manner to cope with emergencies and fluctuating market conditions. By comparing actual electricity prices with forecast data, forecast deviation interval data can help enterprises better adjust energy optimization strategies to better adapt to market changes. The calculation of the response effect based on the energy optimization guidance can help enterprises evaluate the effectiveness of their electricity strategy and provide a basis for future decision-making. According to the response benefit difference data, the strategy optimization parameters are adjusted so that the enterprise's electricity strategy can be continuously improved and optimized to achieve the maximum benefit. The effect of the whole process is to improve energy utilization efficiency, reduce electricity costs, optimize electricity plans, and help enterprises better respond to market changes and realize intelligent and optimized energy management.

[0011] The present invention also provides a power consumption strategy recommendation system based on scientific energy consumption and settlement by plate, which is used to execute the above-mentioned power consumption strategy recommendation method based on scientific energy consumption and settlement by plate, and the power consumption strategy recommendation system based on scientific energy consumption and settlement by plate includes:

[0012] The data collection module is used to obtain climate data; extract the power grid operation status from the announcements, reports and APIs issued by the power trading center to obtain the power grid operation status data; obtain historical energy price data, and perform periodic analysis on the historical energy price data based on time series to obtain primary energy price data;

[0013] The power supply and demand forecasting module is used to select and extract key features based on the changes in power supply and demand relations according to climate data, power grid operation status data, and primary energy price data, thereby obtaining a power supply and demand feature set; the power supply and demand feature set is used to forecast power spot supply and demand using the gradient boosting tree algorithm, thereby obtaining the intraday power spot supply and demand forecast data and its confidence interval;

[0014] The energy optimization strategy formulation module is used to formulate dynamic energy optimization strategies based on the electricity spot supply and demand forecast data and its confidence interval, so as to obtain energy optimization strategy data. The dynamic energy optimization strategy formulation specifically divides the electricity consumption period into safe low-price area, possible low-price area, price fluctuation area and high-price risk area, and formulates differentiated electricity consumption suggestions for each area;

[0015] The energy consumption guidance optimization module is used to monitor the power supply and demand relationship of the energy consumption optimization strategy data in real time, and compare the actual electricity price with the power spot supply and demand forecast data and its confidence interval to obtain the forecast deviation interval data; according to the forecast deviation interval data, the adjustment mechanism of the energy consumption optimization strategy data is triggered to obtain the energy consumption guidance update data;

[0016] The benefit evaluation and adaptive optimization module is used to calculate the benefits generated by energy-consuming enterprises' responses to energy consumption optimization guidelines based on the updated data of energy consumption guidelines, and to compare the estimated response benefits with the actual response benefits to obtain the response benefit difference data; based on the response benefit difference data, the strategy optimization parameters are adjusted to achieve adaptive optimization of the electricity consumption strategy.

[0017] The present invention achieves a comprehensive understanding of the power market environment by acquiring climate data, power grid operation status data and historical energy price data. Time series prediction based on historical energy price data helps to predict future energy price trends in advance and provide a basis for subsequent decision-making. Using climate data, power grid operation status data and primary energy price data to extract power supply and demand features and use gradient boosting tree algorithm to predict power spot supply and demand helps to accurately predict the supply and demand of the power market and provide support for enterprises to formulate reasonable power consumption strategies. According to the power spot supply and demand forecast data, a dynamic energy optimization strategy is formulated, and the power consumption period is subdivided and differentiated power consumption suggestions are formulated, which helps enterprises to flexibly adjust power consumption strategies under different market conditions to reduce power consumption costs and improve efficiency. By real-time monitoring of the power supply and demand relationship, comparing the actual electricity price with the forecast data, identifying the forecast deviation interval and triggering the adjustment mechanism, and updating the energy consumption guidance data, it helps enterprises to adjust the power consumption strategy in time according to the latest market situation and optimize the power consumption effect. Based on the energy optimization strategy data, benefit calculation and response benefit comparison are performed to obtain the response benefit difference data, and then the strategy optimization parameters are adjusted to achieve adaptive optimization of the power consumption strategy, helping enterprises to maximize power consumption benefits and improve competitiveness. The effect of these modules is to establish a complete closed-loop system for electricity market data analysis and electricity usage strategy optimization, which can dynamically adjust electricity usage strategies based on real-time data, improve electricity efficiency, reduce costs, and enhance the company's ability to adapt to market changes; companies can use this system to make more informed decisions and optimize electricity usage behavior, thereby achieving more sustainable development goals.

[0018] The electricity strategy recommendation system based on scientific energy consumption and sub-plate settlement of the present invention provides third-party settlement services specifically for enterprises participating in scientific energy consumption projects. The so-called "sub-plate" is a means of refined energy management for power sales companies. In addition to the traditional power purchase and sales contract, enterprises and power sales companies are also required to sign a wholesale settlement contract, which clarifies the quantity, price and time-sharing decomposition terms of the annual long-term wholesale agreement. At the same time, it follows the trading rules of the electricity market and sets the boundary conditions for wholesale settlement, so as to conduct independent time-sharing wholesale and retail revenue settlement. In combination with the platform, the relevant parties are encouraged to establish a fair profit-sharing mechanism, which can allow energy-consuming enterprises, investment entities of power-related projects and power sales companies to avoid zero-sum games and effectively obtain long-term real-time incentives. Based on the principle of optimal comprehensive benefits, they work closely together to improve the overall rate of return of each related project, thereby realizing a two-way efficient and coordinated coupling response between the source, grid and load, and supporting all market entities to become self-dispatching terminals to effectively participate in the real-time supply and demand balance of the electricity market. Traditional demand-side response only involves the regulation of the company's controllable load. The system applied for in this patent can accurately calculate the contribution of different power supply projects, energy-saving projects or demand-side response projects in the power system, so that the added contribution value of flexible adjustable loads, self-contained power plants, distributed energy storage, photovoltaics, demand-side response, energy-saving transformation, etc., can be realized in the electricity spot market. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments thereof made with reference to the following drawings:

[0020] Figure 1 It is a schematic diagram of the steps of the method for recommending electricity consumption strategy based on scientific energy consumption and settlement according to the present invention;

[0021] Figure 2 for Figure 1 Detailed step flow diagram of step S1;

[0022] Figure 3 for Figure 1 Detailed step flow chart of step S2 in FIG. DETAILED DESCRIPTION

[0023] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0024] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0025] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0026] To achieve this, please refer to Figures 1 to 3 The present invention provides a method for recommending electricity consumption strategies based on scientific energy consumption and settlement, and the method comprises the following steps:

[0027] Step S1: Acquire climate data; extract the grid operation status from the announcements, reports and APIs issued by the power trading center to obtain grid operation status data; obtain historical energy price data, and perform a periodic analysis of the historical energy price data based on time series to obtain primary energy price data;

[0028] Step S2: Select and extract key features based on the change of power supply and demand relationship according to climate data, power grid operation status data and primary energy price data, so as to obtain a power supply and demand feature set; use the gradient boosting tree algorithm to predict the power supply and demand feature set, and obtain the intraday power spot supply and demand forecast data and its confidence interval;

[0029] Step S3: Formulate a dynamic energy optimization strategy based on the electricity spot supply and demand forecast data and its confidence interval, thereby obtaining energy optimization strategy data, wherein the dynamic energy optimization strategy is specifically formulated by subdividing the electricity consumption period into a safe low-price zone, a possible low-price zone, a price fluctuation zone, and a high-price risk zone, and formulating differentiated electricity consumption recommendations for each zone;

[0030] Step S4: Real-time monitoring of the power supply and demand relationship of the energy optimization strategy data, and comparison of the actual electricity price with the power spot supply and demand forecast data and its confidence interval to obtain forecast deviation interval data; triggering an adjustment mechanism for the energy optimization strategy data according to the forecast deviation interval data, thereby obtaining energy guidance update data;

[0031] Step S5: Calculate the benefits of energy-consuming enterprises responding to energy consumption optimization guidelines based on the updated energy consumption guidelines data, and compare the estimated response benefits with the actual response benefits to obtain response benefit difference data; adjust the strategy optimization parameters based on the response benefit difference data to achieve adaptive optimization of the electricity consumption strategy.

[0032] In the embodiment of the present invention, reference Figure 1 The above is a schematic diagram of the steps of a method for recommending a power consumption strategy based on scientific energy consumption and settlement by plate in the present invention. In this example, the method for recommending a power consumption strategy based on scientific energy consumption and settlement by plate includes the following steps:

[0033] Step S1: Acquire climate data; extract the grid operation status from the announcements, reports and APIs issued by the power trading center to obtain grid operation status data; obtain historical energy price data, and perform a periodic analysis of the historical energy price data based on time series to obtain primary energy price data;

[0034] The embodiment of the present invention uses a meteorological data acquisition API, such as OpenWeatherMap or other meteorological service providers, to regularly acquire real-time climate data, including temperature, humidity, wind speed, etc.; the API is called regularly (such as every hour) to acquire the latest climate data, and saved in a structured format (such as JSON or CSV) for subsequent processing. Use Web Scraping technology or API to access announcements, reports and APIs issued by the power trading center, extract power grid operation status information, such as power grid congestion, power transmission and transformation maintenance information, etc.; Use BeautifulSoup, Scrapy or Requests library in Python for Web Scraping, regularly acquire and parse power grid related announcements and reports, and use customized parsing scripts to extract and save information to the database. Obtain historical energy price data through a historical energy price database or market data API, and use time series analysis tools (such as ARIMA models, Prophet, etc.) for periodic analysis. Extract historical energy price data from the energy market API or database, use Pandas for data cleaning and processing, use time series analysis libraries, such as ARIMA in Statsmodels or Facebook's Prophet, to model and predict the data, and generate future energy price data.

[0035] Step S2: Select and extract key features based on the change of power supply and demand relationship according to climate data, power grid operation status data and primary energy price data, so as to obtain a power supply and demand feature set; use the gradient boosting tree algorithm to predict the power supply and demand feature set, and obtain the intraday power spot supply and demand forecast data and its confidence interval;

[0036] The embodiment of the present invention uses a feature selection algorithm, such as feature selection based on statistics (such as variance analysis) or importance measurement based on a tree model, to select the features most relevant to the changes in the power supply and demand relationship; combines climate data, power grid operation status data and primary energy price data, and uses the feature selection tool in the Scikit-learn library to screen out features that have a significant impact on the power supply and demand relationship, forming a power supply and demand feature set. Use a gradient boosting tree algorithm, such as XGBoost, LightGBM or CatBoost, to forecast the spot supply and demand of electricity; use the power supply and demand feature set as input, and use XGBoost or a similar machine learning framework to train the forecast model, and the model output includes the intraday spot supply and demand forecast data and its confidence interval.

[0037] Among them, the specific implementation method of using the gradient boosting tree algorithm to predict the power spot supply and demand of the power supply and demand feature set to obtain the power spot supply and demand forecast data and its confidence interval within the day can be:

[0038] Input training set samples:

[0039] T={(x 1 ,y 1 ),(x 2 ,y 2 ),...,(x m ,y m )};

[0040] Among them, the maximum number of iterations is T, and the loss function is L.

[0041] Initialize the weak learning machine (estimate the constant value that minimizes the loss function, generally the square loss function is the mean of the node, and the absolute loss function is the median of the node sample):

[0042]

[0043] Among them, for the number of iterations t = 1, 2, ..., T there exists (that is, the number of weak learners generated):

[0044] For sample i=1,2,…,m, calculate the negative gradient (the negative gradient of the loss function in the current model is used as an estimate of the residual, for the squared loss function it is the residual; for the general loss function, it is an approximation of the residual (pseudo-residual)):

[0045]

[0046] Among them, using (x i ,γ i )(i=1,2,...,m), that is, (x 1 ,γ y1 ),...,(x i ,γ t1 ), fit a CART regression tree, and get the tth regression tree. Its corresponding leaf node area is R tj , j = 1, 2, ..., J. Where J is the number of leaf nodes of regression tree t.

[0047] Calculate the best fit value for leaf region j=1,2,...,J:

[0048]

[0049] Update the strong learner:

[0050]

[0051] The final regression tree, that is, the expression of the strong learner f(x), is obtained:

[0052]

[0053] Step S3: Formulate a dynamic energy optimization strategy based on the electricity spot supply and demand forecast data and its confidence interval, thereby obtaining energy optimization strategy data, wherein the dynamic energy optimization strategy is specifically formulated by subdividing the electricity consumption period into a safe low-price zone, a possible low-price zone, a price fluctuation zone, and a high-price risk zone, and formulating differentiated electricity consumption recommendations for each zone;

[0054] According to the predicted power supply and demand and price data, the embodiment of the present invention divides a day into multiple time periods, such as safe low-price area, possible low-price area, price fluctuation area, and high-price risk area; uses data analysis tools, such as Pandas in Python, to set thresholds based on the predicted data and mark different time periods as different power consumption intervals. Differentiated power consumption recommendations are formulated for different power consumption intervals based on cost, risk and energy-saving strategies; a rule engine is written to generate corresponding power consumption recommendations for users based on the power price and supply and demand forecast intervals, providing the optimal time and method for power consumption.

[0055] Among them, the safe low-price zone: users are encouraged to increase electricity consumption as much as possible during this period, because the electricity price is at a stable and relatively low level; the possible low-price zone: production plans are adjusted appropriately to balance electricity demand and costs, while reserving a certain degree of flexibility to cope with possible price fluctuations; the price fluctuation zone: users are advised to maintain conventional electricity consumption patterns, while considering the use of energy storage equipment (such as battery energy storage systems) to smooth the electricity load curve; the high-price risk zone: users are explicitly instructed to reduce non-essential electricity consumption, and give priority to ensuring electricity for key businesses or daily necessities, so as to reduce the risk of wholesale high-priced electricity from the electricity spot market.

[0056] Step S4: Real-time monitoring of the power supply and demand relationship of the energy optimization strategy data, and comparison of the actual electricity price with the power spot supply and demand forecast data and its confidence interval to obtain forecast deviation interval data; triggering an adjustment mechanism for the energy optimization strategy data according to the forecast deviation interval data, thereby obtaining energy guidance update data;

[0057] The embodiment of the present invention uses real-time data stream processing tools, such as Apache Kafka and Apache Flink, to monitor real-time data of the power market; sets up a data stream pipeline to continuously monitor actual electricity prices and power supply and demand forecast data, and detects forecast deviations. According to the deviation data between actual and forecast, the adjustment mechanism is triggered to update the energy use strategy; the preset adjustment mechanism, such as regression analysis or rule triggering, is used to automatically adjust the energy use strategy and generate new energy use guidance data.

[0058] Step S5: Calculate the benefits of energy-consuming enterprises responding to energy consumption optimization guidelines based on the updated energy consumption guidelines data, and compare the estimated response benefits with the actual response benefits to obtain response benefit difference data; adjust the strategy optimization parameters based on the response benefit difference data to achieve adaptive optimization of the electricity consumption strategy.

[0059] The embodiment of the present invention calculates the difference between the estimated response benefit and the actual response benefit through data analysis and statistical methods; uses tools such as Pandas and NumPy to calculate the benefits of the user's electricity consumption data, compares the economic benefits under different strategies, and generates response benefit difference data. According to the response benefit difference data, adjust the parameters in the strategy, such as electricity price sensitivity, load adjustment range, etc.; optimize the strategy parameters through machine learning models or heuristic algorithms, so that the system can be adaptively adjusted under different market conditions, and improve the economic benefits and stability of the strategy.

[0060] The climate data of the present invention is crucial to the operation of the power system. Accurate climate data can help the system better predict load demand, optimize power generation plans, and improve energy scheduling. Extracting grid operation status data from information released by the power trading center can help the system monitor the real-time status of the power grid, promptly discover and solve potential problems, and improve the stability and reliability of the power grid; the analysis and prediction of historical energy price data can help enterprises make more informed energy procurement decisions and plan future electricity costs. By selecting and extracting key features of changes in power supply and demand, the spot supply and demand of electricity can be more accurately predicted, helping enterprises make more effective power decisions. The algorithm can effectively handle complex nonlinear relationships, predict power supply and demand feature sets, provide more accurate power spot supply and demand forecast data, and help enterprises optimize power plans. Through the dynamic energy optimization strategy formulated based on the spot supply and demand forecast data, enterprises can more flexibly adjust power plans to obtain the best power benefits. Real-time monitoring of power supply and demand relationships helps enterprises adjust power strategies in a timely manner to cope with emergencies and fluctuating market conditions. By comparing actual electricity prices with forecast data, forecast deviation interval data can help enterprises better adjust energy optimization strategies to better adapt to market changes. The calculation of the response effect based on the energy optimization guidance can help enterprises evaluate the effectiveness of their electricity strategy and provide a basis for future decision-making. According to the response benefit difference data, the strategy optimization parameters are adjusted so that the enterprise's electricity strategy can be continuously improved and optimized to achieve the maximum benefit. The effect of the whole process is to improve energy utilization efficiency, reduce electricity costs, optimize electricity plans, and help enterprises better respond to market changes and realize intelligent and optimized energy management.

[0061] Preferably, step S1 comprises the following steps:

[0062] Step S11: obtaining temperature, weather conditions, wind speed, rainfall, perceived temperature, average humidity and ultraviolet intensity information from the meteorological service platform, thereby obtaining the climate data to be cleaned, wherein the climate data to be cleaned includes historical climate data and forecast climate data;

[0063] Step S12: Clean the climate data to be cleaned and add a time series index, and convert the data format through one-hot encoding to obtain climate data;

[0064] Step S13: Collecting announcements, reports and API information published by the power trading center platform to obtain power grid text description data;

[0065] Step S14: using natural language processing technology to identify power grid entities and extract their status information from the power grid text description data, thereby obtaining power grid entity status data, wherein the status information includes the entity number, location, operation log and functional scope;

[0066] Step S15: removing duplicate, erroneous or irrelevant data from the grid entity status data, and confirming the current operation status, thereby obtaining initial grid operation status data;

[0067] Step S16: digitally processing the initial power grid operation status data to obtain power grid operation status data, wherein the digital processing specifically constructs a feature vector to represent different maintenance states and blocking states;

[0068] Step S17: Obtain historical energy price data, and perform a time series-based periodic analysis on the historical energy price data to obtain primary energy price data.

[0069] As an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 Detailed step flow diagram of step S1 in the embodiment of the present invention, step S1 includes the following steps:

[0070] Step S11: obtaining temperature, weather conditions, wind speed, rainfall, perceived temperature, average humidity and ultraviolet intensity information from the meteorological service platform, thereby obtaining the climate data to be cleaned, wherein the climate data to be cleaned includes historical climate data and forecast climate data;

[0071] The embodiment of the present invention uses a meteorological service API, such as OpenWeatherMap, WeatherAPI, or the open data interface of the National Meteorological Administration, to obtain relevant climate data; a script is written to regularly call the API to obtain historical and predicted temperature, weather conditions, wind speed, rainfall, perceived temperature, average humidity, and ultraviolet intensity data; the data format is generally JSON or XML, which is converted into a structured data format (such as CSV or a database table) for subsequent processing.

[0072] Step S12: Clean the climate data to be cleaned and add a time series index, and convert the data format through one-hot encoding to obtain climate data;

[0073] The embodiment of the present invention uses the Pandas library in Python to clean the data, clean up missing values, outliers or inconsistent data entries; check the completeness and accuracy of the data to ensure that all necessary climate variables are included. Use the pd.to_datetime() method in Pandas to convert the time field into a time series index; add a time series index to the data set to give it a time attribute to facilitate subsequent time series analysis. Use Pandas or Scikit-learn's OneHotEncoder to perform one-hot encoding, perform one-hot encoding on categorical variables such as weather conditions, and convert them into numerical format for processing by machine learning models.

[0074] Step S13: Collecting announcements, reports and API information published by the power trading center platform to obtain power grid text description data;

[0075] The embodiment of the present invention uses Web Scraping technology, such as Python's BeautifulSoup and Requests library, or directly obtains data through an API interface; regularly collects announcements, reports and API information issued by the power trading center, and saves them in text format for subsequent processing.

[0076] Step S14: using natural language processing technology to identify power grid entities and extract their status information from the power grid text description data, thereby obtaining power grid entity status data, wherein the status information includes the entity number, location, operation log and functional scope;

[0077] The embodiment of the present invention uses an NLP toolkit, such as NLTK, SpaCy or BERT, to perform entity recognition; parses text description data to identify and extract power grid entities and their related status information; and uses named entity recognition (NER) technology to extract entity number, location, operation log, functional scope and other information.

[0078] Step S15: removing duplicate, erroneous or irrelevant data from the grid entity status data, and confirming the current operation status, thereby obtaining initial grid operation status data;

[0079] The embodiment of the present invention uses Pandas to clean and filter data, remove duplicate, erroneous or irrelevant data entries, ensure the accuracy and relevance of the data; confirm the current operating status, such as maintenance or blockage, and retain the latest and most relevant status data.

[0080] Step S16: digitally processing the initial power grid operation status data to obtain power grid operation status data, wherein the digital processing specifically constructs a feature vector to represent different maintenance states and blocking states;

[0081] The embodiment of the present invention uses data processing tools, such as Pandas and NumPy, to construct feature vectors; constructs a feature vector for each power grid entity to represent its different states (such as maintenance, blockage); and converts the entity state into numerical data to facilitate subsequent analysis and input of machine learning models.

[0082] Step S17: Obtain historical energy price data, and perform a time series-based periodic analysis on the historical energy price data to obtain primary energy price data.

[0083] The embodiment of the present invention uses energy market data API or database, such as EIA, OPEC, etc., to obtain historical energy price data, including oil prices, coal prices, natural gas prices, etc. Use time series analysis tools, such as ARIMA model, Facebook Prophet, etc., to clean and organize historical price data to make it suitable for time series modeling; use time series models to predict future prices, generate primary energy price data, and provide prediction confidence intervals.

[0084] The present invention obtains meteorological data such as temperature, weather conditions, wind speed, etc., which can help the system more accurately predict load demand and optimize power generation plans. These data are essential for adjusting energy production and distribution to adapt to climate conditions and demand changes. Through data cleaning and format conversion, the accuracy and consistency of climate data are ensured, which helps to avoid misunderstandings and wrong decisions caused by erroneous or inconsistent data. Collecting power grid text description data can provide information about the operating status and potential problems of the power grid, which helps the system to better understand the power grid status and timely discover and solve problems. Using natural language processing technology to perform entity recognition and information extraction on text data can automatically extract key information from a large amount of text, reduce the workload of manual processing, and improve efficiency and accuracy. Deduplication, removal of erroneous data and confirmation of the current operating status in the data cleaning and confirmation stage can improve data quality, reduce redundant information and errors, and ensure the reliability of data used by the system. By constructing feature vectors to represent different power grid states, digital processing can make data easier to process and analyze, which provides a more reliable data basis for system modeling and prediction, and helps the system better understand the power grid state. Energy price forecasting based on time series helps enterprises to make smarter energy procurement plans, optimize electricity costs, and reduce energy expenditures, which can help enterprises better plan electricity strategies and improve economic benefits. Overall, these steps combine the processing and analysis of meteorological data, power grid information, and energy price data, providing multi-dimensional information support for the system, helping enterprises to better understand and manage energy systems, improve efficiency, reduce costs, and optimize decision-making processes.

[0085] Preferably, step S17 comprises the following steps:

[0086] Step S171: Acquire historical energy price data, including market price data of coal and natural gas;

[0087] The embodiment of the present invention uses energy market data API or database, such as EIA (U.S. Energy Information Administration), OPEC (Organization of Petroleum Exporting Countries), etc., to obtain energy market data; historical market price data of coal and natural gas is obtained through API calls or data downloads. The data format is usually CSV or JSON, and it is converted into a structured data format for subsequent processing.

[0088] Step S172: performing time series analysis based on the long short-term memory network on the historical energy price data, thereby obtaining an energy price memory network model;

[0089] The embodiment of the present invention uses Pandas and NumPy in Python to preprocess data, clean up missing values ​​and outliers, and standardize price data to improve the stability and prediction accuracy of the model. Use a deep learning framework such as TensorFlow or PyTorch to build a long short-term memory network (LSTM), build an LSTM network, and design the architecture of the input layer, multiple LSTM layers, a fully connected layer, and an output layer; the input layer accepts energy price data processed by time series, the LSTM layer processes time dependencies, and the fully connected layer is responsible for the final prediction output. Use historical price data as a training set to train the model, divide the data into a training set and a validation set, and use the LSTM model for training; use a loss function (such as mean square error) and an optimization algorithm (such as Adam) to optimize the model until the model converges.

[0090] Step S173: Calculate the price change trend of coal and natural gas and the periodicity of different time periods according to the energy price memory network model, so as to obtain price change trend data and energy price periodicity data;

[0091] The embodiment of the present invention uses the output of the LSTM model to analyze the price change trend, uses the trained LSTM model to backtest the historical data, identifies the rising and falling trends of prices, extracts the trend characteristics of each time point, and forms the price change trend data. Use Fourier transform or autocorrelation analysis to identify the periodicity of energy prices; perform Fourier transform on the price data to extract the periodicity characteristics, determine the main periodic components through spectrum analysis, and generate energy price periodicity data.

[0092] Step S174: determining the price of primary energy based on the price change trend data and the energy price periodicity data, thereby obtaining primary energy price data.

[0093] The embodiment of the present invention uses the trained LSTM model and the trend and periodicity data generated by it to determine future prices; input future environmental variables (such as climate data, supply and demand data) into the LSTM model to predict future coal and natural gas prices; output price determination data taking into account the model's prediction confidence interval, and provide prediction results under various scenarios. Use data visualization tools such as Matplotlib or Seaborn to display the prediction results; compare the primary energy price results with the actual historical prices to verify the accuracy of the model; generate reports or charts to display price change trends, periodicity and future price determination, providing a reference for decision-making.

[0094] The present invention obtains historical energy price data, including coal and natural gas market price data, which helps the system understand the patterns and trends of past price fluctuations, which provides an important data basis for subsequent analysis and prediction. Through time series analysis based on long-term and short-term memory networks, the long-term dependencies and short-term fluctuations in energy price data can be better captured, and the obtained energy price memory network model can more accurately predict future price trends. Through the energy price memory network model, the changing trend of coal and natural gas prices and the periodicity of different time periods are analyzed, which can help the system better understand the law and periodicity of price fluctuations, which is helpful to formulate corresponding response strategies and decisions. By performing a price forecast of energy on the price change trend data and energy price periodicity data, the system can obtain the forecast results of future prices, which helps enterprises make more intelligent energy procurement decisions, plan electricity costs, and optimize energy management strategies. In summary, through a series of operations such as the acquisition of historical energy price data, time series analysis, and price trend forecasting, the system can better understand and predict the trend of energy prices such as coal and natural gas, and provide important support for enterprises to formulate reasonable energy procurement plans and decisions. The execution of these steps can help enterprises reduce energy costs, optimize energy use structures, and improve competitiveness and benefits.

[0095] Preferably, step S2 comprises the following steps:

[0096] Step S21: integrating climate data, power grid operation status data, and primary energy price data based on time alignment to obtain multi-source data;

[0097] Step S22: extracting feature items from multi-source data and calculating the Pearson correlation coefficient between the feature items, thereby generating a feature correlation heat map;

[0098] Step S23: identifying highly correlated feature pairs on the feature correlation heat map according to a preset correlation threshold, thereby obtaining a feature correlation matrix;

[0099] Step S24: extracting features based on changes in power supply and demand relationships from multi-source data according to the feature correlation matrix, thereby obtaining a power supply and demand feature set;

[0100] Step S25: Divide the power supply and demand feature set into a training set, a validation set, and a test set with complete time series;

[0101] Step S26: Optimize the hyperparameters of the preset gradient boosting tree algorithm according to the power supply and demand feature set using a grid search method, evaluate the model performance according to the validation set using a k-fold cross validation method, select the optimal parameter combination, and thus obtain a gradient boosting tree model;

[0102] Step S27: training the gradient boosting tree model pair according to the training set, thereby obtaining an initially trained power supply and demand prediction model;

[0103] Step S28: Use the power supply and demand forecasting model to predict the power spot supply and demand relationship of the test set, and calculate the confidence interval of the predicted value of power spot supply and demand through the quantitative regression forest method, so as to obtain the intraday power spot supply and demand forecast data and its confidence interval.

[0104] As an embodiment of the present invention, refer to Figure 3 As shown, Figure 1 Detailed step flow diagram of step S2 in the embodiment of the present invention, step S2 includes the following steps:

[0105] Step S21: integrating climate data, power grid operation status data, and primary energy price data based on time alignment to obtain multi-source data;

[0106] The embodiment of the present invention uses data processing tools such as Python's Pandas, SQL databases, etc. to align climate data, power grid operation status data, and primary energy price data according to a unified time series; process missing values ​​and outliers in the data to ensure data integrity, and integrate data from different sources into a comprehensive data set, which is called multi-source data.

[0107] Step S22: extracting feature items from multi-source data and calculating the Pearson correlation coefficient between the feature items, thereby generating a feature correlation heat map;

[0108] The embodiment of the present invention uses Python's Pandas and SciPy libraries to perform feature extraction and correlation calculation, extracts key feature items such as temperature, humidity, wind speed, power grid load status, energy price, etc. from multi-source data; calculates the Pearson correlation coefficient between these feature items to generate a feature correlation matrix; uses visualization tools such as Seaborn to generate a feature correlation heat map to intuitively display the relationship between features.

[0109] Step S23: identifying highly correlated feature pairs on the feature correlation heat map according to a preset correlation threshold, thereby obtaining a feature correlation matrix;

[0110] The embodiment of the present invention still uses Pandas and SciPy, sets a correlation threshold (such as 0.8 or -0.8), and screens out feature pairs with strong correlation; these feature pairs may represent important influencing factors or redundant information, and are listed as key analysis objects to generate a feature correlation matrix.

[0111] Step S24: extracting features based on changes in power supply and demand relationships from multi-source data according to the feature correlation matrix, thereby obtaining a power supply and demand feature set;

[0112] The embodiment of the present invention uses feature selection algorithms such as PCA (principal component analysis) or RFE (recursive feature elimination) to extract key features, selects representative features from the feature correlation matrix, and constructs a power supply and demand feature set; ensures that the feature set can comprehensively describe the changes in the power supply and demand relationship, and removes redundant or irrelevant features.

[0113] Step S25: Divide the power supply and demand feature set into a training set, a validation set, and a test set with complete time series;

[0114] The embodiment of the present invention uses the train_test_split function of the Pandas or Scikit-learn library to divide the power supply and demand feature set into a training set, a validation set, and a test set, with the ratios usually being 70%, 15%, and 15%; ensuring the continuity of each data set in the time series to maintain the temporal characteristics of the data.

[0115] Step S26: Optimize the hyperparameters of the preset gradient boosting tree algorithm according to the power supply and demand feature set using a grid search method, evaluate the model performance according to the validation set using a k-fold cross validation method, select the optimal parameter combination, and thus obtain a gradient boosting tree model;

[0116] The embodiment of the present invention uses GridSearchCV and KFold in Scikit-learn to set the hyperparameter search range of the gradient boosting tree model, including the learning rate, the depth of the tree, the minimum number of samples for sample splitting, etc.; uses a grid search method (GridSearch) to search within the set hyperparameter range, uses a k-fold cross-validation method to evaluate the performance of each set of hyperparameter combinations, and selects the optimal parameter combination.

[0117] Step S27: training the gradient boosting tree model pair according to the training set, thereby obtaining an initially trained power supply and demand prediction model;

[0118] The embodiment of the present invention uses GradientBoostingRegressor or XGBoost in Scikit-learn, uses training set data to train the optimized gradient boosting tree model, monitors the error changes during the training process, and prevents overfitting.

[0119] Step S28: Use the power supply and demand forecasting model to predict the power spot supply and demand relationship of the test set, and calculate the confidence interval of the predicted value of power spot supply and demand through the quantitative regression forest method, so as to obtain the intraday power spot supply and demand forecast data and its confidence interval.

[0120] The embodiment of the present invention uses the QuantileRegressor in Scikit-learn to calculate the confidence interval, uses the trained power supply and demand forecasting model to predict the test set data, and obtains the daily power spot supply and demand forecast value; uses the Quantile Regression Forest method to calculate the confidence interval of the forecast value to quantify the uncertainty of the forecast; outputs the daily power spot supply and demand forecast data and its confidence interval to provide data support for the subsequent dynamic energy optimization strategy.

[0121] The present invention integrates climate data, power grid operation status data and primary energy price data in a time-aligned manner, so that the system can obtain multi-source data and provide a more comprehensive and multi-dimensional information basis, which helps the system to more accurately analyze and predict the power supply and demand relationship. Feature item extraction and Pearson correlation coefficient calculation are performed on multi-source data to generate a feature correlation heat map, which helps to discover the correlation relationship between different features. This can help the system understand the correlation between data and provide a basis for subsequent feature selection and modeling. The feature correlation heat map is screened according to a preset correlation threshold, highly correlated feature pairs are identified, and a feature correlation matrix is ​​generated, which helps the system reduce redundant features and improve modeling efficiency and accuracy. According to the feature correlation matrix, feature extraction of changes in power supply and demand relationships is performed on multi-source data to obtain a power supply and demand feature set, which helps the system better capture the characteristics of power supply and demand relationships and provide more targeted features for modeling and prediction. Dividing the power supply and demand feature set into a training set, a validation set and a test set helps to effectively carry out the training, validation and evaluation stages of the model and ensure the generalization ability and accuracy of the model. The hyperparameters of the gradient boosting tree algorithm are optimized by the grid search method, and the optimal parameter combination is selected through cross-validation to establish a gradient boosting tree model. This helps to improve the performance of the model and enable the model to more accurately predict the relationship between electricity supply and demand. The gradient boosting tree model is trained according to the training set to obtain the electricity supply and demand prediction model after initial training, which provides a basis for the subsequent application and prediction of the model. The trained model is used to predict the electricity supply and demand relationship of the test set, and the confidence interval of the predicted value is calculated by the quantitative regression forest method to obtain the daily electricity spot supply and demand forecast data and confidence interval, which helps the system provide more reliable electricity supply and demand forecast results and provide a reference for decision-making. Combining the above steps, through a series of operations such as data integration, feature extraction, model training and prediction, the system can analyze and predict the relationship between electricity supply and demand more comprehensively and accurately, provide important support for the operation and management of the power system, and help enterprises optimize energy utilization and decision-making efficiency.

[0122] Preferably, step S3 comprises the following steps:

[0123] Step S31: forecasting the spot price of electricity based on the spot supply and demand forecast data and its confidence interval, thereby obtaining the spot price forecast data of electricity and the price confidence interval data;

[0124] The embodiment of the present invention uses a time series prediction model, such as LSTM (Long Short-Term Memory Network) or Prophet, inputs electricity spot supply and demand forecast data and its confidence interval, and uses the LSTM model to train and predict electricity spot prices; the output of the calculation model includes the predicted electricity spot price and its corresponding confidence interval, and the output result includes electricity spot price forecast data and price confidence interval data.

[0125] Step S32: Divide the price range for 24 hours a day according to the electricity spot price forecast data, and group the forecast prices to obtain preliminary price range data;

[0126] The embodiment of the present invention uses data analysis tools, such as Pandas or NumPy, to divide the electricity spot price forecast data into 24-hour time periods within a day; uses statistical analysis methods to group the price data and preliminarily divide multiple price ranges. The grouping can be based on simple quantiles (such as quartiles) or custom price thresholds; and obtains preliminary price range data, including information such as the average price, minimum value, and maximum value of each time period.

[0127] Step S33: adjusting the boundary time period of the preliminary price range data according to the price confidence interval data, thereby obtaining the price range classification data of 24 hours within the day, wherein the price range classification specifically includes a safe low price area, a possible low price area, a price fluctuation area, and a high price risk area;

[0128] The embodiment of the present invention uses statistical analysis and data adjustment tools, such as SciPy and Pandas in Python, and refers to the price confidence interval data to adjust the boundary time periods of the preliminary price interval data to reflect the uncertainty in the price forecast; the principle of adjustment is to expand or merge the price intervals corresponding to the wider confidence intervals to reduce the impact of errors. After the adjustment, four main price intervals are classified: safe low price area, possible low price area, price fluctuation area, and high price risk area.

[0129] Step S34: Formulate a dynamic energy optimization strategy based on the price range classification data to obtain energy optimization strategy data.

[0130] The embodiment of the present invention uses decision analysis tools and optimization algorithms, such as linear programming, dynamic programming, etc., to formulate different energy optimization strategies according to the price range classification data. The strategies corresponding to each interval may include recommended electricity consumption, types of energy used, load adjustment, etc.; safe low-price zone: it is recommended to increase electricity consumption as much as possible and utilize low-priced electricity resources; possible low-price zone: provide cautious electricity consumption suggestions and encourage companies to pay close attention to price changes; price fluctuation zone: it is recommended to adopt flexible electricity consumption strategies, such as adjustable loads or backup power supplies; high-price risk zone: it is recommended to reduce non-essential electricity consumption, or turn to backup energy. The generated energy optimization strategy data includes specific electricity consumption recommendations for each time period to maximize economic benefits and the stability of the power system.

[0131] The present invention predicts the spot price of electricity based on the spot supply and demand forecast data and its confidence interval, and can obtain the spot price forecast data and the price confidence interval data, which helps the system predict the price trend of the electricity market and provides a decision-making basis for market participants. According to the spot price forecast data of electricity, the price interval is divided for 24 hours in a day, and the predicted prices are grouped to obtain preliminary price interval data, which helps to segment the price trend and provide a basis for subsequent price analysis and decision-making. According to the price confidence interval data, the boundary time period of the preliminary price interval data is adjusted to obtain the price interval classification data for 24 hours in a day, including a safe low price area, a possible low price area, a price fluctuation area and a high price risk area, which helps to subdivide the price forecast results into different risk areas and help users better understand the price fluctuation situation. According to the price interval classification data, a dynamic energy optimization strategy is formulated to obtain energy optimization strategy data, which can help users formulate flexible energy strategies according to different price intervals, optimize energy consumption and costs, and improve energy utilization efficiency. Combining the above steps, through a series of operations such as electricity spot price forecasting, price range division, price range classification and dynamic energy optimization strategy formulation, the system can provide users with more detailed electricity market price forecasts and energy optimization strategies, helping users to better cope with market price fluctuations, reduce energy costs, and improve energy efficiency, thereby realizing intelligent and optimized energy management.

[0132] Preferably, step S34 includes the following steps:

[0133] Step S341: transferring the adjustable load to the time period displayed as a safe low-price zone by the price range classification data, and performing benefit calculation for increasing the power consumption, thereby obtaining safe time period strategy data;

[0134] The electricity consumption metering tool and economic benefit calculation model of the embodiment of the present invention transfer the adjustable load to the safe low-price zone as much as possible according to the price range classification data, and use the dispatching software or control system to realize the load transfer; calculate the economic benefit of increasing the electricity consumption, take into account the cost savings of low-priced electricity, use the benefit calculation model, input the increased electricity consumption and the corresponding electricity price, calculate the saved cost and potential benefits; generate safe period strategy data, including the recommended load transfer plan and benefit estimation.

[0135] Step S342: For the time period where the price range classification data shows a possible low price zone, a balance adjustment is made to increase the power consumption strategy and adjust the space strategy, and a risk-benefit ratio assessment is performed based on different power consumption levels, so as to obtain possible time period strategy data;

[0136] The power consumption strategy optimization tool and risk-benefit analysis model of the embodiment of the present invention balance the strategy of increasing power consumption and adjusting the space strategy in the possible low-price period based on historical data and predictions, and formulate a strategy of moderately increasing power consumption in consideration of possible electricity price fluctuations and uncertainties; and use the risk-benefit analysis model to evaluate the risk and benefit ratio of different power consumption levels. Different power consumption and predicted price ranges are input, the corresponding risks and potential benefits are calculated, possible time period strategy data is generated, and the best strategy for increasing power consumption in the possible low-price period is recommended.

[0137] Step S343: maintaining a conventional power consumption mode based on the formulation of a neutral power consumption strategy for the time period when the price range classification data is displayed as a price fluctuation zone, and designing a rapid response mechanism based on real-time price changes, thereby obtaining fluctuation period strategy data;

[0138] The real-time monitoring system and rapid response mechanism design tool of the embodiment of the present invention maintain a neutral electricity consumption strategy during the price fluctuation period, that is, maintain the basic electricity demand and avoid excessive adjustment; design a rapid response mechanism based on real-time price changes. Use a real-time monitoring system to monitor changes in electricity prices, and set up an automation system or operation manual to quickly respond to price fluctuations; generate fluctuation period strategy data, including maintaining regular electricity consumption and rapid response plans to deal with price fluctuations.

[0139] Step S344: formulating a strategy for minimizing electricity consumption for the time period displayed as a high-price risk zone by the price range classification data, thereby obtaining high-price time period strategy data;

[0140] During the high-price risk zone period, the embodiment of the present invention formulates a strategy to minimize electricity consumption, uses a cost control model, and identifies electricity demand that can be reduced or delayed, such as the shutdown of non-critical equipment or the use of backup energy; and generates high-price period strategy data, including specific measures to reduce electricity consumption and expected cost savings.

[0141] Step S345: integrating the safe period strategy data, possible period strategy data, fluctuating period strategy data, and high-price period strategy data based on time integrity, thereby obtaining a differentiated energy optimization strategy for different time periods;

[0142] The embodiment of the present invention integrates the strategy data of safe periods, possible periods, fluctuating periods and high-price periods, and uses time series integration tools to ensure the continuity and consistency of the strategy over time; generates differentiated energy optimization strategies for different time periods, and describes in detail the electricity usage recommendations and adjustment measures for each time period.

[0143] Step S346: Analyze the historical electricity consumption patterns of energy-consuming enterprises, identify adjustable and non-adjustable loads, evaluate the adjustment flexibility and response speed of different types of loads, calculate the maximum adjustable load in each period, and obtain load adjustment potential data;

[0144] The embodiment of the present invention analyzes the historical electricity consumption pattern of the enterprise, uses the electricity consumption pattern analysis tool to identify adjustable and non-adjustable loads; uses the load classification and evaluation model to evaluate the adjustment flexibility and response speed of different types of loads, and calculates the maximum adjustable load in each time period; generates load adjustment potential data, including detailed classification and adjustment potential of adjustable loads.

[0145] Step S347: According to the load adjustment potential data, the time-divided differentiated energy optimization strategy is subjected to cross-time load distribution based on production continuity constraints, thereby obtaining energy optimization strategy data.

[0146] The embodiment of the present invention uses a load distribution optimization tool to distribute load across time periods based on load adjustment potential data to ensure continuity and maximize economic benefits; uses a continuity constraint model to ensure that the production process is not interrupted or production efficiency is not affected by load adjustment, and generates energy optimization strategy data, including detailed load distribution plans and optimization suggestions.

[0147] The present invention transfers the adjustable load to the safe low-price zone period and increases the power consumption. The safe period strategy data is obtained through benefit calculation, which helps to utilize the low-cost electricity in the low-price zone period, increase the power consumption, thereby reducing the energy cost and improving the benefit. For the possible low-price zone period, the power consumption increase strategy and the adjustment space strategy are balanced, and the risk-return ratio is evaluated to obtain the possible period strategy data, which helps to flexibly adjust the power consumption strategy in the potential low-price zone period, maximize the benefit and control the risk. For the price fluctuation zone period, a neutral power consumption strategy is formulated and a rapid response mechanism is designed to obtain the fluctuation period strategy data; this helps to maintain the stability of power consumption in the period with large price fluctuations, and can flexibly and quickly respond to price changes to reduce risks. For the high-price risk zone period, a minimum power consumption strategy is formulated to obtain the high-price period strategy data, which helps to minimize the power consumption in the high-price risk period and reduce cost losses. Based on the safe period, possible period, fluctuation period and high-price period strategy data, a time integrity strategy integration is performed to obtain a differentiated energy optimization strategy for different time periods, which helps to formulate a flexible power consumption strategy according to the market conditions and risk characteristics of different time periods, maximize benefits and reduce risks. Through the analysis of historical electricity consumption patterns, the flexibility and response speed of adjustable and non-adjustable loads are evaluated, the maximum adjustable load in each time period is calculated, and the load adjustment potential data is obtained, which helps to determine the load adjustment potential of energy-consuming enterprises and provide a basis for cross-time load distribution. Based on the load adjustment potential data, the cross-time load distribution with continuity constraints is carried out for the differentiated energy optimization strategy for different time periods, and the energy optimization strategy data is obtained, which helps to achieve cross-time energy optimization according to the load adjustment potential, improve energy utilization efficiency, and reduce costs. Through the execution of the above series of steps, the energy strategy can be finely adjusted and optimized, so that enterprises can flexibly adjust the electricity strategy according to market conditions, minimize costs, improve benefits, and ensure the sustainability and stability of energy use.

[0148] Preferably, step S4 comprises the following steps:

[0149] Step S41: Real-time monitoring of the power supply and demand relationship of the energy optimization strategy data to obtain real-time power supply and demand data;

[0150] The real-time monitoring system, smart meter and data collection platform of the embodiment of the present invention use the smart meter and the data collection platform to monitor the power supply and demand relationship in real time, including actual power consumption, power generation and power market transaction data; the collected data is integrated into the monitoring system to form real-time power supply and demand data, and the real-time power supply and demand data is generated for subsequent analysis and adjustment of energy use strategies.

[0151] Step S42: Calculate the real-time deviation between the real-time electricity price and the predicted electricity price based on the real-time electricity supply and demand data and the electricity spot supply and demand forecast data, and evaluate the recent forecast accuracy trend through the sliding window technology to obtain the deviation trend characteristic data;

[0152] The embodiment of the present invention compares the real-time electricity supply and demand data with the electricity spot supply and demand forecast data, calculates the deviation between the real-time electricity price and the forecasted electricity price; uses the sliding window technology to perform trend evaluation on the recent forecast accuracy, selects the appropriate window size to capture the changing trend of the forecast deviation in the short term, generates deviation trend feature data, and reflects the changing trend and volatility of the recent forecast deviation.

[0153] Step S43: performing deviation pattern analysis based on time characteristics on the deviation trend characteristic data, thereby obtaining predicted deviation real-time data;

[0154] The embodiment of the present invention extracts time characteristics from the deviation trend feature data, such as intraday periodicity, seasonal changes, etc., uses a pattern recognition algorithm to analyze the time characteristics of the deviation data, identifies common deviation patterns, such as peak deviation, trough deviation, etc., generates real-time data of predicted deviation, and provides detailed pattern and feature information of the deviation.

[0155] Step S44: judging whether the actual electricity price deviates significantly from the forecast based on the confidence interval of the electricity spot supply and demand forecast data, quantifying the degree of deviation, marking the time periods of continuous deviation, and obtaining forecast deviation interval data;

[0156] The embodiment of the present invention determines whether the real-time electricity price deviates significantly from the forecast based on the confidence interval of the electricity spot supply and demand forecast data. If the actual electricity price exceeds the confidence interval range, it is considered to be significantly deviated. The degree of deviation is quantified, the gap between the actual electricity price and the forecasted electricity price is calculated, and its statistical significance is evaluated. The time periods of continuous deviation are marked, the cycle and characteristics of continuous deviation are identified, and the forecast deviation interval data is generated, and the time periods of significant deviation and continuous deviation are marked.

[0157] Step S45: triggering an adjustment mechanism for the energy usage optimization strategy data according to the predicted deviation interval data to obtain energy usage guidance update data.

[0158] The embodiment of the present invention triggers the adjustment mechanism of the energy consumption optimization strategy according to the predicted deviation interval data, and determines the degree and direction of the adjustment according to the size and trend of the actual deviation; updates the energy consumption guidance data, and provides new electricity consumption suggestions and optimization measures. For example, if the actual electricity price is higher than the forecast, it is recommended to reduce electricity consumption or transfer to backup energy; if it is lower than the forecast, it is recommended to increase electricity consumption or postpone load transfer; generates energy consumption guidance update data, which includes detailed adjustment suggestions and execution guidelines for reference and execution by energy-consuming enterprises.

[0159] The present invention can understand the supply and demand status of the power market in real time by monitoring the real-time data of the power supply and demand relationship, help enterprises make timely decisions to cope with changes in the power market, and ensure stable and economic power consumption. By calculating the deviation between the real-time power price and the predicted power price, and evaluating the trend of the prediction accuracy, the trend of power price fluctuation can be discovered in time, helping enterprises to better predict the power cost and optimize the power consumption strategy. Through the deviation pattern analysis of time characteristics, the change law of deviation data can be better understood, helping enterprises to adjust the power consumption strategy to cope with different market conditions and improve the efficiency of power consumption cost. According to the confidence interval, the real-time data of the prediction deviation is judged, the degree of deviation between the actual power price and the predicted power price can be quantified, and the period of continuous deviation is marked to help enterprises identify abnormal situations and take measures to avoid losses in time. According to the prediction deviation interval data, the adjustment mechanism of the energy optimization strategy is triggered, and the energy guidance data is updated to help enterprises make flexible power consumption decisions according to the real-time market situation, minimize costs and improve benefits. The effect of these steps is to help enterprises monitor the supply and demand of the power market in real time, predict power price fluctuations and evaluate accuracy, analyze deviation data and identify abnormal situations, so as to adjust the power consumption strategy in time to maximize benefits and reduce risks. This refined management approach based on real-time and predictive data can make companies more competitive while improving electricity efficiency and sustainability.

[0160] Preferably, step S45 includes the following steps:

[0161] Step S451: identifying the time interval and degree of deviation of the current electricity price from the forecast according to the forecast deviation interval data, and classifying based on the degree of deviation, thereby obtaining deviation degree classification data;

[0162] The embodiment of the present invention uses a time series prediction model to analyze the prediction deviation interval data, predict the time interval of continuous deviation of the current electricity price and its degree of deviation; according to the prediction result, the degree of deviation is quantified, and a classification algorithm is used to classify it into different levels, such as mild, moderate, severe, etc., to generate deviation degree classification data, identifying the deviation degree and classification information of each time period.

[0163] Step S452: different policy adjustment triggering thresholds are set according to the deviation degree classification data, and a cumulative triggering mechanism for continuous deviation periods is added to obtain policy adjustment triggering condition data;

[0164] The embodiment of the present invention classifies data according to the degree of deviation and sets thresholds for triggering different policy adjustments. For example, for severe deviations, a lower trigger threshold is set to respond quickly. A cumulative trigger mechanism is designed. When a certain number of consecutive deviation periods are reached, even if the degree of deviation is relatively minor, policy adjustment is triggered, and policy adjustment trigger condition data is generated, including trigger thresholds under different degrees of deviation and specific settings of the cumulative trigger mechanism.

[0165] Step S453: Evaluate the benefits of each price range on the energy optimization strategy data, thereby obtaining strategy evaluation report data;

[0166] The embodiment of the present invention conducts benefit evaluation on energy optimization strategy data in each price range, calculates potential benefits or cost savings under different strategies; comprehensively considers factors such as electricity price changes and electricity consumption adjustments, generates a strategy evaluation report, includes the expected benefits and risk analysis of each strategy, generates strategy evaluation report data, and provides detailed benefit analysis results under each price range.

[0167] Step S454: Estimating the adjustment coefficient of each price range according to the deviation degree classification data and the strategy evaluation report data, thereby obtaining dynamic adjustment coefficient data;

[0168] The embodiment of the present invention estimates the adjustment coefficients of each price range based on the strategy evaluation report data and the deviation degree classification data. These coefficients are used to adjust the execution intensity of the electricity consumption strategy, such as increasing the intensity of electricity reduction in the case of large deviations; based on the difference between the actual electricity price and the predicted electricity price, the adjustment coefficients are updated in real time to ensure the flexibility and accuracy of the energy optimization strategy, generate dynamic adjustment coefficient data, and provide adjustment coefficient information for each price range.

[0169] Step S455: triggering an adjustment mechanism for the energy usage optimization strategy data according to the dynamic adjustment coefficient data and the strategy adjustment trigger condition data, thereby obtaining energy usage guidance update data.

[0170] The embodiment of the present invention triggers the adjustment mechanism of the energy consumption optimization strategy according to the dynamic adjustment coefficient data and the strategy adjustment trigger condition data, and the adjustment mechanism is adjusted according to the real-time electricity market and energy consumption conditions; updates the energy consumption guidance data, provides the latest electricity consumption suggestions and strategy adjustment measures, and helps enterprises optimize energy consumption strategies within different electricity price ranges; generates energy consumption guidance update data, including specific adjustment suggestions and implementation plans, to guide the actual operations of energy-consuming enterprises.

[0171] The present invention can more precisely understand the deviation situation by identifying the time interval and degree of deviation of the current electricity price that continuously deviates from the forecast, and classifying based on the degree of deviation, so as to help enterprises take corresponding countermeasures according to different situations. By setting different policy adjustment trigger thresholds and accumulating the trigger mechanism of continuous deviation periods, it is helpful to establish a flexible policy adjustment system, trigger the adjustment in time according to the actual situation, and ensure the effectiveness of the energy optimization strategy. Through the benefit evaluation of each price range, it can help enterprises understand the energy performance under different price ranges, provide a reference basis for future decision-making, and optimize the electricity strategy. According to the deviation degree classification data and the policy evaluation report data, the adjustment coefficient of each price range is calculated, and the dynamic adjustment of the electricity strategy can be realized, the adaptability to market fluctuations can be improved, and the electricity cost can be reduced. According to the dynamic adjustment coefficient data and the policy adjustment trigger condition data, the adjustment mechanism of the energy optimization strategy data is triggered, and the energy guidance data is updated, which helps enterprises to adjust the electricity strategy in real time according to the latest market conditions and improve the efficiency of electricity use. The effect of these steps is to establish a complete closed-loop system for data analysis and strategy adjustment, which can dynamically adjust electricity consumption strategies based on real-time data, improve corporate electricity efficiency, reduce costs, and enhance market competitiveness; through the classification and evaluation of deviations, and dynamic adjustment of strategies and guidance updates, companies can respond more flexibly to changes in the electricity market, maximize energy conservation and emission reduction benefits, and achieve sustainable development.

[0172] Preferably, step S5 comprises the following steps:

[0173] Step S51: Calculate the actual power consumption data in each time period according to the pre-acquired actual power consumption behavior data; calculate the power consumption of the energy-consuming enterprise in response to the suggestions in the optimized energy consumption guide according to the energy consumption guide update data, so as to obtain the recommended power consumption data;

[0174] The embodiment of the present invention uses the power consumption monitoring system to collect the actual power consumption of the enterprise in each period, and compares these data with the recommended power consumption in the energy consumption guidance update data; according to the energy consumption guidance update data, the recommended power consumption of the enterprise in each period is calculated. These suggestions are based on the optimization results of the current market conditions and the power demand of the enterprise, and the actual power consumption data and the recommended power consumption data are generated to provide basic data for subsequent response evaluation and benefit calculation.

[0175] Step S52: evaluating the response degree of the actual power consumption data and the recommended power consumption data, and calculating the real power consumption cost, so as to obtain the actual response benefit data;

[0176] The embodiment of the present invention compares the actual electricity consumption data with the recommended electricity consumption data, evaluates the degree to which the enterprise responds to the energy consumption guidelines, calculates the response deviation, and analyzes the enterprise's implementation of the recommendations; based on the actual electricity consumption and real-time electricity price data, calculates the real electricity cost of each time period, takes into account the impact of electricity price fluctuations and actual electricity consumption behavior, accurately evaluates the cost, generates a response degree assessment report and actual response benefit data, and provides detailed information on the effectiveness of the enterprise's response strategy.

[0177] Step S53: comparing the estimated response benefit with the actual response benefit according to the actual response benefit data, thereby obtaining response benefit difference data;

[0178] The embodiment of the present invention calculates the difference between the actual response revenue data and the estimated response revenue data, analyzes the reasons for the revenue difference, such as actual electricity price fluctuations, response delays, etc.; quantifies the gap between the actual response revenue and the estimated revenue, and generates response revenue difference data, which reflects the degree of difference between the actual execution effect and the expected target; generates response revenue difference data as an important reference for strategy optimization.

[0179] Step S54: adjusting the strategy optimization parameters based on the response revenue difference data to achieve adaptive optimization of the power consumption strategy.

[0180] The embodiment of the present invention adjusts key parameters in the power consumption strategy, such as load adjustment ratio, response speed, etc., based on the response revenue difference data, and uses optimization algorithms to ensure that the adjusted strategy can better adapt to market changes and enterprise needs; implements an adaptive optimization mechanism to update the power consumption strategy regularly or in real time. According to the latest market data and enterprise performance, the strategy is continuously adjusted to achieve the optimal energy use effect, and an updated power consumption strategy is generated to ensure that the enterprise can obtain the best power cost-effectiveness under different market conditions.

[0181] The present invention can help enterprises understand the difference between actual power consumption and recommended power consumption by calculating actual power consumption data and recommended power consumption data, and provide basic data for subsequent response degree evaluation and cost calculation. By evaluating the difference between actual power consumption and recommended power consumption, and calculating the real power consumption cost, the response degree of the enterprise to the energy consumption guidance can be quantified, the benefits brought by the actual response can be evaluated, and the enterprise can optimize the power consumption strategy. By comparing and analyzing the estimated response benefits and the actual response benefits, the difference in response benefits can be found, which helps enterprises understand the gap between actual benefits and expected benefits, and provides a basis for further optimizing strategies. Based on the response benefit difference data, the strategy optimization parameters are adjusted to achieve adaptive optimization of the power consumption strategy, help enterprises better respond to market changes, improve power efficiency, and maximize the savings in power costs. The effect of these steps is to establish a closed-loop system, from actual power consumption data to recommended power consumption data to response degree evaluation and cost calculation, and finally to strategy optimization parameter adjustment, to achieve comprehensive monitoring and optimization of the enterprise's power consumption behavior. By analyzing the response-benefit difference, companies can adjust their strategies in a timely manner, improve their responsiveness to energy consumption guidelines, and thereby optimize electricity consumption behavior, reduce electricity costs, improve efficiency, and achieve sustainable development.

[0182] The present invention also provides a power consumption strategy recommendation system based on scientific energy consumption and settlement by plate, which is used to execute the above-mentioned power consumption strategy recommendation method based on scientific energy consumption and settlement by plate, and the power consumption strategy recommendation system based on scientific energy consumption and settlement by plate includes:

[0183] The data collection module is used to obtain climate data; extract the power grid operation status from the announcements, reports and APIs issued by the power trading center to obtain the power grid operation status data; obtain historical energy price data, and perform periodic analysis on the historical energy price data based on time series to obtain primary energy price data;

[0184] The power supply and demand forecasting module is used to select and extract key features based on the changes in power supply and demand relations according to climate data, power grid operation status data, and primary energy price data, thereby obtaining a power supply and demand feature set; the power supply and demand feature set is used to forecast power spot supply and demand using the gradient boosting tree algorithm, thereby obtaining the intraday power spot supply and demand forecast data and its confidence interval;

[0185] The energy optimization strategy formulation module is used to formulate dynamic energy optimization strategies based on the electricity spot supply and demand forecast data and its confidence interval, so as to obtain energy optimization strategy data. The dynamic energy optimization strategy formulation specifically divides the electricity consumption period into safe low-price area, possible low-price area, price fluctuation area and high-price risk area, and formulates differentiated electricity consumption suggestions for each area;

[0186] The energy consumption guidance optimization module is used to monitor the power supply and demand relationship of the energy consumption optimization strategy data in real time, and compare the actual electricity price with the power spot supply and demand forecast data and its confidence interval to obtain the forecast deviation interval data; according to the forecast deviation interval data, the adjustment mechanism of the energy consumption optimization strategy data is triggered to obtain the energy consumption guidance update data;

[0187] The benefit evaluation and adaptive optimization module is used to calculate the benefits generated by energy-consuming enterprises' responses to energy consumption optimization guidelines based on the updated data of energy consumption guidelines, and to compare the estimated response benefits with the actual response benefits to obtain the response benefit difference data; based on the response benefit difference data, the strategy optimization parameters are adjusted to achieve adaptive optimization of the electricity consumption strategy.

[0188] The present invention achieves a comprehensive understanding of the power market environment by acquiring climate data, power grid operation status data and historical energy price data. Time series prediction based on historical energy price data helps to predict future energy price trends in advance and provide a basis for subsequent decision-making. Using climate data, power grid operation status data and primary energy price data to extract power supply and demand features and use gradient boosting tree algorithm to predict power spot supply and demand helps to accurately predict the supply and demand of the power market and provide support for enterprises to formulate reasonable power consumption strategies. According to the power spot supply and demand forecast data, a dynamic energy optimization strategy is formulated, and the power consumption period is subdivided and differentiated power consumption suggestions are formulated, which helps enterprises to flexibly adjust power consumption strategies under different market conditions to reduce power consumption costs and improve efficiency. By real-time monitoring of the power supply and demand relationship, comparing the actual electricity price with the forecast data, identifying the forecast deviation interval and triggering the adjustment mechanism, and updating the energy consumption guidance data, it helps enterprises to adjust the power consumption strategy in time according to the latest market situation and optimize the power consumption effect. Based on the energy optimization strategy data, benefit calculation and response benefit comparison are performed to obtain the response benefit difference data, and then the strategy optimization parameters are adjusted to achieve adaptive optimization of the power consumption strategy, helping enterprises to maximize power consumption benefits and improve competitiveness. The effect of these modules is to establish a complete closed-loop system for electricity market data analysis and electricity usage strategy optimization, which can dynamically adjust electricity usage strategies based on real-time data, improve electricity efficiency, reduce costs, and enhance the company's ability to adapt to market changes; companies can use this system to make more informed decisions and optimize electricity usage behavior, thereby achieving more sustainable development goals.

[0189] The benefit calculation for energy-consuming enterprises responding to energy optimization guidance according to energy guidance update data in the embodiment of the present invention can be specifically calculated as follows:

[0190] The electricity energy and electricity charges are settled according to the electricity price signed in the contract between the enterprise and the power sales company, which can be roughly divided into two situations: industrial users in Shenzhen area and industrial users outside Shenzhen area.

[0191] Among them, large industrial users outside Shenzhen:

[0192] The comprehensive retail electricity fee for off-peak period = fixed price electricity ratio * fixed flat period retail price * 0.38 + linkage price electricity ratio * linkage market price * 0.38 + floating electricity fee for off-peak period.

[0193] The comprehensive electricity price of the retail electricity in the normal period = the proportion of electricity with fixed price * the fixed retail price in the normal period + the proportion of electricity with linkage price * the linkage market price + the floating electricity price of the electricity.

[0194] Peak-period electricity retail comprehensive electricity fee = fixed price electricity ratio * fixed flat-period retail price * 1.7 + linkage price electricity ratio * linkage market price * 1.7 + floating electricity fee for each kWh.

[0195] Total electricity charges collected by the power sales company = off-peak electricity consumption * off-peak electricity retail comprehensive electricity charges + normal electricity consumption * normal electricity retail comprehensive electricity charges + peak electricity consumption * peak electricity retail comprehensive electricity charges;

[0196] Large industrial users in Shenzhen:

[0197] The comprehensive retail electricity fee for off-peak period = fixed price electricity ratio * fixed flat period retail price * 0.32 + linkage price electricity ratio * linkage market price * 0.32 + floating electricity fee for off-peak period.

[0198] The comprehensive electricity price of the retail electricity in the normal period = the proportion of electricity with fixed price * the fixed retail price in the normal period + the proportion of electricity with linkage price * the linkage market price + the floating electricity price of the electricity.

[0199] Peak-period electricity retail comprehensive electricity fee = fixed price electricity ratio * fixed flat-period retail price * 1.53 + linkage price electricity ratio * linkage market price * 1.53 + floating electricity fee for each kWh.

[0200] The total electricity charge collected by the power sales company = off-peak electricity consumption * comprehensive retail electricity charge for off-peak electricity + normal period electricity consumption * comprehensive retail electricity charge for normal period + peak period electricity consumption * comprehensive retail electricity charge for peak period.

[0201] Actual total wholesale cost = Σ(annual long-term contract electricity * annual long-term contract price + monthly medium- and long-term electricity * monthly medium- and long-term transaction comprehensive price + (day-ahead declared electricity - long-term contract electricity - monthly medium- and long-term electricity) * day-ahead price + (actual electricity - day-ahead declared electricity) * real-time price + deviation profit transfer + peak-valley balance funds + spot sharing costs + monthly assessment electricity charges + transaction fees);

[0202] ① Annual long-term contract electricity: Based on the total electricity consumption of the enterprise in the previous year (based on the time period of the enterprise's historical electricity consumption announced by the trading center), the enterprise and the power sales company need to negotiate the long-term contract ratio for the next year. The total annual long-term contract electricity of the enterprise is equal to the total annual electricity consumption multiplied by the long-term contract ratio. According to the typical reference curves of the year-month, month-day, and day-time published by the power trading center, the total annual long-term contract electricity of the enterprise is decomposed into every hour of the whole year.

[0203] ②Annual long-term contract price: the annual average long-term contract price of the electricity market in that year announced by the power trading center.

[0204] ③ Monthly medium- and long-term electricity volume: The enterprise evaluates whether it needs monthly trading electricity volume, and sends a written transaction demand to the power sales company in accordance with the provincial electricity spot trading rules and electricity market conditions. The power sales company will reply to the enterprise in writing on the specific settlement electricity volume and curve based on the market monthly transaction results.

[0205] ④ Monthly medium- and long-term transaction comprehensive price: the monthly transaction comprehensive settlement price announced by the power trading center.

[0206] ⑤ Electricity volume reported the day before: Enterprises must submit the market electricity purchase volume for each hour of the next day to the system before 12:00 every day, and it will serve as the basis for wholesale settlement in the spot market.

[0207] ⑥ Day-ahead price: The power dispatching agency comprehensively evaluates the unified load forecast, bus load forecast, power generation unit maintenance plan, power transmission and transformation equipment maintenance plan, power generation unit operation constraints and other factors for centralized optimization calculation, and clears the time-sharing node electricity price on the operating day. It can be simply understood that the day-ahead price is the price cleared by the dispatching agency based on the predicted power supply and demand situation of the next day's market, and the hourly day-ahead price is uniformly released by the power trading center system.

[0208] ⑦ Actual electricity: the amount of electricity actually purchased by the enterprise in the market per hour.

[0209] ⑧ Real-time price: The power dispatching agency optimizes and calculates based on the latest grid operation status and ultra-short-term load forecast information, taking into account factors such as the operating constraints of the generator set and the safe operation constraints of the grid, and clears the real-time node electricity price on the operating day. It can be simply understood as the price that is actually cleared based on the actual supply and demand situation of the market on that day. The real-time price per hour is uniformly released by the power trading center system.

[0210] ⑨ Deviation profit: Since there is a deviation between the clearing price published by the trading center every hour and the actual clearing real-time price, there may also be a deviation between the electricity consumption reported by the enterprise every hour and the actual electricity consumption, so deviation profit may be generated every hour, which can be positive or negative.

[0211] ⑩ Deviation profit transfer: Deviation profit may be generated every hour when entering the spot market settlement. In order to avoid huge market risks caused by market players magnifying the speculative arbitrage multiples, the latest policy content for 2023 released by the trading center limits the hours when the deviation profit is positive. Market players can only obtain a profit of the actual electricity consumption in that hour * 20%, and the rest will be recovered uniformly.

[0212] Peak-valley balancing funds: Since there is no set peak-valley coefficient in the wholesale market, the government has formulated peak-valley balancing funds to balance the peak-valley electricity charges of power sales companies. According to the peak-valley coefficients of retail electricity charges of 1.7 and 0.38 (1.53 and 0.32 in Shenzhen), the peak electricity charges are recovered at a unit price of 0.463*(1.7-1)=0.324 yuan / kWh to deduct the overcharged electricity charges during the peak period; the valley electricity charges are recovered at a unit price of 0.463*(1-0.38)=0.287 yuan / kWh to make up for the undercharged electricity charges during the valley period.

[0213] Spot electricity cost sharing: Spot market cost sharing includes but is not limited to: market congestion surplus, system operation compensation, startup compensation, power generation and consumption imbalance, etc. The total spot cost sharing amount is settled every month according to the actual situation of the power market, and divided by the actual power consumption of the whole market that month to obtain the spot cost sharing amount of the market power. The trading center calculates and issues the settlement results.

[0214] The electricity sales company's spot electricity fee = actual total electricity consumption in the month * spot electricity fee amount.

[0215] Monthly assessment of electricity charges by power sales companies: The assessment of electricity charges by power sales companies is divided into two parts, and the larger value is taken for assessment.

[0216] Part I: Monthly electricity consumption declaration and assessment fees:

[0217] Exemption range: Actual power*±5%

[0218] Unit price for electricity assessment: |Monthly centralized competitive transaction comprehensive price - Monthly comprehensive price of the day-ahead market|*0.2

[0219] Part II: Monthly mid- and long-term deviation assessment:

[0220] Exemption from assessment scope: medium and long-term electricity consumption / actual electricity consumption not less than 90%.

[0221] Unit price for electricity assessment per kilowatt-hour: (monthly comprehensive price of centralized competitive transactions - monthly comprehensive price of the day-ahead market)*1.1. If it is less than 0, it will be taken as 0.

[0222] Transaction fees: Market entities conducting wholesale electricity transactions in the electricity market need to pay a certain transaction fee to the trading center based on the amount of electricity they generate. The fee for long-term contracts and monthly electricity is 0.14 cents per kilowatt-hour, and the fee for spot, multi-day and weekly electricity is 0.18 cents per kilowatt-hour (in accordance with the latest policy released by the trading center).

Claims

1. A method for recommending electricity consumption strategies based on scientific energy consumption and settlement, characterized in that: The following steps are involved: Step S1: Obtain climate data; Extract the grid operation status from the announcements, reports and APIs issued by the power trading center to obtain grid operation status data; obtain historical energy price data and perform periodic analysis on the historical energy price data based on time series to obtain primary energy price data; Step S2: Select and extract key features based on the change of power supply and demand relationship according to climate data, power grid operation status data and primary energy price data, so as to obtain a power supply and demand feature set; use the gradient boosting tree algorithm to predict the power supply and demand feature set, and obtain the intraday power spot supply and demand forecast data and its confidence interval; Step S3: Formulate a dynamic energy optimization strategy based on the electricity spot supply and demand forecast data and its confidence interval, thereby obtaining energy optimization strategy data, wherein the dynamic energy optimization strategy is specifically formulated by subdividing the electricity consumption period into a safe low-price zone, a possible low-price zone, a price fluctuation zone, and a high-price risk zone, and formulating differentiated electricity consumption recommendations for each zone; Step S4: Real-time monitoring of the power supply and demand relationship of the energy optimization strategy data, and comparison of the actual electricity price with the power spot supply and demand forecast data and its confidence interval to obtain forecast deviation interval data; triggering an adjustment mechanism for the energy optimization strategy data according to the forecast deviation interval data, thereby obtaining energy guidance update data; Step S5: Calculate the benefits of energy-consuming enterprises responding to the energy optimization guidelines according to the energy consumption guideline update data, and compare the estimated response benefits with the actual response benefits to obtain response benefit difference data; The strategy optimization parameters are adjusted based on the response revenue difference data to achieve adaptive optimization of the power consumption strategy.

2. The method for recommending electricity consumption strategies based on scientific energy consumption and settlement according to claim 1 is characterized in that: Step S1 includes the following steps: Step S11: obtaining temperature, wind speed, rainfall, perceived temperature, average humidity and ultraviolet intensity information from the meteorological service platform, thereby obtaining the climate data to be cleaned, wherein the climate data to be cleaned includes historical climate data and forecast climate data; Step S12: Clean the climate data to be cleaned and add a time series index, and convert the data format through one-hot encoding to obtain climate data; Step S13: Collecting announcements, reports and API information published by the power trading center platform to obtain power grid text description data; Step S14: using natural language processing technology to identify power grid entities and extract their status information from the power grid text description data, thereby obtaining power grid entity status data, wherein the status information includes the entity number, location, operation log and functional scope; Step S15: removing duplicate, erroneous or irrelevant data from the grid entity status data, and confirming the current operation status, thereby obtaining initial grid operation status data; Step S16: digitally processing the initial power grid operation status data to obtain power grid operation status data, wherein the digital processing specifically constructs a feature vector to represent different maintenance states and blocking states; Step S17: Obtain historical energy price data, and perform a time series-based periodic analysis on the historical energy price data to obtain primary energy price data.

3. The method for recommending electricity consumption strategies based on scientific energy consumption and settlement according to claim 2 is characterized in that: Step S17 includes the following steps: Step S171: Acquire historical energy price data, including market price data of coal and natural gas; Step S172: performing time series analysis based on the long short-term memory network on the historical energy price data, thereby obtaining an energy price memory network model; Step S173: performing price change trend analysis of coal and natural gas and periodicity analysis of different time periods according to the energy price memory network model, thereby obtaining price change trend data and energy price periodicity data; Step S174: determining the price of primary energy based on the price change trend data and the energy price periodicity data, thereby obtaining primary energy price data.

4. The method for recommending electricity consumption strategies based on scientific energy consumption and settlement according to claim 3 is characterized in that: Step S2 includes the following steps: Step S21: integrating climate data, power grid operation status data, and primary energy price data based on time alignment to obtain multi-source data; Step S22: extracting feature items from multi-source data and calculating the Pearson correlation coefficient between the feature items, thereby generating a feature correlation heat map; Step S23: identifying highly correlated feature pairs on the feature correlation heat map according to a preset correlation threshold, thereby obtaining a feature correlation matrix; Step S24: extracting features based on changes in power supply and demand relationships from multi-source data according to the feature correlation matrix, thereby obtaining a power supply and demand feature set; Step S25: Divide the power supply and demand feature set into a training set, a validation set, and a test set with complete time series; Step S26: Optimize the hyperparameters of the preset gradient boosting tree algorithm according to the power supply and demand feature set using a grid search method, evaluate the model performance according to the validation set using a k-fold cross validation method, select the optimal parameter combination, and thus obtain a gradient boosting tree model; Step S27: training the gradient boosting tree model according to the training set, thereby obtaining an initially trained power supply and demand prediction model; Step S28: Use the power supply and demand forecasting model to predict the power spot supply and demand relationship according to the test set, and calculate the confidence interval of the predicted value of power spot supply and demand through the quantitative regression forest method, so as to obtain the intraday power spot supply and demand forecast data and its confidence interval.

5. The method for recommending electricity consumption strategies based on scientific energy consumption and settlement according to claim 4 is characterized in that: Step S3 includes the following steps: Step S31: forecasting the spot price of electricity based on the spot supply and demand forecast data and its confidence interval, thereby obtaining the spot price forecast data of electricity and the price confidence interval data; Step S32: dividing the price distribution of 24 hours a day into intervals according to the electricity spot price forecast data, and grouping the forecast prices, thereby obtaining preliminary price interval data; Step S33: adjusting the boundary time period of the preliminary price range data according to the price confidence interval data, thereby obtaining the price range classification data of 24 hours within the day, wherein the price range classification specifically includes a safe low price area, a possible low price area, a price fluctuation area, and a high price risk area; Step S34: Formulate a dynamic energy optimization strategy based on the price range classification data to obtain energy optimization strategy data.

6. The method for recommending electricity consumption strategies based on scientific energy consumption and settlement according to claim 5 is characterized in that: Step S34 includes the following steps: Step S341: transferring the adjustable load to the time period displayed as a safe low-price zone by the price range classification data, and performing benefit calculation for increasing the power consumption, thereby obtaining safe time period strategy data; Step S342: For the time period where the price range classification data shows a possible low price zone, a balance adjustment is made to increase the power consumption strategy and adjust the space strategy, and a risk-benefit ratio assessment is performed based on different power consumption levels, so as to obtain possible time period strategy data; Step S343: maintaining a conventional power consumption mode based on the formulation of a neutral power consumption strategy for the time period when the price range classification data is displayed as a price fluctuation zone, and designing a rapid response mechanism based on real-time price changes, thereby obtaining fluctuation period strategy data; Step S344: formulating a strategy for minimizing electricity consumption for the time period displayed as a high-price risk zone by the price range classification data, thereby obtaining high-price time period strategy data; Step S345: integrating the safe period strategy data, possible period strategy data, fluctuating period strategy data, and high-price period strategy data based on time integrity, thereby obtaining a differentiated energy optimization strategy for different time periods; Step S346: Analyze the historical electricity consumption patterns of energy-consuming enterprises, identify adjustable and non-adjustable loads, evaluate the adjustment flexibility and response speed of different types of loads, calculate the maximum adjustable load in each period, and obtain load adjustment potential data; Step S347: Based on the load adjustment potential data, the time-divided differentiated energy optimization strategy is subjected to cross-time load distribution based on production continuity constraints, thereby obtaining energy optimization strategy data.

7. The method for recommending electricity consumption strategies based on scientific energy consumption and settlement according to claim 6 is characterized in that: Step S4 includes the following steps: Step S41: Real-time monitoring of the power supply and demand relationship of the energy optimization strategy data to obtain real-time power supply and demand data; Step S42: Calculate the real-time deviation between the real-time electricity price and the predicted electricity price based on the real-time electricity supply and demand data and the electricity spot supply and demand forecast data, and evaluate the recent forecast accuracy trend through the sliding window technology to obtain the deviation trend characteristic data; Step S43: performing deviation pattern analysis based on time characteristics on the deviation trend characteristic data, thereby obtaining predicted deviation real-time data; Step S44: judging whether the actual electricity price deviates significantly from the forecast based on the confidence interval of the electricity spot supply and demand forecast data, quantifying the degree of deviation, marking the time periods of continuous deviation, and obtaining forecast deviation interval data; Step S45: triggering an adjustment mechanism for the energy usage optimization strategy data according to the predicted deviation interval data to obtain energy usage guidance update data.

8. The method for recommending electricity consumption strategies based on scientific energy consumption and settlement according to claim 7 is characterized in that: Step S45 includes the following steps: Step S451: identifying the time interval and degree of deviation of the current electricity price from the forecast according to the forecast deviation interval data, and classifying based on the degree of deviation, thereby obtaining deviation degree classification data; Step S452: different policy adjustment triggering thresholds are set according to the deviation degree classification data, and a cumulative triggering mechanism for continuous deviation periods is added to obtain policy adjustment triggering condition data; Step S453: Evaluate the benefits of each price range on the energy optimization strategy data, thereby obtaining strategy evaluation report data; Step S454: Estimating the adjustment coefficient of each price range according to the deviation degree classification data and the strategy evaluation report data, thereby obtaining dynamic adjustment coefficient data; Step S455: triggering an adjustment mechanism for the energy usage optimization strategy data according to the dynamic adjustment coefficient data and the strategy adjustment trigger condition data, thereby obtaining energy usage guidance update data.

9. The method for recommending electricity consumption strategies based on scientific energy consumption and settlement according to claim 8 is characterized in that: Step S5 includes the following steps: Step S51: Calculate the actual power consumption data in each time period according to the pre-acquired actual power consumption behavior data; calculate the power consumption of the energy-consuming enterprise in response to the suggestions in the optimized energy consumption guide according to the energy consumption guide update data, so as to obtain the recommended power consumption data; Step S52: evaluating the response degree of the actual power consumption data and the recommended power consumption data, and calculating the real power consumption cost, so as to obtain the actual response benefit data; Step S53: comparing the estimated response benefit with the actual response benefit according to the actual response benefit data, thereby obtaining response benefit difference data; Step S54: adjusting the strategy optimization parameters based on the response revenue difference data to achieve adaptive optimization of the power consumption strategy.

10. A power consumption strategy recommendation system based on scientific energy consumption and settlement, characterized in that: Used to execute the method for recommending electricity consumption strategies based on scientific energy consumption and settlement by plate as claimed in claim 1, the electricity consumption strategy recommendation system based on scientific energy consumption and settlement by plate comprises: The data collection module is used to obtain climate data; extract the power grid operation status from the announcements, reports and APIs issued by the power trading center to obtain the power grid operation status data; obtain historical energy price data, and perform periodic analysis on the historical energy price data based on time series to obtain primary energy price data; The power supply and demand forecasting module is used to select and extract key features based on the changes in power supply and demand relations according to climate data, power grid operation status data, and primary energy price data, thereby obtaining a power supply and demand feature set; the power supply and demand feature set is used to forecast power spot supply and demand using the gradient boosting tree algorithm, thereby obtaining the intraday power spot supply and demand forecast data and its confidence interval; The energy optimization strategy formulation module is used to formulate dynamic energy optimization strategies based on the electricity spot supply and demand forecast data and its confidence interval, so as to obtain energy optimization strategy data. The dynamic energy optimization strategy formulation specifically divides the electricity consumption period into safe low-price area, possible low-price area, price fluctuation area and high-price risk area, and formulates differentiated electricity consumption suggestions for each area; The energy consumption guidance optimization module is used to monitor the power supply and demand relationship of the energy consumption optimization strategy data in real time, and compare the actual electricity price with the power spot supply and demand forecast data and its confidence interval to obtain the forecast deviation interval data; according to the forecast deviation interval data, the adjustment mechanism of the energy consumption optimization strategy data is triggered to obtain the energy consumption guidance update data; The benefit evaluation and adaptive optimization module is used to calculate the benefits generated by energy-consuming enterprises' responses to energy consumption optimization guidelines based on the updated data of energy consumption guidelines, and to compare the estimated response benefits with the actual response benefits to obtain the response benefit difference data; based on the response benefit difference data, the strategy optimization parameters are adjusted to achieve adaptive optimization of the electricity consumption strategy.

Images