Power optimization scheduling method and system for virtual power plant

Abstract

The invention relates to the technical field of power dispatching, in particular to a power optimization dispatching method and system for a virtual power plant. The method comprises the following steps: acquiring electrical load data of a user in each day by using a virtual power plant; determining the fluctuation trend of the electrical load data at each moment in each day; determining the similarity between the electrical load data of each day and the electrical load data of other days; determining a target sequence and regularity of each day according to the similarity; predicting the average load of the day to be predicted; and according to the average load, regulating and controlling start and stop of a generator set, and completing power optimization scheduling of the virtual power plant. By considering the fluctuation trend of the power consumption load data of the user in each day and combining similarity and regularity analysis, the power consumption load of the target day can be predicted more accurately, so that the load prediction is closer to the actual power consumption behavior, and errors caused by the fact that fluctuation factors are not fully considered in a traditional prediction method are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of power dispatching, and in particular to a method and system for optimizing power dispatching of a virtual power plant. Background Art

[0002] With the continuous increase in global energy demand and the increase in the proportion of renewable energy, traditional power systems are facing more and more challenges, especially in terms of imbalance in power supply and demand, peak-valley compound differences, energy waste, environmental pollution, etc. The limitations of traditional power grids in scheduling optimization are becoming increasingly apparent. Therefore, Virtual Power Plant (VPP), as an emerging method for optimizing power system scheduling, has received much attention and research.

[0003] The patent document with announcement number CN108037667B discloses a base station power optimization scheduling method based on a virtual power plant, including the following steps: (1) firstly establishing an interaction model between a microgrid containing a base station and a virtual power plant and a traditional large power grid; (2) establishing an optimization scheduling model with the purpose of maximizing the scheduling profit of the virtual power plant containing the base station; (3) predicting wind and solar power with the purpose of minimizing the abandoned energy cost caused by the uncertainty of clean energy and establishing a model; (4) using a robust optimization method to handle the output uncertainty of clean energy for constraints and setting prediction coefficients and robust coefficients; (5) finally solving the optimal scheduling result using computer software.

[0004] However, due to the fluctuations in user electricity consumption behavior, when users are active during monitoring of their electricity consumption behavior, the electricity load will fluctuate. Traditional prediction methods use the same prediction weight for all samples, which will lead to inaccurate prediction results, thereby affecting the power dispatch results and causing grid overload or energy waste.

[0005] However, due to certain fluctuations in user electricity consumption behavior, especially during peak periods and special events (such as seasonal changes, holidays or emergencies), traditional forecasting methods often cannot fully consider these complex factors, resulting in increased errors in power load forecasting; when users are active, especially concentrated in certain specific time periods, load fluctuations are often more drastic, which makes it impossible for simple forecasting models based on historical data to accurately capture the dynamic changes in electricity demand; traditional forecasting methods usually assume that all samples have the same prediction weight, but this assumption does not take into account the differences in electricity consumption behavior in each time period or user group. For example, the electricity consumption patterns of some user groups (such as industrial users or commercial users) are quite different from those of household users, and their electricity demand may be affected by factors such as production plans, working hours or weather, which are not fully reflected in traditional forecasting models; therefore, traditional forecasting methods will show greater inaccuracy when dealing with high volatility or abnormal electricity consumption behavior, thereby affecting the optimization effect of power dispatch, and further leading to grid load overload or energy waste. Summary of the invention

[0006] In order to solve the problem that due to fluctuations in users' electricity consumption behavior, traditional prediction methods often cannot fully consider these complex factors, thereby increasing the error of power load prediction, affecting the optimization effect of power dispatch, and further leading to grid overload or energy waste, the present invention provides a power optimization dispatching method and system for a virtual power plant.

[0007] In a first aspect, the present invention provides a method for optimizing power dispatching of a virtual power plant, which adopts the following technical solution: A power optimization scheduling method for a virtual power plant, comprising: obtaining the daily electricity load data of users by using the virtual power plant; recording the electricity load data at the target moment within the target day of each day of the user as the target electricity load data, taking the target electricity load data as the center of a preset fluctuation window, and determining the fluctuation trend of the electricity load data at the target moment within the target day according to the difference between the target electricity load data and the electricity load data at the remaining moments in the fluctuation window; the normalization result of the correlation coefficients of the fluctuation trends of the electricity load data at all corresponding moments of the target day and the remaining days is used as the similarity of the electricity load data between the target day and the remaining days; according to the magnitude of the similarity, determining a target sequence including multiple dates for the target day, obtaining the day difference values between all adjacent two days in the target sequence, and determining the regularity of the target day according to the day difference values and the similarity of the electricity load data between the target day and the remaining days in the target sequence; obtaining the regularity and the average electricity load data of each day within a preset prediction window, predicting the average load of the day to be predicted according to the regularity and the average electricity load data; and regulating the start and stop of the generator set according to the average load to complete the power optimization scheduling of the virtual power plant.

[0008] The beneficial effects are as follows: By considering the fluctuation trend of the daily electricity load data of users and combining similarity and regularity analysis, the electricity load of the target day can be predicted more accurately, making the load prediction closer to the actual electricity consumption behavior and reducing the errors caused by the failure to fully consider the fluctuation factors in the traditional prediction method; Based on the analysis of the similarity of the load data between the target day and other dates, the system can identify similar electricity load patterns and further improve the prediction accuracy through regularity analysis, which helps to ensure the stability and efficiency of power supply; By accurately predicting the electricity load, more scientific and reasonable power scheduling can be achieved. By regulating the start and stop of the generator set, unnecessary power generation can be reduced, the energy use efficiency can be optimized, thereby avoiding the risk of load overloading in the power grid and reducing energy waste; By dynamically obtaining the daily load data, fluctuation trend and regularity analysis, the scheduling strategy of the virtual power plant can be adjusted in time to adapt to the changes in electricity demand in different time periods and seasons, and improve the overall response ability of the power system.

[0009] Further, the fluctuation trend satisfies: ; where is the fluctuation trend of the electricity load data at the th moment within the th day, is the number of moments within the preset fluctuation window, is the electricity load data at the th moment within the th day, is the the th moment within the preset fluctuation window corresponding to the th moment, and the electricity consumption load data at this moment, is the standard normalization function, and

[0010] The beneficial effect is that it can accurately identify the instantaneous fluctuations in the electricity consumption load. By comparing the differences between the current moment and the data of other moments within the fluctuation window, the fluctuation intensity of the electricity consumption load can be judged. Through normalization, the fluctuation trends can be compared on the same scale, avoiding the deviation caused by different data scales and making the analysis of the fluctuation trend more accurate.

[0011] Further, the correlation coefficient is the Pearson correlation coefficient.

[0012] Further, the normalization method is extreme value normalization.

[0013] Further, determining the target sequence of the target day including multiple dates according to the magnitude of the similarity includes: sorting the remaining days corresponding to the target day with a similarity greater than the preset similarity threshold according to the date to obtain the target sequence.

[0014] The beneficial effect is that by sorting according to the similarity between the target day and other days, other days with a strong correlation with the target day can be more accurately identified, providing a more accurate prediction or judgment. By setting the similarity threshold and sorting, the interference of irrelevant or low-similarity data is reduced, thereby improving the efficiency of data processing. When dealing with a large amount of historical data, valuable parts can be more quickly screened out, avoiding unnecessary calculations and redundant information.

[0015] Further, the regularity satisfies: ; where is the regularity of the th day, is the total number of days in the target sequence of the th day, is the th difference value of the number of days in the target sequence of the th day, is the average value of all the difference values of the number of days in the target sequence of the th day, is the th day and the th day, and is the average value of the similarities of the electricity consumption load data of the remaining days in the target sequences of these two days, is the natural exponential function, is the standard normalization function.

[0016] The beneficial effects are as follows: By integrating multiple factors such as the variance of difference values, mean, similarity, and normalization, the regularity can more accurately reflect the strength of the regularity of a target sequence; through the design of the exponential function, the model can more sensitively identify data with low regularity or large fluctuations.

[0017] Further, the average load satisfies: ; where The average load of the day to be predicted , is the regularity of the th day within the prediction window corresponding to the day to be predicted , is the number of days within the preset prediction window, is the ordinal number of the day of the th day within the prediction window corresponding to the day to be predicted , is the ordinal number of the day of the day to be predicted , is the average power consumption load data of the th day within the prediction window corresponding to the day to be predicted , is the natural exponential function.

[0018] The beneficial effects are as follows: By utilizing the regularity of each day in the historical load data and the time series relationship, a more accurate prediction of the load of the day to be predicted can be made; through the weighting of the natural exponential function, the time difference between the historical load data and the day to be predicted can be considered, ensuring that the historical data closer to the current time difference has a greater impact on the prediction result, improving the timeliness; by using the regularity parameter, the impact of periodic or seasonal changes on the load can be captured, thereby improving the accuracy of the prediction.

[0019] In a second aspect, the present invention provides a power optimization scheduling system for a virtual power plant, adopting the following technical solution: A power optimization scheduling system for a virtual power plant includes: a processor and a memory, where the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned power optimization scheduling method for a virtual power plant is implemented.

[0020] By adopting the above technical solution, the above-mentioned power optimization scheduling method for a virtual power plant is generated into a computer program and stored in the memory to be loaded and executed by the processor, so as to manufacture a terminal device according to the memory and the processor, which is convenient to use.

[0021] The present invention has the following technical effects: By taking the target power load data as the center and combining the fluctuation window to determine the fluctuation trend of power load data, the fluctuation characteristics of user power consumption behavior can be captured. Compared with traditional prediction methods, complex power consumption factors are considered more comprehensively. The similarity is measured by normalizing the correlation coefficient of the power load data fluctuation trend at the corresponding time of the target day and other days. It can more accurately reflect the degree of correlation between power loads on different dates and provide a reliable basis for the subsequent determination of the target sequence. The target sequence is determined based on the similarity, and the regularity of the target day is determined by combining the difference value of the days and the similarity, which helps to explore the potential rules of user power consumption in the time series and make the prediction more in line with the actual power consumption. The average load of the day to be predicted is predicted based on the daily regularity and average power load data in the preset prediction window. A variety of factors are comprehensively considered. Compared with traditional prediction methods, it can effectively reduce the power load prediction error and improve the prediction accuracy. The start and stop of the generator set is regulated according to the accurately predicted average load, which helps to avoid the problem of power grid overload or energy waste, realize the optimal dispatch of virtual power plant power, and improve the efficiency of power resource utilization and the stability of power grid operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-restrictive manner, and the same or corresponding numbers are the same or corresponding parts.

[0023] Figure 1 It is a method flow chart of a method for optimizing power dispatching of a virtual power plant in an embodiment of the present invention. DETAILED DESCRIPTION

[0024] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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 those skilled in the art without creative work are within the scope of protection of the present invention.

[0025] It should be understood that when the terms "first", "second", etc. are used in the claims, descriptions, and drawings of the present invention, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the description and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their collections.

[0026] The embodiment of the present invention discloses a method for optimizing power dispatching of a virtual power plant, referring toFigure 1 , including steps S1 - S6: S1: Obtain the daily electricity load data of users using a virtual power plant.

[0027] It should be noted that a virtual power plant combines distributed energy sources (such as solar energy, wind energy, energy storage devices, demand response devices, etc.) with the smart grid through informatization and digital means to form an integrated power supply system composed of multiple independent power generation devices and consumers; through centralized management and optimized dispatching, these distributed power resources participate in the power grid's power supply dispatching like traditional large power plants, thereby achieving the maximum utilization of energy, reducing power costs, and enhancing the stability and flexibility of the power grid operation. The electricity load data of users refers to the data on electricity usage obtained from various end-users (households, commercial buildings, industrial users, etc.), which are usually collected using monitoring devices such as smart meters, sensors, and other monitoring devices, and then transmitted to the virtual power plant system for optimizing power dispatching and management. The users referred to in the present invention are any one of the various end-users.

[0028] Implementers can set the acquisition frequency according to the specific implementation situation. For example, 1Hz.

[0029] S2: Determine the fluctuation trend of the electricity load data at each moment within each day.

[0030] It should be noted that the degree of fluctuation of the electricity load during the low-load period and the high-load period of users is different. For example, during the low-load period (small load value) of electricity usage, the fluctuation range of the electricity load data is small, and at this time, the electricity load fluctuates within a small range; while during the high-load period (large load value) of electricity usage, the electricity load data of users generally shows a large upward or downward trend as a whole; that is, during the low-load period and the high-load period of users' electricity usage, the performance of the electricity load data is different. If not distinguished, it is easy to have misjudgment phenomena, further leading to mistakes in power dispatching. Therefore, in this step, based on the daily electricity load data of users, the fluctuation trend of the electricity load data at any moment within any day of a certain user is obtained.

[0031] Record the electricity load data at the target moment within the target day of each day of the user as the target electricity load data, use the target electricity load data as the center of the preset fluctuation window, and determine the fluctuation trend of the electricity load data at the target moment within the target day according to the difference between the target electricity load data and the electricity load data at the other moments in the fluctuation window.

[0032] Implementers can set the size of the fluctuation window according to the specific implementation situation. For example, 1×51.

[0033] Specifically, the fluctuation trend satisfies: ; In the formula, is the fluctuation trend of the electricity consumption load data at the -th moment within the -th day. is the number of moments within the preset fluctuation window. is the electricity consumption load data at the -th moment within the -th day. is the electricity consumption load data at the -th moment within the -th day, corresponding to the -th moment within the preset fluctuation window. is the standard normalization function. is the absolute value symbol.

[0034] Among them, the greater the fluctuation trend, the greater the possibility that the electricity consumption load data at this moment is in the peak electricity consumption period. represents the relative difference between the remaining electricity consumption load data and the electricity consumption load data at this moment within the corresponding fluctuation window for the electricity consumption load data at this moment. The larger its value, the greater the difference, that is, the greater the fluctuation trend of the electricity consumption load data at this moment. The fluctuation trend of the electricity consumption load data at this moment is measured by calculating the average cumulative relative change difference.

[0035] S3: Determine the similarity between the electricity consumption load data of each day and that of the other days.

[0036] It should be noted that due to the influence of users' daily behavior habits, users' daily electricity consumption usually shows a periodic pattern. In order to avoid directly calculating through the electricity consumption load data, which may lead to inaccurate calculation results, in this step, the similarity between the electricity consumption load data of any two days is calculated by using the fluctuation trends of the electricity consumption load data at the corresponding moments of the two days. Through this similarity, the similarity of the electricity consumption pattern can be more effectively identified, avoiding misjudgment caused by simply relying on the difference between the electricity consumption load data.

[0037] The normalized result of the correlation coefficient of the fluctuation trends of the electricity consumption load data at all corresponding moments between the target day and the other days is used as the similarity between the electricity consumption load data of the target day and that of the other days.

[0038] Specifically, the correlation coefficient is the Pearson correlation coefficient.

[0039] Specifically, the normalization method is extreme value normalization.

[0040] In another embodiment, the normalization method is standard normalization.

[0041] S4: Determine the target sequence and regularity for each day based on the similarity.

[0042] It should be noted that usually, there is a regularity in the user's electricity consumption pattern. In case of emergencies, such as holding events, entertaining guests, etc., it will cause a sudden change in the electricity consumption situation. When predicting the user's electricity consumption data, this type of data will have an adverse impact on the prediction result. Therefore, it is necessary to calculate the regularity of the electricity load data for any day to facilitate subsequent prediction based on the result of the regularity of the electricity load data and make the prediction result more accurate.

[0043] Determine the target sequence of the target day containing multiple dates according to the magnitude of the similarity.

[0044] Specifically, the determining the target sequence of the target day containing multiple dates according to the magnitude of the similarity includes: In response to the target day and the other days corresponding to the similarity greater than the preset similarity threshold being sorted by date, a target sequence is obtained.

[0045] Implementers can set the similarity threshold according to the specific implementation situation. For example, 0.8.

[0046] Obtain the difference value of the number of days between all adjacent two days in the target sequence (the numerical value of the number of days between adjacent two days in the target sequence). According to the difference value of the number of days and the similarity of the electricity load data between the target day and the other days in the target sequence, determine the regularity of the target day.

[0047] Specifically, the regularity satisfies: ; In the formula, is the regularity of the th day, is the total number of days in the target sequence of the th day, is the th difference value of the number of days in the target sequence of the th day, is the average value of all difference values of the number of days in the target sequence of the th day, is the average value of the similarities of the electricity load data between the th day and the other days in the target sequence of the th day, is the natural exponential function, is the standard normalization function.

[0048] Among them, Indicates the difference of all daily difference values in the target sequence of the target date. The smaller its value, the smaller the difference, that is, the closer the daily difference values are, the stronger the regularity of the target date; the numerator uses for limitation to avoid inaccurate calculation results due to the small number of days in the target sequence of the target date; the mean value of the similarity , the larger its value, the more similar the target date is to all days in its target sequence, and the stronger the regularity of the target date.

[0049] S5: Predict the average load of the day to be predicted.

[0050] It should be noted that for the electricity load data of any day, when using it as a data sample for prediction, if the regularity of the electricity load data of that day is stronger, the corresponding prediction weight of that day is larger, avoiding using a unified weight, which may cause the electricity load data of a day with poor regularity to affect the prediction result.

[0051] Obtain the regularity and average electricity load data of each day within a preset prediction window, and predict the average load of the day to be predicted according to the regularity and average electricity load data.

[0052] Specifically, the average load satisfies: ; In the formula, the average load of the day to be predicted, is the regularity of the th day in the prediction window corresponding to the day to be predicted, is the number of days in the preset prediction window, is the day ordinal number of the th day in the prediction window corresponding to the day to be predicted, is the day ordinal number of the day to be predicted, is the average electricity load data of the th day in the prediction window corresponding to the day to be predicted, is the natural exponential function. Implementers can set the size of the prediction window according to the specific implementation situation. For example, 1×30. Among them, for any day in the prediction window corresponding to the day to be predicted, if the regularity of that day is stronger and it is closer to the day to be predicted, the prediction weight of that day is larger.

[0053]

[0054]

[0055] ​​​​​S6: Regulate the start and stop of the generating units according to the average load to complete the optimal power dispatching of the virtual power plant.

[0056] Obtain the sum of the average loads of all end-users on the day to be predicted to get the final prediction result. According to the prediction result, regulate the start-stop plan of the generating units and determine the standby units. For example, during high load periods, high-efficiency but costly generating units can be enabled to ensure power demand; during low load periods, some large-load units can be shut down to reduce resource waste. It is also possible to combine energy storage systems (such as pumped-storage power stations, lithium battery energy storage, etc.), charge during low load periods and discharge during high load periods according to the prediction result to balance load fluctuations, so as to improve the flexibility of the power system and complete the optimal power dispatching of the virtual power plant.

[0057] An embodiment of the present invention also discloses an optimal power dispatching system for a virtual power plant, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, an optimal power dispatching method for a virtual power plant according to the present invention is implemented.

[0058] The above system also includes other components well-known to those skilled in the art such as a communication bus and a communication interface, and their settings and functions are known in the art, so they will not be elaborated here.

[0059] In the present invention, the aforementioned memory can be any tangible medium that contains or stores a program, and this program can be used by or in combination with an instruction execution system, device, or device. For example, a computer-readable storage medium can be any suitable magnetic storage medium or magneto-optical storage medium, or any other medium that can be used to store the required information and can be accessed by an application program, module, or both. Any such computer storage medium can be part of the device or accessible or connectable to the device.

[0060] Although this specification has shown and described multiple embodiments of the present invention, it is obvious to those skilled in the art that such embodiments are provided only by way of example. Those skilled in the art will think of many changes, alterations, and alternative ways without departing from the spirit and idea of the present invention. It should be understood that various alternative solutions to the embodiments of the present invention described herein can be adopted during the practice of the present invention.

[0061] The above are all the preferred embodiments of the present invention. The protection scope of the present invention is not limited thereby. Therefore, all equivalent changes made according to the structure, shape, and principle of the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for optimizing power dispatching of a virtual power plant, characterized in that: include: Use virtual power plants to obtain users’ daily electricity load data; The power load data at the target time within the target day of each day of the user is recorded as the target power load data, and the target power load data is used as the center of the preset fluctuation window. According to the difference between the target power load data and the power load data at other times in the fluctuation window, the fluctuation trend of the power load data at the target time within the target day is determined; The normalized result of the correlation coefficient of the fluctuation trend of the power load data at all corresponding moments of the target day and the remaining days is used as the similarity of the power load data of the target day and the remaining days; According to the magnitude of the similarity, a target sequence of a target day including multiple dates is determined, and the day difference values ​​of all two adjacent days in the target sequence are obtained. According to the day difference values ​​and the similarity between the power load data of the target day and the remaining days in the target sequence, the regularity of the target day is determined; Obtaining the regularity and average power load data of each day within a preset prediction window, and predicting the average load of the prediction day based on the regularity and average power load data; According to the average load, the start and stop of the generator sets are regulated to complete the power optimization scheduling of the virtual power plant.

2. The power optimization dispatching method of a virtual power plant according to claim 1, characterized in that: The fluctuation trend satisfies: ； In the formula, For the Day The fluctuation trend of power load data at each moment, is the number of moments in the preset fluctuation window, For the Day The power load data at each moment, For the Day The first time in the preset fluctuation window corresponding to the The power load data at each moment, is the standard normalization function, is the absolute value symbol.

3. The power optimization dispatching method of a virtual power plant according to claim 1, characterized in that: The correlation coefficient is the Pearson correlation coefficient.

4. The power optimization dispatching method of a virtual power plant according to claim 1, characterized in that: The normalization method is extreme value normalization.

5. The power optimization dispatching method of a virtual power plant according to claim 1, characterized in that: Determining a target sequence of target days including multiple dates according to the magnitude of the similarity includes: In response to the target day and the corresponding remaining days having similarities greater than a preset similarity threshold, the target sequence is obtained by sorting them according to the date.

6. The power optimization dispatching method of a virtual power plant according to claim 1, characterized in that: The regularity meets the following requirements: ； In the formula, For the The regularity of the day, For the The total number of days in the target sequence of days, For the The target sequence of the day The difference in days, For the The mean of all day difference values ​​in the target sequence of the day, For the Day and Day The mean of the similarity of the power load data of the remaining days in the target sequence of the day, is the natural exponential function, is the standard normalization function.

7. The power optimization dispatching method of a virtual power plant according to claim 1, characterized in that: The average load satisfies: ； In the formula, To be predicted date The average load, To be predicted The corresponding prediction window The regularity of the day, The number of days in the preset forecast window, To be predicted day The corresponding prediction window The ordinal number of the day, To be predicted day The ordinal number of the day, To be predicted day The corresponding prediction window The average daily electricity load data, is a natural exponential function.

8. A power optimization dispatching system for a virtual power plant, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a power optimization scheduling method for a virtual power plant according to any one of claims 1-7 is implemented.

Images