Energy regulation method and system of virtual power plant based on prediction model

Abstract

This invention discloses an energy regulation method and system for virtual power plants based on a predictive model, relating to the field of virtual power plant operation control. The method includes the following steps: collecting energy data from the virtual power plant; integrating wind power data and load data, and extracting features from the integrated data; constructing a preliminary energy dynamic balance prediction model based on predicted generation data of distributed energy sources and predicted demand data of load-side resources; using the preliminary energy dynamic balance prediction model to predict day-ahead load demand and distributed energy; adjusting the operating status of each distributed energy source and load-side resource in real time using the niche gray wolf optimization algorithm according to the preliminary dispatch strategy; dynamically adjusting and optimizing the energy dynamic balance prediction model; and re-predicting load demand and distributed energy using the optimized energy dynamic balance prediction model. This invention can respond to changes in the power system in advance, achieving more efficient energy dispatch.

Description

Technical Field

[0001] This invention relates to the field of virtual power plant operation control, and more particularly to an energy regulation method and system for a virtual power plant based on a predictive model. Background Technology

[0002] With the continued growth of global energy demand, the finite reserves of fossil fuels, and the negative environmental impacts of fossil fuel use, many countries have adopted strategies such as developing renewable energy and increasing distributed generation to address this energy crisis, including wind turbines and solar photovoltaic panels. This has led to a high penetration rate of renewable energy in existing power systems, and large-scale, high-proportion grid integration presents a series of challenges in energy control and system operation. Furthermore, renewable energy typically exhibits characteristics such as naturalness, intermittency, randomness, and volatility, making it difficult to directly schedule its production, which can adversely affect grid operation. With the rapid development of smart grid technology, the concept of virtual power plants has gradually emerged in the global academic and industrial communities.

[0003] A Virtual Power Plant (VPP) is a new type of power grid operation mode. Through advanced information technology, it centrally schedules and operates various distributed energy sources, load-side resources (such as electric vehicles, energy storage facilities, and combined heat and power facilities), and renewable energy sources (such as wind and solar power) scattered throughout the power grid, making it behave like a unified and controllable power plant in the electricity market.

[0004] The main functions of a virtual power plant include two parts: first, it has the ability to forecast load and renewable energy generation; second, it has the ability to optimize and dispatch in real time, so that the operating status of distributed energy and load-side resources is always in the optimal state, thereby improving the economy and reliability of power grid operation.

[0005] Existing power plants face the following major problems in energy: wind power and load data are generally random, uncertain, and volatile, which makes it difficult for power plants to plan and dispatch energy supply; if wind power and load data cannot be accurately predicted, it is difficult to make reasonable production plans, which may lead to an imbalance between energy supply and demand in the power grid; power plants usually adopt a centralized system structure and control method, which makes it difficult for power plants to make full use of distributed energy and energy storage resources to achieve more flexible and economical operation.

[0006] No effective solutions have yet been proposed to address the problems in the relevant technologies. Summary of the Invention

[0007] To overcome the above problems, this invention aims to propose an energy regulation method and system for virtual power plants based on predictive models, with the goal of solving the problem that it is difficult to make reasonable production plans if wind power and load data cannot be accurately predicted.

[0008] Therefore, the specific technical solution adopted by the present invention is as follows:

[0009] According to one aspect of the present invention, an energy regulation method for a virtual power plant based on a predictive model is provided, the energy regulation method for the virtual power plant comprising the following steps:

[0010] S1. Collect energy data from the virtual power plant and extract wind power data and load data from the energy data;

[0011] S2. Use a linear regression model to integrate wind power data and load data, and extract features from the integrated data to obtain the predicted power generation data of distributed energy and the predicted demand data of load-side resources.

[0012] S3. Based on the predicted power generation data of the distributed energy source and the predicted demand data of the load-side resources, a preliminary energy dynamic balance prediction model is constructed using a recurrent neural network.

[0013] S4. Use a preliminary energy dynamic balance forecasting model to forecast day-ahead load demand and distributed energy resources, and formulate preliminary dispatching strategies based on the forecast results.

[0014] S5. Based on the preliminary scheduling strategy, the operating status of each distributed energy source and load-side resource is adjusted in real time using the niche gray wolf optimization algorithm.

[0015] S6. Based on real-time operating data, dynamically adjust and optimize the energy dynamic balance prediction model;

[0016] S7. Re-predict load demand and distributed energy using the optimized energy dynamic balance prediction model, and optimize the scheduling strategy based on the new prediction results.

[0017] S8. Repeat steps S5-S7 until optimal energy regulation is achieved.

[0018] Optionally, the step of integrating wind power data and load data using a linear regression model, and extracting features from the integrated data to obtain predicted generation data for distributed energy resources and predicted demand data for load-side resources includes the following steps:

[0019] S21. Obtain wind power data from the virtual power plant, wherein the wind power data includes at least wind speed, wind direction, temperature, humidity, and atmospheric pressure;

[0020] S22. Obtain load data from the virtual power plant, wherein the load data includes at least load demand, equipment operating status, and environmental parameters;

[0021] S23. Perform linear regression integration on the wind speed, wind direction, temperature, humidity and atmospheric pressure of wind power data and the load demand, equipment operating status and environmental parameters of load data to obtain a comprehensive dataset that includes both wind power data and load data.

[0022] S24. Extract features from the aggregated data of the comprehensive dataset, wherein the features include at least statistical features, frequency domain features, and time domain features;

[0023] S25. Based on the extracted features, a linear regression model is used to obtain the predicted power generation data of distributed energy and the predicted demand data of load-side resources.

[0024] Optionally, obtaining the predicted generation data of distributed energy resources and the predicted demand data of load-side resources using a linear regression model based on the extracted features includes the following steps:

[0025] S251. Based on the characteristics of the centralized data in the comprehensive dataset, and combined with historical power generation data and load demand data, construct a linear regression model;

[0026] S252. Divide the comprehensive dataset into a training set and a test set. The training set is used to train the linear regression model, and the test set is used to evaluate the predictive ability of the linear regression model.

[0027] S253. Use the training set to train the linear regression model and learn the characteristics of the data, the relationship between historical power generation data and load demand data;

[0028] S254. Use the trained linear regression model to predict the test set and evaluate the model's predictive performance.

[0029] S255. If the predictive performance of the linear regression model meets the requirements, it can be used to predict future distributed energy generation data and load-side resource demand data.

[0030] If the predictive performance of the linear regression model does not meet the requirements, then reselect the features of the data in the comprehensive dataset and adjust the parameters of the linear regression model.

[0031] Optionally, the step of using a preliminary energy dynamic balance forecasting model to forecast day-ahead load demand and distributed energy resources, and formulating a preliminary dispatching strategy based on the forecast results, includes the following steps:

[0032] S41. Using a preliminary energy dynamic balance forecasting model, input the statistical characteristics, frequency domain characteristics, and time domain characteristics of the load data to forecast the day-ahead load demand.

[0033] S42. Using a preliminary energy dynamic balance prediction model, input the statistical characteristics, frequency domain characteristics, and time domain characteristics of wind power data to predict day-ahead distributed energy generation.

[0034] S43. Based on the day-ahead load demand forecast and the day-ahead distributed energy generation forecast, formulate a preliminary dispatch strategy.

[0035] Optionally, the step of adjusting the operating status of each distributed energy source and load-side resource in real time using the niche gray wolf optimization algorithm based on the preliminary scheduling strategy includes the following steps:

[0036] S51. Based on the preliminary scheduling strategy, determine the initial operating state of each distributed energy source and load-side resource, and regard the initial operating state as a gray wolf individual in the niche gray wolf optimization algorithm.

[0037] S52. Set a fitness function, evaluate the merits of each individual gray wolf, and calculate the fitness value of each individual gray wolf.

[0038] S53. Sort all gray wolf individuals according to their fitness values, select the three best gray wolf individuals, and denot them as α gray wolf, β gray wolf, and δ gray wolf. The running state corresponding to the best fitness value is denoted as the current best solution.

[0039] S54. Based on the running status and corresponding fitness values ​​of α gray wolf, β gray wolf and δ gray wolf, calculate the update strategy and update the running status of the remaining gray wolf individuals.

[0040] S55. According to the update strategy, update the running state of the remaining gray wolf individuals to obtain the new running state, and re-evaluate the fitness values ​​of the remaining gray wolf individuals to find a new optimal solution.

[0041] S56. Determine if the maximum number of iterations has been reached;

[0042] If so, output the optimal solution;

[0043] If not, proceed to step S52;

[0044] S57. Based on the optimal solution output by the niche gray wolf optimization algorithm, adjust the operating status of each distributed energy source and load-side resource in real time.

[0045] Optionally, the process of setting a fitness function, evaluating the quality of each individual gray wolf, and calculating the fitness value of each individual gray wolf includes the following steps:

[0046] S521. Define the fitness function of the niche gray wolf optimization algorithm based on the scheduling problem of distributed energy and load-side resources, and ensure the computability and interpretability of the fitness function.

[0047] S522. Identify each individual gray wolf, obtain the operational status data of each individual gray wolf, and ensure the accuracy and timeliness of the data;

[0048] S523. Calculate and record the fitness value of each individual gray wolf through the fitness function, and set up a mechanism to automatically recalculate the fitness value to update the running status data of each individual gray wolf in real time.

[0049] S524. Regularly check the effectiveness of the fitness function and adjust the fitness function as needed.

[0050] Optionally, the step of calculating the update strategy based on the operational states and corresponding fitness values ​​of α gray wolves, β gray wolves, and δ gray wolves, and updating the operational states of the remaining gray wolf individuals, includes the following steps:

[0051] S541. Set the step size parameter A and distance adjustment parameter C for the niche gray wolf optimization algorithm;

[0052] S542. Based on the positions of α gray wolf, β gray wolf and δ gray wolf, the movement step size parameter A and the distance adjustment parameter C, calculate the new position of each non-optimal gray wolf;

[0053] S543. Perform range detection on the newly calculated non-optimal gray wolf position to determine whether it exceeds the preset solution space range;

[0054] If the solution exceeds the solution space range, the parameters need to be adjusted to bring it back into the solution space.

[0055] S544. Substitute the adjusted new position into the expression for the running state to obtain the new running state, and update the running state of the non-optimal gray wolf.

[0056] Optionally, calculating the new position for each non-optimal gray wolf includes:

[0057] Calculate the distance D between the non-optimal gray wolf and the optimal gray wolf α. α The distance D between the non-optimal gray wolf and the optimal gray wolf β β The distance D between the non-optimal gray wolf and the optimal gray wolf δ δ ;

[0058] According to distance D α Distance D β and distance D δ Calculate the new position of the non-optimal gray wolf after it moves according to the behaviors of the optimal gray wolf α, the optimal gray wolf δ, and the optimal gray wolf δ.

[0059] Wherein, the distance D α The calculation formula is as follows:

[0060]

[0061] The distance D β The calculation formula is as follows:

[0062]

[0063] The distance D δ The calculation formula is as follows:

[0064]

[0065] In the formula, Represents the non-optimal gray wolf and the optimal gray wolf. The distance between them;

[0066] Represents the non-optimal gray wolf and the optimal gray wolf. The distance between them;

[0067] Represents the non-optimal gray wolf and the optimal gray wolf. The distance between them;

[0068] , and These represent the randomly generated distance adjustment parameters;

[0069] , and These represent the calculated movement step size parameters;

[0070] This indicates the current position of the non-optimal gray wolf.

[0071] Optionally, the step of re-forecasting load demand and distributed energy resources using the optimized energy dynamic balance forecasting model, and optimizing the scheduling strategy based on the new forecast results, includes the following steps:

[0072] S71. Using the optimized energy dynamic balance prediction model, input historical and current load data to predict future load demand;

[0073] S72. Using the optimized energy dynamic balance prediction model, input historical and current wind power data to predict future distributed energy supply.

[0074] S73. Compare the predicted load demand with the distributed energy supply to identify periods of energy shortage or surplus.

[0075] S74. If there are periods of insufficient energy, increase the output of distributed energy sources, activate energy storage devices to supplement the supply, and implement load-side management to reset the operating status of each distributed energy source.

[0076] S75. If there is a period of energy surplus, reduce the output of distributed energy, send excess energy to the grid, store excess energy, and reset the operating status of each load-side resource.

[0077] S76. Evaluate the effectiveness of the new scheduling strategy and compare key energy indicators before and after scheduling.

[0078] If there is a significant difference between the assessment results and the actual results of the scheduling strategy, the energy dynamic balance prediction model will be adjusted.

[0079] According to another aspect of the present invention, an energy control system for a virtual power plant based on a prediction model is also provided. The system includes: a virtual power plant data collection and extraction module, a data integration and feature extraction module, a preliminary energy dynamic balance prediction model construction module, a prediction and scheduling strategy formulation module, a real-time scheduling optimization module, an energy dynamic balance prediction model optimization module, an optimized prediction and scheduling strategy formulation module, and a feedback module.

[0080] The virtual power plant data collection and extraction module is used to collect energy data from the virtual power plant and extract wind power data and load data from the energy data.

[0081] The data integration and feature extraction module is used to integrate wind power data and load data using a linear regression model, and to extract features from the integrated data to obtain predicted power generation data of distributed energy and predicted demand data of load-side resources.

[0082] The preliminary energy dynamic balance prediction model construction module is used to construct a preliminary energy dynamic balance prediction model based on the predicted power generation data of the distributed energy and the predicted demand data of the load-side resources using a recurrent neural network.

[0083] The forecasting and scheduling strategy formulation module is used to forecast day-ahead load demand and distributed energy using a preliminary energy dynamic balance forecasting model, and formulate a preliminary scheduling strategy based on the forecast results.

[0084] The real-time scheduling optimization module is used to adjust the operating status of each distributed energy source and load-side resource in real time based on the preliminary scheduling strategy and using the niche gray wolf optimization algorithm.

[0085] The energy dynamic balance prediction model optimization module is used to dynamically adjust and optimize the energy dynamic balance prediction model based on real-time operating data.

[0086] The optimization prediction and scheduling strategy formulation module is used to re-predict load demand and distributed energy using the optimized energy dynamic balance prediction model, and optimize the scheduling strategy based on the new prediction results.

[0087] The feedback module is used to cyclically execute the real-time scheduling optimization module, the energy dynamic balance prediction model optimization module, and the optimization prediction and scheduling strategy formulation module until optimal energy regulation is achieved.

[0088] Compared with the prior art, this application has the following beneficial effects:

[0089] 1. This invention uses an energy dynamic balance prediction model to predict future load demand and distributed energy supply, and then formulates a dispatch strategy based on the prediction results. This allows for a proactive response to changes in the power system and enables more efficient energy dispatch.

[0090] 2. This invention acquires wind power data and load data, and then integrates them using a linear regression model to obtain a comprehensive dataset containing richer information. This helps us to understand the operating status of the power system and environmental changes more comprehensively. By extracting statistical features, frequency domain features, and time domain features, we can analyze the data from different perspectives, thereby better understanding the inherent laws of the data. The linear regression model is a simple and effective prediction method that can make predictions quickly. Moreover, the model has strong interpretability and is easy to understand the prediction results.

[0091] 3. This invention utilizes the niche gray wolf optimization algorithm to adjust the operating status of distributed energy resources and load-side resources in real time to achieve optimal operating results. It can quickly respond to system changes and improve system operating efficiency. The fitness function setting allows individual gray wolves to adaptively adjust according to the actual system state, ensuring that the system always operates in an optimal state and improving system stability. The niche gray wolf optimization algorithm is a global search algorithm that avoids getting trapped in local optima. This method ensures that the solution found is the global optimum, thereby achieving the optimal scheduling strategy.

[0092] 4. By setting a maximum number of iterations, this invention can avoid infinite loops in the algorithm and improve its running efficiency. This approach can ensure efficient operation of the algorithm while satisfying the optimization effect. In the process of setting the fitness function and update strategy, parameters can be flexibly adjusted according to actual needs, such as the step size parameter and distance adjustment parameter, so as to better adapt to various different operating environments and conditions. Attached Figure Description

[0093] The above-mentioned features, characteristics, and advantages of the present invention, as well as their implementation methods, will become clearer and more readily understood in conjunction with the following description of the embodiments, which are illustrated in detail with reference to the accompanying drawings. Schematic diagrams are shown here:

[0094] Figure 1 This is a flowchart of an energy regulation method for a virtual power plant based on a predictive model according to an embodiment of the present invention. Detailed Implementation

[0095] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0096] According to embodiments of the present invention, an energy regulation method and system for a virtual power plant based on a predictive model is provided.

[0097] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments, such as... Figure 1 As shown, according to an embodiment of the present invention, an energy regulation method for a virtual power plant based on a prediction model is provided. This energy regulation method for a virtual power plant includes the following steps:

[0098] S1. Collect energy data from the virtual power plant and extract wind power data and load data from the energy data.

[0099] It should be noted that when collecting energy data in the virtual power plant, the primary focus is on the power generation of various distributed energy sources, including but not limited to wind power, photovoltaic power, and hydropower, while also collecting the power plant's load data. This data can be acquired in real time through various sensors and monitoring equipment and transmitted to a data center for processing and analysis via advanced communication technologies.

[0100] Wind power data mainly includes parameters such as wind speed, wind direction, wind turbine rotation speed, and power. Wind speed and wind direction are important factors affecting the power generation of wind turbines. By monitoring and analyzing this data in real time, wind power generation can be predicted, and the operation strategy of the power plant can be adjusted accordingly.

[0101] Load data primarily includes parameters such as the real-time load of power plants and the demand load of the power grid. The real-time load of a power plant represents its current power supply capacity, while the demand load of the power grid represents its current power demand. By monitoring and analyzing this data in real time, the operating status of power plants can be assessed, the power demand of the power grid can be predicted, and the power supply strategy of the power plants can be adjusted accordingly. By extracting and analyzing energy data from virtual power plants, the operating status of power plants can be understood, the power demand of the power grid can be predicted, the operating strategies of power plants can be optimized, and a more efficient and economical energy supply can be achieved.

[0102] S2. Use a linear regression model to integrate wind power data and load data, and extract features from the integrated data to obtain predicted power generation data for distributed energy and predicted demand data for load-side resources.

[0103] Preferably, the process of integrating wind power data and load data using a linear regression model, and extracting features from the integrated data to obtain predicted generation data for distributed energy resources and predicted demand data for load-side resources includes the following steps:

[0104] S21. Obtain wind power data from the virtual power plant, wherein the wind power data includes at least wind speed, wind direction, temperature, humidity, and atmospheric pressure;

[0105] S22. Obtain load data from the virtual power plant, wherein the load data includes at least load demand, equipment operating status, and environmental parameters;

[0106] S23. Perform linear regression integration on the wind speed, wind direction, temperature, humidity and atmospheric pressure of wind power data and the load demand, equipment operating status and environmental parameters of load data to obtain a comprehensive dataset that includes both wind power data and load data.

[0107] S24. Extract features from the aggregated data of the comprehensive dataset, wherein the features include at least statistical features, frequency domain features, and time domain features;

[0108] S25. Based on the extracted features, a linear regression model is used to obtain the predicted power generation data of distributed energy and the predicted demand data of load-side resources.

[0109] Preferably, the step of obtaining predicted generation data of distributed energy resources and predicted demand data of load-side resources using a linear regression model based on the extracted features includes the following steps:

[0110] S251. Based on the characteristics of the centralized data in the comprehensive dataset, and combined with historical power generation data and load demand data, construct a linear regression model;

[0111] S252. Divide the comprehensive dataset into a training set and a test set. The training set is used to train the linear regression model, and the test set is used to evaluate the predictive ability of the linear regression model.

[0112] S253. Use the training set to train the linear regression model and learn the characteristics of the data, the relationship between historical power generation data and load demand data;

[0113] S254. Use the trained linear regression model to predict the test set and evaluate the model's predictive performance.

[0114] S255. If the predictive performance of the linear regression model meets the requirements, it can be used to predict future distributed energy generation data and load-side resource demand data.

[0115] If the predictive performance of the linear regression model does not meet the requirements, then reselect the features of the data in the comprehensive dataset and adjust the parameters of the linear regression model.

[0116] It's important to explain that in step S25, the extracted features are used to train the linear regression model. These features include statistical features, frequency domain features, and time domain features. Statistical features help understand the basic distribution characteristics of the data, such as mean and variance; frequency domain features reveal the periodicity and frequency of the data; and time domain features help understand the trend of data changes over time. Furthermore, to improve the model's predictive performance, the dataset is typically divided into a training set and a test set. The training set is used to train the model, and the test set is used to evaluate the model's predictive performance. During model training, it's crucial to prevent overfitting, i.e., the model performing well on the training set but poorly on the test set. If the predictive performance of the linear regression model does not meet the requirements, it may be necessary to reselect features or adjust the parameters of the linear regression model. This requires judgment based on specific circumstances, such as the characteristics of the data and the complexity of the problem.

[0117] S3. Based on the predicted power generation data of the distributed energy source and the predicted demand data of the load-side resources, a preliminary energy dynamic balance prediction model is constructed using a recurrent neural network.

[0118] It's important to explain that in this process, RNNs are used to build a preliminary energy dynamic balance prediction model. This model needs to predict the generation of distributed energy resources and the demand for resources on the load side; since this data is time-series data, RNNs are a suitable choice. Building an RNN model typically involves the following steps: Defining the RNN structure, including the input layer, hidden layers, and output layer. The hidden layer contains recurrent connections to store historical information. Selecting activation functions, such as tanh or ReLU, for the hidden and output layers. Selecting optimization algorithms, such as SGD or Adam, for training the model. Defining loss functions, such as MSE or cross-entropy, to measure the model's predictive performance. Compiling the model and setting relevant parameters, such as the learning rate and number of iterations. Training the model using training data, continuously updating the model's parameters to minimize the loss function. Evaluating the model using test data to assess its predictive performance. If the performance is unsatisfactory, the model structure needs to be redefined or the parameters adjusted. Using the model for prediction, inputting new time-series data to obtain the model's prediction results.

[0119] S4. Utilize a preliminary energy dynamic balance forecasting model to forecast day-ahead load demand and distributed energy resources, and formulate preliminary dispatching strategies based on the forecast results.

[0120] Preferably, the step of using a preliminary energy dynamic balance forecasting model to forecast day-ahead load demand and distributed energy resources, and formulating a preliminary dispatching strategy based on the forecast results, includes the following steps:

[0121] S41. Using a preliminary energy dynamic balance forecasting model, input the statistical characteristics, frequency domain characteristics, and time domain characteristics of the load data to forecast the day-ahead load demand.

[0122] S42. Using a preliminary energy dynamic balance prediction model, input the statistical characteristics, frequency domain characteristics, and time domain characteristics of wind power data to predict day-ahead distributed energy generation.

[0123] S43. Based on the day-ahead load demand forecast and the day-ahead distributed energy generation forecast, formulate a preliminary dispatch strategy.

[0124] It's important to clarify that this process involves day-ahead dispatching of the energy system, which involves developing a reasonable power generation dispatching strategy based on predicted load demand and distributed energy generation data. This is a crucial step in power system operation, aiming to rationally allocate power resources to meet user electricity demands while ensuring the safety and economic efficiency of power system operation. First, a dynamic energy balance forecasting model is used to predict day-ahead load demand and distributed energy generation. This process requires inputting the statistical, frequency, and time domain characteristics of load and wind power data. A preliminary dispatching strategy is then developed based on the forecast results. Power system dispatching strategies generally need to consider various factors, including the grid's operating status, power plant output, electricity market prices, and renewable energy generation forecasts. The goal of developing a dispatching strategy is to ensure the stable operation of the power system while maximizing economic efficiency.

[0125] S5. Based on the preliminary scheduling strategy, the operating status of each distributed energy source and load-side resource is adjusted in real time using the niche gray wolf optimization algorithm.

[0126] Preferably, the step of adjusting the operating status of each distributed energy source and load-side resource in real time using the niche gray wolf optimization algorithm according to the preliminary scheduling strategy includes the following steps:

[0127] S51. Based on the preliminary scheduling strategy, determine the initial operating state of each distributed energy source and load-side resource, and regard the initial operating state as a gray wolf individual in the niche gray wolf optimization algorithm.

[0128] S52. Set a fitness function to evaluate the merits of each individual gray wolf (i.e., the operating status of distributed energy and load-side resources) and calculate the fitness value of each individual gray wolf.

[0129] S53. Sort all gray wolf individuals according to their fitness values, select the three best gray wolf individuals, and denot them as α gray wolf, β gray wolf, and δ gray wolf. The running state corresponding to the best fitness value is denoted as the current best solution.

[0130] S54. Based on the running states (i.e., their positions in the solution space) and corresponding fitness values ​​of α gray wolves, β gray wolves, and δ gray wolves, calculate the update strategy and update the running states of the remaining gray wolf individuals.

[0131] S55. According to the update strategy, update the running state of the remaining gray wolf individuals to obtain the new running state, and re-evaluate the fitness values ​​of the remaining gray wolf individuals to find a new optimal solution.

[0132] S56. Determine if the maximum number of iterations has been reached;

[0133] If so, output the optimal solution;

[0134] If not, proceed to step S52;

[0135] S57. Based on the optimal solution output by the niche gray wolf optimization algorithm, adjust the operating status of each distributed energy source and load-side resource in real time.

[0136] Preferably, the steps of setting a fitness function, evaluating the quality of each individual gray wolf, and calculating the fitness value of each individual gray wolf include the following:

[0137] S521. Define the fitness function of the niche gray wolf optimization algorithm based on the scheduling problem of distributed energy and load-side resources, and ensure the computability and interpretability of the fitness function.

[0138] S522. Identify each individual gray wolf, obtain the operational status data of each individual gray wolf, and ensure the accuracy and timeliness of the data;

[0139] S523. Calculate and record the fitness value of each individual gray wolf through the fitness function, and set up a mechanism to automatically recalculate the fitness value to update the running status data of each individual gray wolf in real time.

[0140] S524. Regularly check the effectiveness of the fitness function and adjust the fitness function as needed.

[0141] Preferably, the step of calculating the update strategy based on the operational status and corresponding fitness values ​​of α gray wolves, β gray wolves, and δ gray wolves, and updating the operational status of the remaining gray wolf individuals, includes the following steps:

[0142] S541. Set the step size parameter A and distance adjustment parameter C for the niche gray wolf optimization algorithm;

[0143] S542. Based on the positions of α gray wolf, β gray wolf and δ gray wolf, the movement step size parameter A and the distance adjustment parameter C, calculate the new position of each non-optimal gray wolf;

[0144] S543. Perform range detection on the newly calculated non-optimal gray wolf position to determine whether it exceeds the preset solution space range;

[0145] If the solution exceeds the solution space range, the parameters need to be adjusted to bring it back into the solution space.

[0146] S544. Substitute the adjusted new position into the expression for the running state to obtain the new running state, and update the running state of the non-optimal gray wolf.

[0147] Preferably, calculating the new position of each non-optimal gray wolf includes:

[0148] Calculate the distance D between the non-optimal gray wolf and the optimal gray wolf α. αThe distance D between the non-optimal gray wolf and the optimal gray wolf β β The distance D between the non-optimal gray wolf and the optimal gray wolf δ δ ;

[0149] According to distance D α Distance D β and distance D δ Calculate the new position of the non-optimal gray wolf after it moves according to the behaviors of the optimal gray wolf α, the optimal gray wolf δ, and the optimal gray wolf δ.

[0150] Wherein, the distance D α The calculation formula is as follows:

[0151]

[0152] The distance D β The calculation formula is as follows:

[0153]

[0154] The distance D δ The calculation formula is as follows:

[0155]

[0156] In the formula, Represents the non-optimal gray wolf and the optimal gray wolf. The distance between them;

[0157] Represents the non-optimal gray wolf and the optimal gray wolf. The distance between them;

[0158] Represents the non-optimal gray wolf and the optimal gray wolf. The distance between them;

[0159] , and These represent the randomly generated distance adjustment parameters;

[0160] , and These represent the calculated movement step size parameters;

[0161] This indicates the current position of the non-optimal gray wolf.

[0162] The formula for calculating the new position of the non-optimal gray wolf after moving according to the behaviors of the optimal gray wolf α, optimal gray wolf δ, and optimal gray wolf δ is as follows:

[0163]

[0164] In the formula, Indicates the new position after the move;

[0165] , and These represent the calculated movement step size parameters;

[0166] Represents the non-optimal gray wolf and the optimal gray wolf. The distance between them;

[0167] Represents the non-optimal gray wolf and the optimal gray wolf. The distance between them;

[0168] Represents the non-optimal gray wolf and the optimal gray wolf. The distance between them.

[0169] It's important to explain that, firstly, all individual gray wolves need to be evaluated, typically using a function called the fitness function. This function assigns a value to each individual gray wolf based on its current state; this value is its fitness value. A higher fitness value indicates a better individual gray wolf. Next, the fitness values ​​of all gray wolves are sorted. This reveals which gray wolves are better and which are relatively weaker. Then, the three gray wolves with the highest fitness values ​​are selected; these are α, β, and δ. These three gray wolves can be considered leaders in the current optimization process, and other gray wolves will try to learn or imitate their behavior. Finally, the running state corresponding to the optimal fitness value is the current optimal solution. That is, the current state is considered the best among all possible states. This optimal solution is recorded as the current best result.

[0170] S6. Based on real-time operating data, dynamically adjust and optimize the energy dynamic balance prediction model.

[0171] It's important to clarify that in actual operation, the predictive performance of the prediction model may change due to variations in the power system's operating status, environmental conditions, and equipment performance. Therefore, it's necessary to dynamically adjust and optimize the energy dynamic balance prediction model based on real-time operational data to ensure its predictive performance. This is a common method for dynamically adjusting and optimizing models. When new data is received, it can be used to update the model's parameters. For example, online learning methods can be used, training the model each time new data is received. Alternatively, a certain amount of new data can be collected periodically, and then used to retrain the model. Over time, some features may lose their predictive power, while new features may emerge. Therefore, feature selection methods can be used to periodically update the model's features. This can be achieved by evaluating the impact of each feature on predictive performance. If multiple prediction models are available, model fusion methods can be used to improve predictive performance.

[0172] S7. Re-predict load demand and distributed energy using the optimized energy dynamic balance prediction model, and optimize the scheduling strategy based on the new prediction results.

[0173] Preferably, the step of re-forecasting load demand and distributed energy resources using the optimized energy dynamic balance prediction model, and optimizing the scheduling strategy based on the new prediction results, includes the following steps:

[0174] S71. Using the optimized energy dynamic balance prediction model, input historical and current load data to predict future load demand;

[0175] S72. Using the optimized energy dynamic balance prediction model, input historical and current wind power data to predict future distributed energy supply.

[0176] S73. Compare the predicted load demand with the distributed energy supply to identify periods of energy shortage or surplus.

[0177] S74. If there are periods of insufficient energy, increase the output of distributed energy sources, activate energy storage devices to supplement the supply, and implement load-side management to reset the operating status of each distributed energy source.

[0178] S75. If there is a period of energy surplus, reduce the output of distributed energy, send excess energy to the grid, store excess energy, and reset the operating status of each load-side resource.

[0179] S76. Evaluate the effectiveness of the new scheduling strategy and compare key energy indicators before and after scheduling.

[0180] If there is a significant difference between the assessment results and the actual results of the scheduling strategy, the energy dynamic balance prediction model will be adjusted.

[0181] It's important to explain that the process begins with using an optimized energy dynamic balance prediction model. Historical and current load and wind power data are input to predict future load demand and distributed energy supply. Next, the predicted load demand and distributed energy supply are compared to identify periods of energy shortage or surplus. This is a crucial step as it helps identify the time periods and locations requiring adjustment. Then, based on the comparison results, corresponding dispatch strategies are developed. Specifically, if there are periods of energy shortage, the output of distributed energy sources needs to be increased, energy storage devices activated to supplement supply, load-side management implemented, and the operating status of each distributed energy source reset. Conversely, if there are periods of energy surplus, the output of distributed energy sources needs to be reduced, excess energy supplied to the grid and stored, and the operating status of each load-side resource reset. Finally, the effectiveness of the new dispatch strategy is evaluated, and key energy indicators before and after the dispatch are compared. If the evaluation results differ significantly from the actual results of the dispatch strategy, it indicates a potential problem with the prediction model, requiring further adjustment and optimization.

[0182] S8. Repeat steps S5-S7 until optimal energy regulation is achieved.

[0183] According to another embodiment of the present invention, an energy control system for a virtual power plant based on a prediction model is also provided. The system includes: a virtual power plant data collection and extraction module, a data integration and feature extraction module, a preliminary energy dynamic balance prediction model construction module, a prediction and scheduling strategy formulation module, a real-time scheduling optimization module, an energy dynamic balance prediction model optimization module, an optimized prediction and scheduling strategy formulation module, and a feedback module.

[0184] The virtual power plant data collection and extraction module is used to collect energy data from the virtual power plant and extract wind power data and load data from the energy data.

[0185] The data integration and feature extraction module is used to integrate wind power data and load data using a linear regression model, and to extract features from the integrated data to obtain predicted power generation data of distributed energy and predicted demand data of load-side resources.

[0186] The preliminary energy dynamic balance prediction model construction module is used to construct a preliminary energy dynamic balance prediction model based on the predicted power generation data of the distributed energy and the predicted demand data of the load-side resources using a recurrent neural network.

[0187] The forecasting and scheduling strategy formulation module is used to forecast day-ahead load demand and distributed energy using a preliminary energy dynamic balance forecasting model, and formulate a preliminary scheduling strategy based on the forecast results.

[0188] The real-time scheduling optimization module is used to adjust the operating status of each distributed energy source and load-side resource in real time based on the preliminary scheduling strategy and using the niche gray wolf optimization algorithm.

[0189] The energy dynamic balance prediction model optimization module is used to dynamically adjust and optimize the energy dynamic balance prediction model based on real-time operating data.

[0190] The optimization prediction and scheduling strategy formulation module is used to re-predict load demand and distributed energy using the optimized energy dynamic balance prediction model, and optimize the scheduling strategy based on the new prediction results.

[0191] The feedback module is used to cyclically execute the real-time scheduling optimization module, the energy dynamic balance prediction model optimization module, and the optimization prediction and scheduling strategy formulation module until optimal energy regulation is achieved.

[0192] Specifically, to facilitate a better understanding by those skilled in the art, the technical terms or some nouns that may be involved in this application are explained below in the relevant embodiments:

[0193] Linear regression model: A linear regression model is a predictive model based on the linear relationship between input and output variables. The parameters of the model are determined by minimizing the squared error between the actual output value and the predicted output value.

[0194] A Recurrent Neural Network (RNN) is a type of neural network capable of processing sequential data. It possesses memory capabilities, remembering previous computation results and applying them to current calculations. This makes RNNs well-suited for processing time-series data, such as speech and text.

[0195] In summary, by utilizing the above-mentioned technical solutions of this invention, the present invention employs an energy dynamic balance prediction model to predict future load demand and distributed energy supply, and then formulates scheduling strategies based on the prediction results. This allows for proactive responses to changes in the power system, achieving more efficient energy dispatch. By acquiring wind power and load data and integrating them using a linear regression model, this invention obtains a comprehensive dataset containing richer information, helping us to more fully understand the operating status of the power system and environmental changes. By extracting statistical, frequency domain, and time domain features, data can be analyzed from different perspectives, thereby better understanding the inherent patterns of the data. The linear regression model is a simple and effective prediction method that can quickly make predictions, and its strong interpretability facilitates understanding the prediction results. Furthermore, by using the niche gray wolf optimization algorithm, this invention can adjust distributed energy supply in real time. The system monitors the operational status of both source and load-side resources to achieve optimal performance, enabling rapid response to system changes and improved efficiency. The fitness function allows individual gray wolves to adaptively adjust based on the system's actual state, ensuring optimal system operation and enhancing stability. The niche gray wolf optimization algorithm, a global search algorithm, avoids getting trapped in local optima, ensuring the found solution is globally optimal and achieving the best scheduling strategy. Setting a maximum number of iterations prevents infinite loops, improving efficiency. This approach ensures efficient operation while meeting optimization requirements. Parameters, such as step size and distance adjustment, can be flexibly adjusted based on actual needs during fitness function and update strategy settings to better adapt to various operating environments and conditions.

[0196] Although the present invention has been disclosed above with reference to preferred embodiments, the embodiments are merely examples for illustrative purposes and are not intended to limit the present invention. Those skilled in the art can make various modifications and refinements without departing from the spirit and scope of the present invention. The scope of protection claimed by the present invention should be determined by the claims.

Claims

1. An energy regulation method for a virtual power plant based on a predictive model, characterized in that, The energy regulation method of this virtual power plant includes the following steps: S1. Collect energy data from the virtual power plant and extract wind power data and load data from the energy data; the energy data includes wind power, photovoltaic power, and hydropower data, the wind power data includes wind speed, wind direction, wind turbine speed and power; the load data includes the real-time load of the power plant and the demand load of the power grid; S2. Use a linear regression model to integrate wind power data and load data, and extract features from the integrated data to obtain the predicted power generation data of distributed energy and the predicted demand data of load-side resources. S3. Based on the predicted power generation data of the distributed energy source and the predicted demand data of the load-side resources, a preliminary energy dynamic balance prediction model is constructed using a recurrent neural network. S4. Use a preliminary energy dynamic balance forecasting model to forecast day-ahead load demand and distributed energy resources, and formulate preliminary dispatching strategies based on the forecast results. S5. Based on the preliminary scheduling strategy, the operating status of each distributed energy source and load-side resource is adjusted in real time using the niche gray wolf optimization algorithm. S6. Based on real-time operating data, dynamically adjust and optimize the energy dynamic balance prediction model; the real-time operating data refers to the energy data in the virtual power plant. S7. Re-predict load demand and distributed energy using the optimized energy dynamic balance prediction model, and optimize the scheduling strategy based on the new prediction results. S8. Repeat steps S5-S7 until optimal energy regulation is achieved.

2. The energy regulation method for a virtual power plant based on a predictive model according to claim 1, characterized in that, The process of integrating wind power and load data using a linear regression model, and extracting features from the integrated data to obtain predicted generation data for distributed energy resources and predicted demand data for load-side resources includes the following steps: S21. Obtain wind power data from the virtual power plant, wherein the wind power data includes at least wind speed, wind direction, temperature, humidity, and atmospheric pressure; S22. Obtain load data from the virtual power plant, wherein the load data includes at least load demand, equipment operating status, and environmental parameters; S23. Perform linear regression integration on the wind speed, wind direction, temperature, humidity and atmospheric pressure of wind power data and the load demand, equipment operating status and environmental parameters of load data to obtain a comprehensive dataset that includes both wind power data and load data. S24. Extract features from the aggregated data of the comprehensive dataset, wherein the features include at least statistical features, frequency domain features, and time domain features; S25. Based on the extracted features, a linear regression model is used to obtain the predicted power generation data of distributed energy and the predicted demand data of load-side resources.

3. The energy regulation method for a virtual power plant based on a predictive model according to claim 2, characterized in that, The process of obtaining predicted generation data for distributed energy resources and predicted demand data for load-side resources using a linear regression model based on extracted features includes the following steps: S251. Based on the characteristics of the centralized data in the comprehensive dataset, and combined with historical power generation data and load demand data, construct a linear regression model; S252. Divide the comprehensive dataset into a training set and a test set. The training set is used to train the linear regression model, and the test set is used to evaluate the predictive ability of the linear regression model. S253. Use the training set to train the linear regression model and learn the characteristics of the data, the relationship between historical power generation data and load demand data; S254. Use the trained linear regression model to predict the test set and evaluate the model's predictive performance. S255. If the predictive performance of the linear regression model meets the requirements, it can be used to predict future distributed energy generation data and load-side resource demand data. If the predictive performance of the linear regression model does not meet the requirements, then reselect the features of the data in the comprehensive dataset and adjust the parameters of the linear regression model.

4. The energy regulation method for a virtual power plant based on a predictive model according to claim 1, characterized in that, The process of using a preliminary energy dynamic balance forecasting model to predict day-ahead load demand and distributed energy resources, and formulating preliminary dispatching strategies based on the forecast results, includes the following steps: S41. Using a preliminary energy dynamic balance forecasting model, input the statistical characteristics, frequency domain characteristics, and time domain characteristics of the load data to forecast the day-ahead load demand. S42. Using a preliminary energy dynamic balance prediction model, input the statistical characteristics, frequency domain characteristics, and time domain characteristics of wind power data to predict day-ahead distributed energy generation. S43. Based on the day-ahead load demand forecast and the day-ahead distributed energy generation forecast, formulate a preliminary dispatch strategy.

5. The energy regulation method for a virtual power plant based on a predictive model according to claim 1, characterized in that, The process of adjusting the operating status of distributed energy sources and load-side resources in real time based on the initial scheduling strategy using the niche gray wolf optimization algorithm includes the following steps: S51. Based on the preliminary scheduling strategy, determine the initial operating state of each distributed energy source and load-side resource, and regard the initial operating state as a gray wolf individual in the niche gray wolf optimization algorithm. S52. Set a fitness function, evaluate the merits of each individual gray wolf, and calculate the fitness value of each individual gray wolf. S53. Sort all gray wolf individuals according to their fitness values, select the three best gray wolf individuals, and denot them as α gray wolf, β gray wolf, and δ gray wolf. The running state corresponding to the best fitness value is denoted as the current best solution. S54. Based on the running status and corresponding fitness values ​​of α gray wolf, β gray wolf and δ gray wolf, calculate the update strategy and update the running status of the remaining gray wolf individuals. S55. According to the update strategy, update the running state of the remaining gray wolf individuals to obtain the new running state, and re-evaluate the fitness values ​​of the remaining gray wolf individuals to find a new optimal solution. S56. Determine if the maximum number of iterations has been reached; If so, output the optimal solution; If not, proceed to step S52; S57. Based on the optimal solution output by the niche gray wolf optimization algorithm, adjust the operating status of each distributed energy source and load-side resource in real time.

6. The energy regulation method for a virtual power plant based on a predictive model according to claim 5, characterized in that, The process of setting a fitness function to evaluate the quality of each individual gray wolf and calculating the fitness value of each individual gray wolf includes the following steps: S521. Define the fitness function of the niche gray wolf optimization algorithm based on the scheduling problem of distributed energy and load-side resources, and ensure the computability and interpretability of the fitness function. S522. Identify each individual gray wolf, obtain the operational status data of each individual gray wolf, and ensure the accuracy and timeliness of the data; S523. Calculate and record the fitness value of each individual gray wolf through the fitness function, and set up a mechanism to automatically recalculate the fitness value to update the running status data of each individual gray wolf in real time. S524. Regularly check the effectiveness of the fitness function and adjust the fitness function as needed.

7. The energy regulation method for a virtual power plant based on a predictive model according to claim 5, characterized in that, The process of calculating the update strategy based on the operational status and corresponding fitness values ​​of α gray wolves, β gray wolves, and δ gray wolves, and updating the operational status of the remaining gray wolf individuals, includes the following steps: S541. Set the step size parameter A and distance adjustment parameter C for the niche gray wolf optimization algorithm; S542. Based on the positions of α gray wolf, β gray wolf and δ gray wolf, the movement step size parameter A and the distance adjustment parameter C, calculate the new position of each non-optimal gray wolf; S543. Perform range detection on the newly calculated non-optimal gray wolf position to determine whether it exceeds the preset solution space range; If the solution exceeds the solution space range, the parameters need to be adjusted to bring it back into the solution space. S544. Substitute the adjusted new position into the expression for the running state to obtain the new running state, and update the running state of the non-optimal gray wolf.

8. The energy regulation method for a virtual power plant based on a predictive model according to claim 7, characterized in that, The calculation of the new position for each non-optimal gray wolf includes: Calculate the distance D between the non-optimal gray wolf and the optimal gray wolf α. α The distance D between the non-optimal gray wolf and the optimal gray wolf β β The distance D between the non-optimal gray wolf and the optimal gray wolf δ δ ; According to distance D α Distance D β and distance D δ Calculate the new position of the non-optimal gray wolf after it moves according to the behaviors of the optimal gray wolf α, the optimal gray wolf δ, and the optimal gray wolf δ. Wherein, the distance D α The calculation formula is as follows: The distance D β The calculation formula is as follows: The distance D δ The calculation formula is as follows: In the formula, Represents the non-optimal gray wolf and the optimal gray wolf. The distance between them; Represents the non-optimal gray wolf and the optimal gray wolf. The distance between them; Represents the non-optimal gray wolf and the optimal gray wolf. The distance between them; , and These represent the randomly generated distance adjustment parameters; , and These represent the calculated movement step size parameters; This indicates the current position of the non-optimal gray wolf.

9. The energy regulation method for a virtual power plant based on a predictive model according to claim 1, characterized in that, The process of re-forecasting load demand and distributed energy resources using the optimized energy dynamic balance prediction model, and optimizing the scheduling strategy based on the new prediction results, includes the following steps: S71. Using the optimized energy dynamic balance prediction model, input historical and current load data to predict future load demand; S72. Using the optimized energy dynamic balance prediction model, input historical and current wind power data to predict future distributed energy supply. S73. Compare the predicted load demand with the distributed energy supply to identify periods of energy shortage or surplus. S74. If there are periods of insufficient energy, increase the output of distributed energy sources, activate energy storage devices to supplement the supply, and implement load-side management to reset the operating status of each distributed energy source. S75. If there is a period of energy surplus, reduce the output of distributed energy, send excess energy to the grid, store excess energy, and reset the operating status of each load-side resource. S76. Evaluate the effectiveness of the new scheduling strategy and compare key energy indicators before and after scheduling. If the assessment results differ from the actual results of the scheduling strategy, the energy dynamic balance prediction model will be adjusted.

10. An energy control system for a virtual power plant based on a predictive model, used to implement the energy control method for a virtual power plant based on a predictive model as described in any one of claims 1-9, characterized in that, The system includes: a virtual power plant data collection and extraction module, a data integration and feature extraction module, a preliminary energy dynamic balance prediction model construction module, a prediction and dispatch strategy formulation module, a real-time dispatch optimization module, an energy dynamic balance prediction model optimization module, an optimized prediction and dispatch strategy formulation module, and a feedback module; The virtual power plant data collection and extraction module is used to collect energy data from the virtual power plant and extract wind power data and load data from the energy data. The data integration and feature extraction module is used to integrate wind power data and load data using a linear regression model, and to extract features from the integrated data to obtain predicted power generation data of distributed energy and predicted demand data of load-side resources. The preliminary energy dynamic balance prediction model construction module is used to construct a preliminary energy dynamic balance prediction model based on the predicted power generation data of the distributed energy and the predicted demand data of the load-side resources using a recurrent neural network. The forecasting and scheduling strategy formulation module is used to forecast day-ahead load demand and distributed energy using a preliminary energy dynamic balance forecasting model, and formulate a preliminary scheduling strategy based on the forecast results. The real-time scheduling optimization module is used to adjust the operating status of each distributed energy source and load-side resource in real time based on the preliminary scheduling strategy and using the niche gray wolf optimization algorithm. The energy dynamic balance prediction model optimization module is used to dynamically adjust and optimize the energy dynamic balance prediction model based on real-time operating data. The optimization prediction and scheduling strategy formulation module is used to re-predict load demand and distributed energy using the optimized energy dynamic balance prediction model, and optimize the scheduling strategy based on the new prediction results. The feedback module is used to cyclically execute the real-time scheduling optimization module, the energy dynamic balance prediction model optimization module, and the optimization prediction and scheduling strategy formulation module until optimal energy regulation is achieved.

Images