Electricity selling control system and method for classifying medium and long term load prediction

Abstract

The invention provides an electricity selling control system and method for classifying medium and long term load prediction. The system comprises a data acquisition module, a data processing module, a prediction algorithm module, an electricity selling control module and a display interaction module. The data acquisition module is used for collecting power consumption data of a user; the data processing module is used for performing data preprocessing on the power utilization data and classifying and dividing the power utilization data; the prediction algorithm module is used for selecting a statistical algorithm model or a machine learning algorithm model according to a preset data volume threshold value, and predicting the power consumption of the user based on the classified and divided power consumption data; the electricity selling control module is used for forming electricity selling strategy result data; and the display interaction module is used for man-machine interaction and power utilization data information display. According to the method, relatively accurate power consumption prediction can be carried out for users of different power consumption types and grades, power consumption characteristics and change trends of different user groups can be captured, the accuracy of medium and long term load prediction is improved, and the user experience is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power consumption control, and in particular relates to a power sales control system and method for type-based medium- and long-term load forecasting. Background Art

[0002] User load forecasting refers to learning from existing user electricity consumption data to predict user electricity consumption behavior in the future. In recent years, with the continuous deepening of power market reform and the continuous opening of power sales business, user load forecasting has become an important link for power sales companies to reduce deviation assessment and enhance profitability. For power sales companies, it is beneficial for them to decide their own medium- and long-term contract signing strategies, spot market quotations, trading strategies, and economic calculations of individual users. Therefore, load forecasting has become one of the important contents of the modernization of power system control and the transformation of power sales companies to spot marketization.

[0003] At present, the structure and operation mode of the power sales side of the power market have become more complex with the opening of the power market. The competition in the market has become very fierce, and a situation of competition and cooperation between multiple sellers and multiple buyers has been formed. In this case, the user's power consumption behavior, load curve, response law and intensity have changed greatly compared with the traditional power market. The power sales side needs to explore and study the pattern recognition technology and cluster analysis method that match the analysis of user power consumption behavior, and on this basis, establish a mathematical model of user power consumption behavior, master the key power consumption characteristics of users, and make refined load forecasts according to different power consumption scenarios and habits of users. Based on the refined load forecast, provide users with reference suggestions for energy-saving transformation and personalized, scientific intelligent power consumption strategies; formulate personalized and competitive power sales plans for themselves.

[0004] There are many classification criteria for load forecasting. The forecast time period is a common one. Electricity load forecasting can be divided into ultra-short-term load forecasting, short-term load forecasting, medium-term load forecasting and long-term load forecasting. The ultra-short-term load forecasting time ranges from a few seconds or minutes to a few hours, the short-term load forecasting time ranges from hours to weeks, and the medium-term load forecasting and long-term load forecasting time range from months to years. Among them, ultra-short-term load forecasting and short-term load forecasting can be classified as short-term forecasting, and medium-term load forecasting and long-term load forecasting are collectively referred to as medium- and long-term load forecasting. At present, due to the short cycle of short-term forecasting, it can be easily controlled and adjusted in system control. However, for medium- and long-term load forecasting, due to the long cycle of electricity consumption and the increase in uncontrollable factors, the reliability and stability of the current power sales control system still have a lot of room for improvement, and the user experience needs to be further optimized to achieve the rationality of power sales control. Summary of the invention

[0005] The purpose of the present invention is to provide a power sales control system and method for medium- and long-term load forecasting by type, so as to solve the problem that the reliability and stability of the power sales control system for medium- and long-term load forecasting still have a lot of room for improvement, and the user experience needs to be further optimized.

[0006] In order to achieve the above object, the present invention adopts the following technical scheme:

[0007] The present invention provides a power selling control system for medium- and long-term load forecasting by type, comprising a data acquisition module, a data processing module, a forecasting algorithm module, a power selling control module and a display interaction module; wherein:

[0008] The data acquisition module is used to collect the user's electricity consumption data and transmit the electricity consumption data to the data processing module;

[0009] The data processing module is used to pre-process the power consumption data, and classify the data and divide the power consumption level into different levels by combining the historical data stored in the system and the pre-processed power consumption data, so as to obtain the classified and divided power consumption data;

[0010] The prediction algorithm module is used to select a statistical algorithm model or a machine learning algorithm model according to a preset data volume threshold, and predict the user's power consumption based on the classified and divided power consumption data to obtain prediction result data;

[0011] The power selling control module is used to form power selling strategy result data based on the prediction result data;

[0012] The display interaction module is used for human-computer interaction and displaying electricity consumption data information.

[0013] In some implementations, the electricity consumption data includes peak electricity consumption data, peak electricity consumption data, valley electricity consumption data, fixed price electricity consumption data, and total retail electricity consumption data.

[0014] In some implementations, data preprocessing is used to remove abnormal data, specifically using the following formulas (1) and (2):

[0015] H ni =H nj η ij iF ni =φ&H nj ≠φ(1);

[0016] G K =H ni ifi∈K&H ni >0(2);

[0017] Among them: K is five different electricity consumption levels, i is 1 to 12 different months, G Kis the historical data set of all years belonging to the Kth electricity consumption level, H ni is the historical data set for the ith month in the nth year, H nj is the historical data set of the jth month in the nth year, η ij is the ratio of electricity consumption in month i to that in month j.

[0018] In some implementations, the human-computer interaction includes user electricity settlement.

[0019] In some embodiments, the statistical algorithm model calculates the average daily electricity consumption on weekdays and holidays, predicts the total electricity consumption on corresponding weekdays and holidays respectively, and adds the two together to obtain the final electricity consumption prediction result.

[0020] In some embodiments, the statistical algorithm model uses a seasonal ARIMA model, and is specifically predicted according to the following formulas (3), (4), and (5):

[0021]

[0022]

[0023]

[0024] Where: N is the total number of historical data in the Kth electricity consumption level, is the average electricity consumption during holidays in the Kth electricity consumption level, is the average power consumption on weekdays in the Kth power consumption level, G Kl is the electricity consumption of the month corresponding to the lth data in the Kth electricity consumption level, D Klw is the number of working days in the month corresponding to the lth data in the Kth electricity consumption level, D Klh is the number of holidays in the month corresponding to the lth data in the Kth electricity consumption level, is the predicted value of electricity consumption in the forecast month, D w is the number of working days corresponding to the predicted month, D h The corresponding number of holidays in the predicted month.

[0025] In some embodiments, the machine learning algorithm model automatically learns the data characteristics of time series, increase and decrease trends, and seasonal changes based on electricity consumption data, and then obtains electricity consumption prediction results.

[0026] In some implementations, the machine learning algorithm model uses a long short-term memory network model, and prediction is performed specifically according to the following formula (7):

[0027]

[0028] Where: M is the total number of users in the i-th month of the m-th year, H′ mi is the total electricity consumption of all users in the i-th month of the m-th year, H mij is the electricity consumption of the jth user in the i-th month of the m-th year.

[0029] In some implementations, the prediction result data includes a monthly total electricity consumption forecast for all users and a monthly total electricity consumption forecast for a single user.

[0030] The present invention also provides a method for controlling electricity sales based on medium- and long-term load forecasting by type, which is specifically performed according to the following steps:

[0031] S1. The data acquisition module collects the user's electricity consumption data and transmits the electricity consumption data to the data processing module;

[0032] S2, the data processing module performs data preprocessing on the electricity consumption data, and combines the historical data and the preprocessed electricity consumption data to perform data classification and electricity consumption grade division;

[0033] S3, the prediction algorithm module predicts the user's electricity consumption through the classified and divided electricity consumption data to obtain the prediction result data;

[0034] S4. The power selling control module forms power selling strategy result data based on the prediction result data.

[0035] Compared with the prior art, the power sales control system and method for medium- and long-term load forecasting by type in the present invention have the following beneficial technical effects.

[0036] The present invention provides a type-based medium- and long-term load forecasting power sales control system, including a data acquisition module, a data processing module, a prediction algorithm module, a power sales control module, and a display interaction module; wherein: the data acquisition module is used to collect the user's power consumption data and transmit the power consumption data to the data processing module; the data processing module is used to pre-process the power consumption data, and classify and classify the data in combination with the historical data stored in the system and the pre-processed power consumption data to obtain the classified and classified power consumption data; the prediction algorithm module is used to select a statistical algorithm model or a machine learning algorithm model according to a preset data volume threshold, and predict the user's power consumption based on the classified and classified power consumption data to obtain the prediction result data; the power sales control module is used to form the power sales strategy result data based on the prediction result data; the display interaction module is used for human-computer interaction and display of power consumption data information. Based on the above, the present invention collects comprehensive user power consumption data through the data acquisition module, and performs fine data pre-processing and classification through the data processing module, so as to make more accurate power consumption predictions for user groups of different types and different power consumption levels. Compared with the common unified prediction model, the system of the present invention can better capture the electricity consumption characteristics and change trends of different user groups, thereby improving the accuracy of medium- and long-term load forecasting. The present invention provides strong data support for the electricity sales control module based on the accurate classification prediction results through the prediction algorithm module. The electricity sales control module can flexibly adjust the electricity sales strategy according to the prediction result data. The present invention provides a human-computer interaction interface through the display interaction module, so that users can clearly understand their own electricity consumption, prediction results, electricity sales strategy and other information, enhance the user experience of power services, and has certain practicality. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The drawings in the specification are used to provide further understanding of the present invention and constitute a part of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0038] Figure 1 A schematic diagram of the model structure of the system in the power sales control system and method for medium- and long-term load forecasting by type according to the present invention;

[0039] Figure 2 It is a schematic diagram of the system application scenario and architecture of the power sales control system and method for medium- and long-term load forecasting by type in the present invention;

[0040] Figure 3 It is a schematic diagram of the model structure of the settlement management in the power sales control system and method for medium- and long-term load forecasting by type according to the present invention;

[0041] Figure 4 A schematic diagram of the highest and lowest typical load curves for each day of the year in an embodiment of the power sales control system and method for medium- and long-term load forecasting by type of the present invention;

[0042] Figure 5 A schematic diagram of a typical load curve on a working day in an embodiment of a power sales control system and method for medium- and long-term load forecasting by type according to the present invention;

[0043] Figure 6 It is a schematic diagram of typical load curves on holidays in an embodiment of a power sales control system and method for medium- and long-term load forecasting by type according to the present invention. DETAILED DESCRIPTION

[0044] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only 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 ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0045] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0046] How to improve the effectiveness and accuracy of medium- and long-term load forecasts of power sales systems, how to improve the reliability of the entire power sales system, and improve user experience.

[0047] Based on this, Figure 1 As shown, the present invention provides a power selling control system for medium- and long-term load forecasting by type, including a data acquisition module, a data processing module, a prediction algorithm module, a power selling control module and a display interaction module; wherein:

[0048] The data acquisition module is used to collect the user's electricity consumption data and transmit the electricity consumption data to the data processing module;

[0049] The data processing module is used to pre-process the power consumption data, and classify the data and divide the power consumption level into different levels by combining the historical data stored in the system and the pre-processed power consumption data, so as to obtain the classified and divided power consumption data;

[0050] The prediction algorithm module is used to select a statistical algorithm model or a machine learning algorithm model according to a preset data volume threshold, and predict the user's power consumption based on the classified and divided power consumption data to obtain prediction result data;

[0051] The power selling control module is used to form power selling strategy result data based on the prediction result data;

[0052] The display interaction module is used for human-computer interaction and displaying electricity consumption data information.

[0053] Based on the above, the power sales control system of the present invention can take into account the power consumption patterns on weekdays and holidays as well as the power consumption patterns in different months of the year, and then accurately carry out medium- and long-term load forecasting. The application of this system can improve the accuracy of power consumption forecasting, optimize power sales strategies, enhance user experience, promote the rational allocation of resources, help power companies plan power production and transmission in advance, rationally allocate power grid resources, and reduce power waste and supply-demand imbalance.

[0054] Specifically, the demand for medium- and long-term load forecasting mainly includes two parts: the forecast of the total monthly electricity consumption of all users and the forecast of the total monthly electricity consumption of a single user. For each part, it also includes the forecast of five small parts: peak electricity, peak electricity, valley electricity, one-price electricity and total retail electricity. The power sales control system of the present invention adopts a type-based forecasting method for the forecast of the peak electricity, peak electricity, valley electricity, one-price electricity and total retail electricity of each user. According to the size of historical data and the difference in data basis, the present invention proposes two different electricity consumption forecasting algorithm models. The first is a statistical algorithm model, which is suitable for the situation of insufficient data. The algorithm takes into account the electricity consumption rules of working days and holidays and the electricity consumption rules of different months throughout the year; the second is a machine learning algorithm model, which is suitable for the situation of good data basis and sufficient data volume. The machine learning algorithm model will automatically learn to consider the data characteristics of different aspects such as time series, increase and decrease trends, and seasonal changes. The system of the present invention presets the threshold of the data volume, and then the system automatically determines which model to use for calculation and processing.

[0055] The statistical algorithm used in the present invention mainly uses each user's historical electricity consumption data to classify electricity consumption and data, and then calculates the average daily electricity consumption on weekdays and holidays. The implementation logic of the statistical algorithm is divided into three parts: electricity consumption classification, data statistical processing, and data prediction.

[0056] As a specific embodiment, the statistical algorithm model used in the present invention may be, but not limited to, a seasonal ARIMA model. The machine learning algorithm model used in the present invention may be, but not limited to, a long short-term memory network model. The present invention uses the above-mentioned model algorithm to statistically predict the total electricity consumption on the corresponding working days of the month and the total electricity consumption on holidays based on the mean, and finally adds these two parts to obtain the final electricity consumption prediction result.

[0057] Based on the aforementioned power sales control system, the present invention proposes a power sales control method for type-based medium- and long-term load forecasting, which includes the following steps:

[0058] S1. The data acquisition module collects the user's electricity consumption data and transmits the electricity consumption data to the data processing module;

[0059] S2, the data processing module performs data preprocessing on the electricity consumption data, and combines the historical data and the preprocessed electricity consumption data to perform data classification and electricity consumption grade division;

[0060] S3, the prediction algorithm module predicts the user's electricity consumption through the classified and divided electricity consumption data to obtain the prediction result data;

[0061] S4. The power selling control module forms power selling strategy result data based on the prediction result data.

[0062] Through medium- and long-term load forecasting, the present invention can provide the power system with forecast data on electricity demand in the future, which helps the power system to plan and dispatch power production and transmission in advance, and reduce resource waste or shortage problems caused by mismatch between supply and demand. Accurate forecasting and differentiated power sales strategies can help power companies better balance power supply and demand. It effectively improves the rationality of power sales and further enhances user experience.

[0063] like Figure 2As shown, in the actual application scenario and architecture, the electricity sales system of the present invention is implemented through the constructed Web site and management background. The electricity sales system accesses the system through the website and the background, in which customer information, contract information, electricity reminder, policy query, electricity forecast, etc. are designed as independent microservice structures to form a microservice department. The microservice accesses the business database and is finally aggregated to the data warehouse. The management background is responsible for handling key tasks such as user interaction, order management and data storage. Below the management background, there is a microservice layer for providing more fine-grained services; customer information query, contract query service, electricity reminder service, policy query service and electricity forecast service. The above services directly or indirectly rely on the data source of customer information data, and make policy decisions through the policy database. In addition, the policy database itself also relies on the Oracle database to store policy data. Finally, through the module of the big data platform, it is used to process the data from all the above components and convert it into a form that can be used for analysis and decision-making.

[0064] like Figure 3 As shown, the present invention adds a settlement management part to the part of the interactive module in the system, which is used to set the settlement information between the power sales company and the customer, mainly including the design and setting of the power sales system for the profit sharing information. The profit sharing information is mainly that the power sales company participates in the power transaction on behalf of the enterprise or extracts a part of the profit from the electricity bill saved for the customer according to the long-term power purchase contract signed with the power plant, so as to improve the rights and interests of the customer.

[0065] The following is a detailed description of a power sales control system and method for medium- and long-term load forecasting by type according to a specific embodiment of the present invention.

[0066] like Figure 4 As shown, according to the typical curve of the highest and lowest daily loads in a certain area throughout the year (such as Figure 1 ) divides each user's annual electricity consumption into months with no electricity consumption levels. It is generally believed that July and August are the first level, December and January are the second level, June and September are the third level, March, April, May, October, November are the fourth level, and February is the fifth level.

[0067] In order to ensure that the data of all months in the historical data are not empty, the electricity consumption ratio between different months in the typical curve of the highest and lowest daily loads throughout the year can be statistically analyzed; then, the electricity consumption data of the corresponding month with empty data is indirectly calculated using the data of the known month and the ratio between the month and the month with empty data, so as to ensure that the data in all electricity consumption levels are not empty. The algorithm model is shown in formula (1). According to the electricity consumption level divided in the first step, all historical electricity consumption data of all users in each electricity consumption level are statistically analyzed and the abnormal data with negative electricity values ​​are eliminated. The specific algorithm model is shown in formula (2). And (2):

[0068] H ni =H nj η ij iF ni =φ&H nj ≠φ(1);

[0069] G K =H ni ifi∈K&H ni >0(2);

[0070] Among them: K is five different electricity consumption levels, i is 1 to 12 different months, G K is the historical data set of all years belonging to the Kth electricity consumption level, H ni is the historical data set for the ith month in the nth year, H nj is the historical data set of the jth month in the nth year, η ij is the ratio of electricity consumption in month i to that in month j.

[0071] For large industrial and general commercial users, if the user has signed a contract and has at least one month of historical electricity consumption data, the algorithm will treat it as an ordinary user and predict its electricity consumption in the next few months. The specific prediction method is as follows.

[0072] like Figure 5 and Figure 6 As shown, the ratio δ of the total daily electricity consumption between working days and holidays is calculated based on the typical load curve of working days and the typical load curve of holidays. According to the month to be predicted and the classification result of the first step, the electricity consumption level corresponding to the month is matched, and the ratio δ is used to respectively count the electricity consumption data of working days and holidays corresponding to all months in the level; then, the average daily electricity consumption of holidays and the average daily electricity consumption of working days are calculated respectively, and the specific algorithm model is shown in formula (3) and formula (4); finally, according to the month to be predicted, the number of days of holidays and working days corresponding to the predicted month is respectively counted, and multiplied by the statistical average daily electricity consumption of working days and the average daily electricity consumption of holidays and then added, so as to obtain the total electricity consumption data of the predicted month, and the specific algorithm model is shown in formula (5):

[0073]

[0074]

[0075]

[0076] Where: N is the total number of historical data in the Kth electricity consumption level, is the average electricity consumption during holidays in the Kth electricity consumption level, is the average power consumption on weekdays in the Kth power consumption level, G Kl is the electricity consumption of the month corresponding to the lth data in the Kth electricity consumption level, D Klw is the number of working days in the month corresponding to the lth data in the Kth electricity consumption level, D Klh is the number of holidays in the month corresponding to the lth data in the Kth electricity consumption level, is the predicted value of electricity consumption in the forecast month, D w is the number of working days corresponding to the predicted month, D h The corresponding number of holidays in the predicted month.

[0077] For new users who sign contracts for the first time, the algorithm regards them as newly added users. For such users, the power consumption algorithm will uniformly set their predicted results to zero. The specific algorithm model is shown in formula (6). The specific predicted results need to be filled in manually later.

[0078] P m =0(6);

[0079] For the prediction of the total electricity consumption of all users, firstly, the total electricity consumption data corresponding to each month in the historical data is obtained, and the algorithm is as shown in formula (7), and then the data is predicted as a common user.

[0080]

[0081] Where M is the total number of users in the i-th month of the m-th year; H′ mi is the total electricity consumption of all users in the i-th month of the m-th year; H mij is the electricity consumption of the jth user in the i-th month of the m-th year.

[0082] In the electricity sales control system of the present invention for medium- and long-term load forecasting by type, the default forecast result for new users is zero, which can be optimized to use historical data of users of the same type for forecasting in the later stage. For the forecast of the total power consumption of all users, the total power consumption data corresponding to each month in the historical data can be obtained and used as a common user for data forecasting.

[0083] To sum up, the present invention makes separate predictions for the electricity consumption data of each user. The prediction principle is simple, easy to implement, short in prediction time, low in data demand, and small in prediction deviation for stable users. It can improve the overall operating accuracy of the electricity sales system, improve the convenience of the electricity consumption system, and improve the user experience, and has good applicability.

[0084] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Any ordinary technician in the industry can smoothly implement the present invention as shown in the specification and described above. However, any equivalent changes, modifications and evolutions made by technicians familiar with the profession without departing from the scope of the technical solution of the present invention using the technical content disclosed above are all equivalent embodiments of the present invention. At the same time, any equivalent changes, modifications and evolutions made to the above embodiments based on the essential technology of the present invention are still within the protection scope of the technical solution of the present invention.

Claims

1. A power sales control system for medium- and long-term load forecasting by type, characterized in that: It includes data acquisition module, data processing module, prediction algorithm module, power sales control module and display interaction module; among which: The data acquisition module is used to collect the user's electricity consumption data and transmit the electricity consumption data to the data processing module; The data processing module is used to pre-process the power consumption data, and classify the data and divide the power consumption level into different levels by combining the historical data stored in the system and the pre-processed power consumption data, so as to obtain the classified and divided power consumption data; The prediction algorithm module is used to select a statistical algorithm model or a machine learning algorithm model according to a preset data volume threshold, and predict the user's power consumption based on the classified and divided power consumption data to obtain prediction result data; The power selling control module is used to form power selling strategy result data based on the prediction result data; The display interaction module is used for human-computer interaction and displaying electricity consumption data information.

2. The electricity sales control system for medium- and long-term load forecasting by type according to claim 1 is characterized in that: Electricity consumption data includes peak electricity data, peak electricity data, valley electricity data, fixed price electricity data and total retail electricity data.

3. The electricity sales control system for medium- and long-term load forecasting by type according to claim 1 is characterized in that: Data preprocessing is used to remove abnormal data, and the following formulas (1) and (2) are used: H ni =H nj η ij ifH ni =φ&H nj ≠φ(1); G K =H ni ifi∈K&H ni >0(2); Among them: K is five different electricity consumption levels, i is 1 to 12 different months, G K is the historical data set of all years belonging to the Kth electricity consumption level, H ni is the historical data set for the ith month in the nth year, H nj is the historical data set of the jth month in the nth year, η ij is the ratio of electricity consumption in month i to that in month j.

4. The electricity sales control system for medium- and long-term load forecasting by type according to claim 1 is characterized in that: Human-computer interaction includes user electricity settlement.

5. The electricity sales control system for medium- and long-term load forecasting by type according to claim 1 is characterized in that: The statistical algorithm model calculates the average daily electricity consumption on weekdays and holidays, predicts the total electricity consumption on the corresponding weekdays and holidays in the month, and adds the two together to obtain the final electricity consumption prediction result.

6. The electricity sales control system for medium- and long-term load forecasting by type according to claim 1 is characterized in that: The statistical algorithm model uses the seasonal ARIMA model and makes predictions according to the following formulas (3), (4), and (5): Where: N is the total number of historical data in the Kth electricity consumption level, is the average electricity consumption during holidays in the Kth electricity consumption level, is the average power consumption on weekdays in the Kth power consumption level, G Kl is the electricity consumption of the month corresponding to the lth data in the Kth electricity consumption level, D Klw is the number of working days in the month corresponding to the lth data in the Kth electricity consumption level, D Klh is the number of holidays in the month corresponding to the lth data in the Kth electricity consumption level, is the predicted value of electricity consumption in the forecast month, D w is the number of working days corresponding to the predicted month, D h The corresponding number of holidays in the predicted month.

7. The electricity sales control system for medium- and long-term load forecasting by type according to claim 1 is characterized in that: The machine learning algorithm model automatically learns the data characteristics of time series, increase and decrease trends, and seasonal change patterns based on electricity consumption data, and then obtains electricity consumption prediction results.

8. The electricity sales control system for medium- and long-term load forecasting by type according to claim 1 is characterized in that: The machine learning algorithm model uses a long short-term memory network model, and the prediction is made according to the following formula (7): Where: M is the total number of users in the i-th month of the m-th year, H′ mi is the total electricity consumption of all users in the i-th month of the m-th year, H mij is the electricity consumption of the jth user in the i-th month of the m-th year.

9. The electricity sales control system for medium- and long-term load forecasting by type according to claim 1 is characterized in that: The forecast result data includes the monthly total electricity consumption forecast for all users and the monthly total electricity consumption forecast for a single user.

10. A control method for a power sales control system for medium- and long-term load forecasting by type as described in any one of claims 1 to 9, characterized in that: The steps include: S1. The data acquisition module collects the user's electricity consumption data and transmits the electricity consumption data to the data processing module; S2, the data processing module performs data preprocessing on the electricity consumption data, and combines the historical data and the preprocessed electricity consumption data to perform data classification and electricity consumption grade division; S3, the prediction algorithm module predicts the user's electricity consumption through the classified and divided electricity consumption data to obtain the prediction result data; S4. The power selling control module forms power selling strategy result data based on the prediction result data.

Images