Multi-mode coordination control method and system based on wind and light storage new energy power station

Abstract

The invention relates to the technical field of power regulation and control, and discloses a multi-mode coordination control method and system based on a wind and light storage new energy power station, and the method comprises the following steps: obtaining the real-time operation data of a wind power, photovoltaic power generation and energy storage system; predicting power grid load and wind power and photovoltaic power generation conditions in a period of time in the future; utilizing a collaborative recommendation algorithm to recommend an energy scheduling scheme set; an optimal energy scheduling scheme is determined by using an improved Bayesian optimization algorithm based on an immune mechanism; performing corresponding control operation on wind energy, light energy and energy storage equipment according to the obtained optimal energy scheduling scheme to realize efficient cooperative work of various energy sources; the invention further discloses a multi-mode coordination control system based on the wind and light storage new energy power station. According to the invention, the optimal energy scheduling scheme can be found more quickly, the coordination control of the wind and light storage new energy power station is realized efficiently, and the utilization efficiency of energy is improved to the greatest extent while the load demand of the power grid is better met.

Description

Technical Field

[0001] The present invention relates to the field of electric power regulation technology, and in particular to a multi-mode coordinated control method and system based on a wind, solar and energy storage new energy power station. Background Art

[0002] With the rapid depletion of fossil energy such as coal, oil and natural gas, and the continuous deterioration of the ecological environment, especially the increasingly serious global climate change caused by greenhouse gas emissions, the sustainable development of human society has been threatened unprecedentedly. Therefore, energy technology is gradually turning to low-carbon and carbon-free development. In this context, the power system, as an aspect closely related to human life, is also facing a series of challenges.

[0003] Smart grid is a widely distributed and highly automated energy exchange system, which is characterized by the two-way flow of electricity and information. In order to promote the development of smart grid, a large number of demonstration practices are being carried out all over the world. To truly realize smart grid, it is necessary to achieve the goals of reliability, safety, economy, efficiency, environmental protection and safe use of the power grid in all links such as power generation, transmission, distribution, power consumption, energy storage and microgrid. Therefore, accelerating the construction of a smart grid multi-level energy system with two-way interaction in all links including power generation, transmission, distribution and power consumption is an important stage and foundation for my country to comprehensively promote the construction of smart grid.

[0004] In recent years, my country's new energy industry has developed rapidly, playing an important role in improving energy structure, protecting the environment, and promoting economic development. However, new energy power generation such as wind power and photovoltaic power generation is intermittent and volatile. Large-scale use of new energy power generation will cause large fluctuations in the voltage and frequency of the power grid, which will bring difficulties to stable operation. Therefore, the introduction of a large-capacity energy storage system is an important method to solve the stable operation of large-scale grid-connected power generation of new energy. The application of energy storage systems combined with wind power and photovoltaic power generation systems to build an integrated wind, solar and storage system can coordinate the power output of wind power and photovoltaic power generation and reduce their output fluctuations. This control scheme is gradually gaining attention. It is an important means to ensure the stable operation of large-scale grid-connected power generation of new energy, and this idea is becoming a research hotspot.

[0005] However, the operation mode of the traditional wind, solar and energy storage new energy power station is relatively single and extensive, lacking mutual coordination and system optimization operation mode. Therefore, the present invention proposes a multi-mode coordinated control method and system based on the wind, solar and energy storage new energy power station. Summary of the invention

[0006] In response to the problems in the related technology, the present invention proposes a multi-mode coordinated control method and system based on a wind, solar and energy storage new energy power station to overcome the above-mentioned technical problems existing in the existing related technology.

[0007] To this end, the specific technical solution adopted by the present invention is as follows:

[0008] According to one aspect of the present invention, a multi-mode coordinated control method based on a wind, solar and energy storage new energy power station is provided, the method comprising the following steps:

[0009] S1. Obtain real-time operation data of wind power, photovoltaic power generation and energy storage systems through external protocol access;

[0010] S2. Use real-time operation data combined with weather data to predict the grid load and wind power and photovoltaic power generation in the future;

[0011] S3, using a collaborative recommendation algorithm to recommend an energy scheduling solution according to the predicted data, and obtaining an energy scheduling solution set;

[0012] S4, determining the best energy dispatching scheme from the energy dispatching scheme set by using an improved Bayesian optimization algorithm based on an immune mechanism according to the power generation capacity, energy storage capacity and grid load demand of the power station;

[0013] S5. Carrying out corresponding control operations on wind energy, solar energy and energy storage equipment according to the obtained optimal energy scheduling plan is the key to achieving efficient coordinated work of various energy sources.

[0014] Preferably, the method of using the real-time operation data in combination with the weather data to respectively predict the grid load and wind power and photovoltaic power generation conditions in the future period includes the following steps:

[0015] S21. Obtain historical operation data of wind power generation, photovoltaic power generation, and energy storage systems and corresponding weather data from the database;

[0016] S22, using historical load data and corresponding time and weather data to build a power grid load prediction model, and using the trained power grid load prediction model to predict the power grid load in the future;

[0017] S23, building a wind power generation prediction model based on historical wind power generation operation data, and using the trained wind power generation prediction model to predict wind power generation data in the future;

[0018] S24. Use historical photovoltaic power generation operation data to build a photovoltaic power generation prediction model, and use the trained photovoltaic power generation prediction model to predict photovoltaic power generation data in the future.

[0019] Preferably, the grid load prediction model, wind power generation prediction model and photovoltaic power generation prediction model are all any one of a time series model, a regression model, a machine learning model, a deep learning model or an integrated learning model.

[0020] Preferably, the step of using a collaborative recommendation algorithm to recommend an energy scheduling solution according to the prediction data to obtain an energy scheduling solution set comprises the following steps:

[0021] S31, obtaining historical data and forecast data and performing preprocessing, the historical data including energy data, load demand data and weather conditions, and the forecast data including grid load forecast data, wind power generation forecast data and photovoltaic power generation forecast data;

[0022] S32, constructing a scoring matrix of time / load demand and energy type based on the acquired historical data and forecast data;

[0023] S33, using a clustering algorithm to calculate the scoring matrix of time / load demand and energy type to obtain several different categories, and cutting the scoring matrix in the time / load demand dimension to obtain several scoring matrices of time / load demand and energy type;

[0024] S34, using a similarity measurement method to find the neighboring time / load demand of the target time / load demand in each category according to the clustering result, and analyzing to obtain a recommended energy type set of the time / load demand in the category;

[0025] S35. Comprehensively recommend energy types in all categories, sort the predicted score values, and use a preset number of recommended energy types as a set of energy scheduling solutions recommended by the prediction data based on the sorting results from high to low.

[0026] Preferably, the step of calculating the scoring matrix of time / load demand and energy type by using a clustering algorithm to obtain a number of different categories, and cutting the scoring matrix in the time / load demand dimension to obtain a number of scoring matrices of time / load demand and energy type includes the following steps:

[0027] S331, taking each time / load demand and each energy type score in the time / load demand and energy type score matrix as a separate sample and sample feature, respectively, to construct a sample feature matrix;

[0028] S332, setting the number k of cluster centers, and randomly selecting k samples from the sample feature matrix as initial cluster centers;

[0029] S333, based on the initial cluster center, using the Euclidean distance calculation method to calculate the distance between two time / load demand samples, and classifying the samples into the category with the closest distance;

[0030] S334, when a round of clustering is completed, a new center of each category is obtained, and the coordinates of the new cluster center are defined as the centroid of the category;

[0031] S335, repeat S333 and S334, when the distance between the latest cluster center and the original cluster center is less than or equal to the preset distance threshold, clustering is completed, and k different categories are obtained;

[0032] S336. Cut the scoring matrix in the time / load demand dimension, extract the scores corresponding to the samples of each category, and obtain k new scoring matrices.

[0033] Preferably, the calculation formula of the predicted score value is:

[0034]

[0035] Where P(u,i) represents the prediction score of target time / load demand u for energy type i;

[0036] Sim(u,v) represents the similarity between the target time / load demand u and the neighbor time / load demand v;

[0037] R(v,i) represents the actual score of the neighbor time / load demand v for energy type i.

[0038] Preferably, the method of determining the best energy scheduling scheme from the energy scheduling scheme set by using an improved Bayesian optimization algorithm based on an immune mechanism according to the power generation capacity, energy storage capacity and grid load demand of the power station comprises the following steps:

[0039] S41, obtaining data on power generation capacity, energy storage capacity and grid load demand of the power station, encoding them, and creating an initial energy scheduling plan set;

[0040] S42, setting a fitness function based on the power generation cost, environmental impact, grid load matching, grid demand satisfaction and energy consumption of the power station;

[0041] S43, establishing a Bayesian network for predicting the fitness of a new energy scheduling scheme by analyzing the initial energy scheduling scheme set;

[0042] S44, using the maximum likelihood estimation algorithm in combination with the energy scheduling scheme set to calculate the conditional probability of each node in the Bayesian network, and using the conditional probability of the node as a parameter of the Bayesian network;

[0043] S45, sampling the Bayesian network according to the conditional probability of the Bayesian network to generate a new energy scheduling scheme individual;

[0044] S46, selecting the energy scheduling scheme individual with the highest fitness value to make a vaccine, inoculating the newly generated energy scheduling scheme individuals, and selecting the next generation of energy scheduling scheme sets according to the fitness of the energy scheduling scheme individuals;

[0045] S47, determine whether the number of iterations reaches the maximum number of iterations or whether the fitness of the new energy scheduling scheme individual reaches the preset threshold. If so, output the current energy scheduling scheme as the optimal energy scheduling scheme, otherwise, return to S42 and continue iterating.

[0046] Preferably, the calculation formula of the fitness function is:

[0047] y=ω1C1+ω2C2+ω3C3+ω4C4-ω5C5

[0048] ω1+ω2+ω3+ω4+ω5=1;

[0049] In the formula, y represents the fitness value, C1 represents the standardized carbon dioxide emissions, ω1 represents the weight of the standardized carbon dioxide emissions, C2 represents the standardized power generation cost, ω2 represents the weight of the standardized power generation cost, C3 represents the standardized grid load matching degree, ω3 represents the weight of the standardized grid load matching degree, C4 represents the standardized grid demand satisfaction, ω4 represents the weight of the standardized grid demand satisfaction, C5 represents the standardized energy consumption, and ω5 represents the weight of the standardized energy consumption.

[0050] Preferably, the energy scheduling scheme individual with the highest fitness value is selected to prepare a vaccine, the newly generated energy scheduling scheme individuals are vaccinated, and the next generation of energy scheduling scheme set is obtained by selecting according to the fitness of the energy scheduling scheme individuals, which includes the following steps:

[0051] S461, selecting an energy scheduling scheme individual with the highest fitness value, and randomly selecting a number of characteristic gene positions from the energy scheduling scheme individual, and preparing a vaccine using information of the extracted characteristic gene positions;

[0052] S462, randomly selecting a number of individuals in the energy scheduling scheme for vaccination according to the vaccination probability, and mutating the information on the characteristic gene positions of the selected individuals in the energy scheduling scheme according to the information of the vaccine;

[0053] S463. Test the energy scheduling plan individuals that have been vaccinated. If their affinity is less than that of the parent energy scheduling plan individuals, the energy scheduling plan individuals will be replaced by the parent energy scheduling plan individuals. Conversely, if the affinity of the energy scheduling plan individuals after vaccination is greater than that of the parent energy scheduling plan individuals, the energy scheduling plan individuals will replace the parent energy scheduling plan individuals and enter the next generation population.

[0054] According to another aspect of the present invention, a multi-mode coordinated control system based on a wind, solar and energy storage new energy power station is provided, the system comprising an operation data acquisition module, an information prediction module, an energy scheduling scheme set recommendation module, an energy scheduling scheme determination module and a multi-mode coordinated control module;

[0055] The operation data acquisition module is used to obtain the real-time operation data of wind power, photovoltaic power generation and energy storage systems through external protocol access;

[0056] The information prediction module is used to use real-time operation data combined with weather data to predict the power grid load and wind power and photovoltaic power generation in the future period of time;

[0057] The energy scheduling solution set recommendation module is used to recommend energy scheduling solutions according to the predicted data using a collaborative recommendation algorithm to obtain an energy scheduling solution set;

[0058] The energy scheduling scheme determination module is used to determine the best energy scheduling scheme from the energy scheduling scheme set according to the power generation capacity, energy storage capacity and grid load demand of the power station by using an improved Bayesian optimization algorithm based on an immune mechanism;

[0059] The multi-mode coordinated control module is used to perform corresponding control operations on wind energy, solar energy and energy storage equipment according to the obtained optimal energy scheduling plan, so as to achieve efficient coordinated work of various energy sources.

[0060] Compared with the prior art, the present invention provides a multi-mode coordinated control method and system based on a wind, solar and energy storage new energy power station, which has the following beneficial effects:

[0061] (1) The present invention can accurately predict the grid load and wind power and photovoltaic power generation conditions in the future based on the real-time operation data of wind power, photovoltaic power generation and energy storage systems combined with weather data, and analyze and obtain a set of energy scheduling plans based on the prediction results, so that the best energy scheduling plan can be efficiently and quickly determined from the energy scheduling plan set according to the power generation capacity, energy storage capacity and grid load demand of the power station, and then the best energy scheduling plan can be used to accurately control wind energy, light energy and energy storage equipment, realize efficient coordination among various energy sources, better meet the grid load demand and maximize the energy utilization efficiency.

[0062] (2) By utilizing a collaborative recommendation algorithm to recommend a set of energy scheduling solutions based on the predicted data, a set of energy scheduling solutions corresponding to the target time or load demand can be recommended based on a similarity algorithm before determining the optimal energy scheduling solution. This allows the subsequent determination of the optimal energy scheduling solution to be obtained by analyzing only the recommended energy scheduling solution set. Since the solutions in the recommended set are relatively excellent candidate solutions, a search can be performed in a smaller set when determining the optimal energy scheduling solution. Compared with the traditional global search method, the present invention can effectively reduce the search space, improve the search efficiency, and find the optimal energy scheduling solution more quickly, thereby efficiently realizing the coordinated control of wind, solar and energy storage new energy power stations.

[0063] (3) Compared with traditional optimization algorithms, the present invention can determine the optimal energy scheduling plan according to the power generation capacity, energy storage capacity and grid load demand of the power station by utilizing an improved Bayesian optimization algorithm based on the immune mechanism, thereby effectively reducing the amount of calculation, shortening the operation time, and finding the optimal energy scheduling plan more quickly. By comprehensively considering multiple factors such as the power generation cost of the power station, environmental impact, grid load matching, grid demand satisfaction and energy consumption, a more optimized energy scheduling plan under multiple objectives can be found. Then, the operation of the wind, solar and storage new energy power station can be automatically controlled according to the determined optimal energy scheduling plan, reducing manual intervention and achieving efficient and continuous operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0065] Figure 1 It is a flow chart of a multi-mode coordinated control method based on a wind, solar and energy storage new energy power station according to an embodiment of the present invention. DETAILED DESCRIPTION

[0066] 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 only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0067] According to an embodiment of the present invention, a multi-mode coordinated control method and system based on a wind, solar and energy storage new energy power station are provided.

[0068] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1As shown, according to one embodiment of the present invention, a multi-mode coordinated control method and system based on a wind, solar and energy storage new energy power station is provided, and the method includes the following steps:

[0069] S1. Obtain real-time operation data of wind power, photovoltaic power generation and energy storage systems through external protocol access;

[0070] Among them, real-time operation data includes:

[0071] Wind power data: wind turbine speed, power generation, wind speed, wind direction and other information.

[0072] Photovoltaic power generation data: current, voltage, power of photovoltaic panels, as well as ambient temperature, light intensity and other information.

[0073] Energy storage system data: voltage, current, status (such as charging, discharging, standby, etc.), remaining storage energy and other information of energy storage equipment.

[0074] Other shared data: In addition to the above specific operating data, some shared data can also be obtained, such as the voltage, frequency, load and other information of the power grid;

[0075] S2. Use real-time operation data combined with weather data to predict the grid load and wind power and photovoltaic power generation in the future;

[0076] Grid load prediction: Common influencing parameters include historical load data, time (such as season, day of the week, time of day, etc.), weather factors (such as temperature, humidity, etc.), special events (such as holidays, large-scale events, etc.). These will affect the user's power consumption behavior, thereby affecting the grid load.

[0077] Prediction of wind power generation: The main influencing factors are climatic conditions such as wind speed, wind direction, air density, humidity, as well as technical parameters of wind turbines, real-time operating status, etc.

[0078] Prediction of photovoltaic power generation: The main influencing factors are weather factors such as solar radiation intensity, temperature, illumination time, cloud cover, rainfall, as well as the angle, orientation, cleanliness of the photovoltaic panels, and the configuration of the photovoltaic array.

[0079] The method of using real-time operation data combined with weather data to predict the power grid load and wind power and photovoltaic power generation in the future period includes the following steps:

[0080] S21. Obtain historical operation data of wind power generation, photovoltaic power generation, and energy storage systems and corresponding weather data from the database;

[0081] S22, using historical load data and the corresponding time and weather data to build a power grid load forecasting model, and using the trained power grid load forecasting model combined with real-time operation data and weather data to predict the power grid load in the future;

[0082] S23, constructing a wind power generation prediction model based on historical wind power generation operation data, and using the trained wind power generation prediction model in combination with real-time operation data and weather data to predict wind power generation data in the future;

[0083] S24. Use historical photovoltaic power generation operation data to build a photovoltaic power generation prediction model, and use the trained photovoltaic power generation prediction model in combination with real-time operation data and weather data to predict photovoltaic power generation data in the future.

[0084] Specifically, there are many models for predicting grid load, wind power and photovoltaic power generation, and different models can be used according to specific circumstances. The following are some commonly used models:

[0085] Time series models: such as the Autoregressive Integrated Moving Average Model (ARIMA), the Seasonal Autoregressive Integrated Moving Average Model (SARIMA), etc. This type of model is mainly used to predict data with time correlation.

[0086] Regression models: such as linear regression, multivariate linear regression, ridge regression, Lasso regression, etc. This type of model can be used to predict continuous numerical variables.

[0087] Machine learning models: such as support vector machine (SVM), random forest, gradient boosted decision tree (GBDT), etc. These models can process nonlinear and high-dimensional data.

[0088] Deep learning models: such as long short-term memory networks (LSTM), convolutional neural networks (CNN), neural networks, etc. These models can handle complex, nonlinear standard and non-standard problems.

[0089] Ensemble learning models: such as Bagging, Boosting, Stacking, etc. This type of model can improve the accuracy of prediction by integrating multiple models together.

[0090] Optimization algorithm model: such as genetic algorithm, particle swarm optimization, etc. This type of model is mainly used to solve optimization problems and can also be used for prediction.

[0091] When choosing a model, you need to consider factors such as the characteristics of the data, the complexity of the problem, and the limitations of computing resources. At the same time, the choice of model is not a one-time decision, and it requires continuous testing and optimization in practical applications.

[0092] S3, using a collaborative recommendation algorithm to recommend an energy scheduling solution according to the predicted data, and obtaining an energy scheduling solution set;

[0093] The method of using the collaborative recommendation algorithm to recommend an energy scheduling solution according to the predicted data to obtain an energy scheduling solution set includes the following steps:

[0094] S31. Obtain historical data and forecast data and perform preprocessing. Historical data includes energy data, load demand data and weather conditions. Forecast data includes grid load forecast data, wind power forecast data and photovoltaic power forecast data. Clean these data, process missing values, outliers and noise data, and perform some data conversion, such as normalization or standardization.

[0095] S32. In the energy scheduling scenario, "users" can be regarded as different time periods or different load demands, and "items" can be regarded as different energy types. Based on historical data and forecast data, a user-item scoring matrix is ​​constructed, in which each element represents the effect or benefit of using a certain energy in a certain time period or under a certain load demand. Specifically, a scoring matrix of time / load demand and energy type is constructed based on the acquired historical data and forecast data. The rows of this matrix can be different time periods or load demands, and the columns are different energy types. Each element of the matrix represents the efficiency or satisfaction of using a certain energy in a certain time period or load demand.

[0096] S33, using a clustering algorithm to calculate the scoring matrix of time / load demand and energy type, obtaining several different categories, and cutting the scoring matrix in the time / load demand dimension to obtain several scoring matrices of time / load demand and energy type; the purpose of this step is to find similar time / load demand patterns based on historical data for subsequent recommendations;

[0097] Specifically, the use of a clustering algorithm to calculate the scoring matrix of time / load demand and energy type to obtain a number of different categories, and cutting the scoring matrix in the time / load demand dimension to obtain a number of scoring matrices of time / load demand and energy type includes the following steps:

[0098] S331, taking each time / load demand and each energy type score in the time / load demand and energy type score matrix as a separate sample and sample feature, respectively, to construct a sample feature matrix;

[0099] Specifically, the time / load demand and energy type score matrix is ​​constructed as input data suitable for the clustering algorithm. Each time / load demand can be regarded as a sample, and the score of each energy type can be regarded as a sample feature to construct a sample feature matrix.

[0100] S332, setting the number of cluster centers k, and randomly selecting k samples from the sample feature matrix as initial cluster centers;

[0101] S333, based on the initial cluster center, using the Euclidean distance calculation method to calculate the distance between two time / load demand samples, and classifying the samples into the category with the closest distance;

[0102] S334, when a round of clustering is completed, a new center of each category is obtained, and the coordinates of the new cluster center are defined as the centroid of the category;

[0103] S335, repeat S333 and S334, when the distance between the latest cluster center and the original cluster center is less than or equal to the preset distance threshold, clustering is completed, and k different categories are obtained;

[0104] S336. Cut the scoring matrix in the time / load demand dimension, extract the scores corresponding to the samples of each category, and obtain k new scoring matrices.

[0105] S34. Use similarity measurement methods (such as cosine similarity, Pearson correlation coefficient, etc.) to find the neighboring time / load demands of the target time / load demands in each category according to the clustering results, and analyze to obtain the recommended energy type set for the time / load demands in this category; the purpose of this step is to find out which energy types have higher utilization efficiency or satisfaction under similar circumstances.

[0106] S35. Comprehensively recommend energy types in all categories, sort the predicted score values, and use a preset number of recommended energy types as a set of energy scheduling solutions recommended by the prediction data based on the sorting results from high to low.

[0107] Specifically, the calculation formula of the predicted score value is:

[0108]

[0109] Where P(u,i) represents the prediction score of target time / load demand u for energy type i;

[0110] Sim(u,v) represents the similarity between the target time / load demand u and the neighboring time / load demand v;

[0111] R(v,i) represents the actual score of the neighbor time / load demand v for energy type i.

[0112] S4, determining the best energy dispatching scheme from the energy dispatching scheme set by using an improved Bayesian optimization algorithm based on an immune mechanism according to the power generation capacity, energy storage capacity and grid load demand of the power station;

[0113] The method of determining the best energy scheduling scheme from the energy scheduling scheme set by using an improved Bayesian optimization algorithm based on an immune mechanism according to the power generation capacity, energy storage capacity and grid load demand of the power station includes the following steps:

[0114] S41, obtaining data on power generation capacity, energy storage capacity and grid load demand of the power station, encoding them, and creating an initial energy scheduling plan set;

[0115] S42, setting a fitness function based on the power generation cost, environmental impact, grid load matching, grid demand satisfaction and energy consumption of the power station;

[0116] Specifically, in this fitness function, the roles of various influencing factors are as follows:

[0117] Power generation cost: This refers to the total cost required to implement a certain energy scheduling plan, including fuel cost, maintenance cost, equipment depreciation cost, etc. The lower the power generation cost, the better the economic efficiency of the energy scheduling plan.

[0118] CO2 emissions: This refers to the amount of carbon dioxide emissions generated by implementing a certain energy scheduling plan, which is used to measure the environmental friendliness of the energy scheduling plan. The smaller the CO2 emissions, the smaller the impact of the energy scheduling plan on the environment.

[0119] Grid load matching: This refers to the degree of matching between the actual load of the grid and the load predicted in the energy dispatching plan. The higher the matching degree, the more accurate the energy dispatching plan is in predicting the load of the grid, and the better the stability of the grid operation.

[0120] Grid demand satisfaction: This refers to the degree of match between the actual power supply of the grid and the grid demand. The higher the demand satisfaction, the better the energy scheduling scheme can meet the load demand of the grid.

[0121] Energy consumption: This refers to the total amount of energy required to implement a certain energy scheduling solution. The lower the energy consumption, the higher the energy efficiency of the energy scheduling solution and the more energy-saving it is.

[0122] The calculation formula of the fitness function is:

[0123] y=ω1C1+ω2C2+ω3C3+ω4C4-ω5C5

[0124] ω1+ω2+ω3+ω4+ω5=1;

[0125] In the formula, y represents the fitness value, C1 represents the standardized carbon dioxide emissions, ω1 represents the weight of the standardized carbon dioxide emissions, C2 represents the standardized power generation cost, ω2 represents the weight of the standardized power generation cost, C3 represents the standardized grid load matching degree, ω3 represents the weight of the standardized grid load matching degree, C4 represents the standardized grid demand satisfaction, ω4 represents the weight of the standardized grid demand satisfaction, C5 represents the standardized energy consumption, and ω5 represents the weight of the standardized energy consumption.

[0126] Among them, standardized power generation cost = (power generation cost - minimum power generation cost) / (maximum power generation cost - minimum power generation cost);

[0127] Standardized CO2 emissions = (CO2 emissions - minimum CO2 emissions) / (maximum CO2 emissions - minimum CO2 emissions);

[0128] Standardized grid load matching degree = (grid load matching degree - minimum matching degree) / (maximum matching degree - minimum matching degree);

[0129] Standardized grid demand satisfaction = (actual power supply - minimum power supply) / (grid demand - minimum power supply);

[0130] Normalized energy consumption = (energy consumption - minimum energy consumption) / (maximum energy consumption - minimum energy consumption).

[0131] S43, establishing a Bayesian network for predicting the fitness of a new energy scheduling scheme by analyzing the initial energy scheduling scheme set;

[0132] S44, using the maximum likelihood estimation algorithm in combination with the energy scheduling scheme set to calculate the conditional probability of each node in the Bayesian network, and using the conditional probability of the node as a parameter of the Bayesian network, specifically including:

[0133] Calculate conditional probability: Use the maximum likelihood estimation algorithm to calculate the conditional probability of each node. For a node with a parent node, it is necessary to calculate the conditional probability of the node given the state of the parent node.

[0134] Specifically, if the parent of a node A is B, then the conditional probability of A P(A|B) can be calculated by the following formula: P(A|B) = P(A,B) / P(B);

[0135] Among them, P(A,B) is the probability of nodes A and B occurring at the same time, and P(B) is the probability of node B occurring. Both probabilities can be calculated by statistically analyzing the data in the energy scheduling solution set.

[0136] Determine parameters: Finally, we use the calculated conditional probability of each node as the parameters of the Bayesian network;

[0137] S45, sampling the Bayesian network according to the conditional probability of the Bayesian network to generate a new energy scheduling scheme individual;

[0138] S46, selecting the energy scheduling scheme individual with the highest fitness value to make a vaccine, inoculating the newly generated energy scheduling scheme individuals, and selecting the next generation of energy scheduling scheme sets according to the fitness of the energy scheduling scheme individuals;

[0139] Specifically, the process of selecting the energy scheduling scheme individual with the highest fitness value to prepare a vaccine, inoculating the newly generated energy scheduling scheme individual, and selecting the next generation energy scheduling scheme set according to the fitness of the energy scheduling scheme individual includes the following steps:

[0140] S461, selecting the energy scheduling scheme individual with the highest fitness value, and randomly selecting a number of characteristic gene positions (i.e., key information, such as specific energy ratio, scheduling time, etc.) from the energy scheduling scheme individual, and making the information of the extracted characteristic gene positions into a vaccine;

[0141] S462, randomly selecting a number of individuals in the energy scheduling scheme for vaccination according to the vaccination probability, and mutating the information on the characteristic gene positions of the selected individuals in the energy scheduling scheme according to the information of the vaccine;

[0142] S463. Test the energy scheduling plan individuals that have been vaccinated. If their affinity is less than that of the parent energy scheduling plan individuals, the energy scheduling plan individuals will be replaced by the parent energy scheduling plan individuals. Conversely, if the affinity of the energy scheduling plan individuals after vaccination is greater than that of the parent energy scheduling plan individuals, the energy scheduling plan individuals will replace the parent energy scheduling plan individuals and enter the next generation population.

[0143] S47, determine whether the number of iterations reaches the maximum number of iterations or whether the fitness of the new energy scheduling scheme individual reaches the preset threshold. If so, output the current energy scheduling scheme as the optimal energy scheduling scheme, otherwise, return to S42 and continue iterating.

[0144] S5. Carrying out corresponding control operations on wind energy, solar energy and energy storage equipment according to the obtained optimal energy scheduling plan is the key to achieving efficient coordinated work of various energy sources.

[0145] Among them, the corresponding control operation of wind energy, light energy and energy storage equipment according to the obtained optimal energy scheduling scheme is to achieve efficient coordination of various energy sources, including:

[0146] Wind energy control operation: adjust the operating status of wind turbines according to the optimal energy scheduling plan. For example, if the plan recommends increasing the utilization rate of wind energy, the speed of the wind turbine can be increased; if the plan recommends reducing the utilization rate of wind energy, the speed of the wind turbine can be reduced or part of the wind turbine can be shut down.

[0147] Light energy control operation: According to the optimal energy scheduling plan, adjust the operating state of the photovoltaic power generation system. For example, if the plan recommends improving the utilization rate of light energy, the maximum power point tracking (MPPT) technology can be used to optimize the working state of the photovoltaic array; if the plan recommends reducing the utilization rate of light energy, the power generation can be reduced by changing the tilt angle of the photovoltaic array or partially blocking the photovoltaic array.

[0148] Energy storage device control operation: According to the optimal energy scheduling plan, the charging and discharging strategy of the energy storage device is adjusted. For example, if the plan recommends that the energy storage device discharge to meet the grid load, the energy storage device can be controlled to discharge; if the plan recommends that the energy storage device charge to store excess wind energy or solar energy, the energy storage device can be controlled to charge.

[0149] Coordinated control: The control operations of wind energy, solar energy and energy storage equipment need to be coordinated to ensure the overall operation efficiency and stability of the power station. For example, when both wind energy and solar energy are sufficient, wind energy and solar energy can be used to generate electricity at the same time, and the energy storage equipment can be controlled to charge; when both wind energy and solar energy are insufficient, wind energy, solar energy and energy storage equipment can be used to discharge at the same time to meet the grid load.

[0150] According to one embodiment of the present invention, a multi-mode coordinated control system based on a wind, solar and energy storage new energy power station is provided, the system comprising an operation data acquisition module, an information prediction module, an energy scheduling scheme set recommendation module, an energy scheduling scheme determination module and a multi-mode coordinated control module;

[0151] The operation data acquisition module is used to obtain the real-time operation data of wind power, photovoltaic power generation and energy storage systems through external protocol access;

[0152] The information prediction module is used to use real-time operation data combined with weather data to predict the power grid load and wind power and photovoltaic power generation in the future period of time;

[0153] The energy scheduling solution set recommendation module is used to recommend energy scheduling solutions according to the predicted data using a collaborative recommendation algorithm to obtain an energy scheduling solution set;

[0154] The energy scheduling scheme determination module is used to determine the best energy scheduling scheme from the energy scheduling scheme set according to the power generation capacity, energy storage capacity and grid load demand of the power station by using an improved Bayesian optimization algorithm based on an immune mechanism;

[0155] The multi-mode coordinated control module is used to perform corresponding control operations on wind energy, solar energy and energy storage equipment according to the obtained optimal energy scheduling plan, so as to achieve efficient coordinated work of various energy sources.

[0156] To sum up, with the help of the above-mentioned technical scheme of the present invention, the grid load and wind power and photovoltaic power generation conditions in the future period can be accurately predicted based on the real-time operation data of wind power, photovoltaic power generation and energy storage systems combined with weather data, and an energy scheduling plan set can be analyzed based on the prediction results, so that the optimal energy scheduling plan can be efficiently and quickly determined from the energy scheduling plan set according to the power generation capacity, energy storage capacity and grid load demand of the power station, and then the optimal energy scheduling plan can be used to accurately control wind energy, light energy and energy storage equipment, to achieve efficient coordination between various energy sources, which can better meet the grid load demand while maximizing the energy utilization efficiency.

[0157] In addition, by utilizing a collaborative recommendation algorithm to recommend an energy scheduling solution set based on the predicted data, a corresponding energy scheduling solution set can be recommended for the target time or load demand based on a similarity algorithm before determining the optimal energy scheduling solution, so that when the optimal energy scheduling solution is subsequently determined, it only needs to be analyzed from the recommended energy scheduling solution set. Since the solutions in the recommended set are already relatively excellent candidate solutions, the optimal energy scheduling solution can be searched in a smaller set when determining the optimal energy scheduling solution. Compared with the traditional global search method, the present invention can effectively reduce the search space, improve search efficiency, and find the optimal energy scheduling solution more quickly, thereby efficiently realizing the coordinated control of wind, solar and energy storage new energy power stations.

[0158] In addition, compared with traditional optimization algorithms, the present invention can determine the optimal energy scheduling plan according to the power generation capacity, energy storage capacity and grid load demand of the power station by utilizing an improved Bayesian optimization algorithm based on the immune mechanism, thereby effectively reducing the amount of calculation, shortening the operation time, and finding the optimal energy scheduling plan more quickly. By comprehensively considering multiple factors such as the power generation cost of the power station, environmental impact, grid load matching, grid demand satisfaction and energy consumption, a more optimized energy scheduling plan under multiple objectives can be found. Then, the operation of the wind, solar and storage new energy power station can be automatically controlled according to the determined optimal energy scheduling plan, reducing manual intervention and achieving efficient and continuous operation.

[0159] The technical features of the above-mentioned embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification. A person of ordinary skill in the art can understand that all or part of the steps in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a program. The program can be stored in a computer-readable storage medium. When the program is executed, it includes the steps described in the above method, and the storage medium, such as ROM / RAM, a disk, an optical disk, etc.

[0160] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A multi-mode coordinated control method based on a wind, solar and energy storage new energy power station, characterized in that: The method comprises the following steps: S1. Obtain real-time operation data of wind power, photovoltaic power generation and energy storage systems through external protocol access; S2. Use real-time operation data combined with weather data to predict the grid load and wind power and photovoltaic power generation in the future; S3, using a collaborative recommendation algorithm to recommend an energy scheduling solution according to the predicted data, and obtaining an energy scheduling solution set; S4, determining the best energy dispatching scheme from the energy dispatching scheme set by using an improved Bayesian optimization algorithm based on an immune mechanism according to the power generation capacity, energy storage capacity and grid load demand of the power station; S5. Carrying out corresponding control operations on wind energy, solar energy and energy storage equipment according to the obtained optimal energy scheduling plan is the key to achieving efficient coordinated work of various energy sources.

2. According to claim 1, a multi-mode coordinated control method based on a wind, solar and energy storage new energy power station is characterized in that: The method of using real-time operation data combined with weather data to respectively predict the power grid load and wind power and photovoltaic power generation conditions in the future period includes the following steps: S21. Obtain historical operation data of wind power generation, photovoltaic power generation, and energy storage systems and corresponding weather data from the database; S22, using historical load data and corresponding time and weather data to build a power grid load prediction model, and using the trained power grid load prediction model to predict the power grid load in the future; S23, building a wind power generation prediction model based on historical wind power generation operation data, and using the trained wind power generation prediction model to predict wind power generation data in the future; S24. Use historical photovoltaic power generation operation data to build a photovoltaic power generation prediction model, and use the trained photovoltaic power generation prediction model to predict photovoltaic power generation data in the future.

3. A multi-mode coordinated control method based on a wind, solar and energy storage new energy power station according to claim 2, characterized in that: The power grid load prediction model, wind power generation prediction model and photovoltaic power generation prediction model are all any one of a time series model, a regression model, a machine learning model, a deep learning model or an integrated learning model.

4. The multi-mode coordinated control method based on a wind, solar and energy storage new energy power station according to claim 1 is characterized in that: The method of using the collaborative recommendation algorithm to recommend an energy scheduling solution according to the prediction data to obtain an energy scheduling solution set includes the following steps: S31, obtaining historical data and forecast data and performing preprocessing, the historical data including energy data, load demand data and weather conditions, and the forecast data including grid load forecast data, wind power generation forecast data and photovoltaic power generation forecast data; S32, constructing a scoring matrix of time / load demand and energy type based on the acquired historical data and forecast data; S33, using a clustering algorithm to calculate the scoring matrix of time / load demand and energy type to obtain several different categories, and cutting the scoring matrix in the time / load demand dimension to obtain several scoring matrices of time / load demand and energy type; S34, using a similarity measurement method to find the neighboring time / load demand of the target time / load demand in each category according to the clustering result, and analyzing to obtain a recommended energy type set of the time / load demand in the category; S35. Comprehensively recommend energy types in all categories, sort the predicted score values, and use a preset number of recommended energy types as a set of energy scheduling solutions recommended by the prediction data based on the sorting results from high to low.

5. A multi-mode coordinated control method based on a wind, solar and energy storage new energy power station according to claim 4, characterized in that: The method of calculating the scoring matrix of time / load demand and energy type by using a clustering algorithm to obtain a number of different categories and cutting the scoring matrix in the time / load demand dimension to obtain a number of scoring matrices of time / load demand and energy type includes the following steps: S331, taking each time / load demand and each energy type score in the time / load demand and energy type score matrix as a separate sample and sample feature, respectively, to construct a sample feature matrix; S332, setting the number k of cluster centers, and randomly selecting k samples from the sample feature matrix as initial cluster centers; S333, based on the initial cluster center, using the Euclidean distance calculation method to calculate the distance between two time / load demand samples, and classifying the samples into the category with the closest distance; S334, when a round of clustering is completed, a new center of each category is obtained, and the coordinates of the new cluster center are defined as the centroid of the category; S335, repeat S333 and S334, when the distance between the latest cluster center and the original cluster center is less than or equal to the preset distance threshold, clustering is completed, and k different categories are obtained; S336. Cut the scoring matrix in the time / load demand dimension, extract the scores corresponding to the samples of each category, and obtain k new scoring matrices.

6. A multi-mode coordinated control method based on a wind, solar and energy storage new energy power station according to claim 4, characterized in that: The calculation formula of the predicted score value is: Where P(u,i) represents the prediction score of target time / load demand u for energy type i; Sim(u,v) represents the similarity between the target time / load demand u and the neighbor time / load demand v; R(v,i) represents the actual score of the neighbor time / load demand v for energy type i.

7. The multi-mode coordinated control method based on a wind, solar and energy storage new energy power station according to claim 1 is characterized in that: The method of determining the best energy dispatching scheme from the energy dispatching scheme set by using an improved Bayesian optimization algorithm based on an immune mechanism according to the power generation capacity, energy storage capacity and grid load demand of the power station comprises the following steps: S41, obtaining data on power generation capacity, energy storage capacity and grid load demand of the power station, encoding them, and creating an initial energy scheduling plan set; S42, setting a fitness function based on the power generation cost, environmental impact, grid load matching, grid demand satisfaction and energy consumption of the power station; S43, establishing a Bayesian network for predicting the fitness of a new energy scheduling scheme by analyzing the initial energy scheduling scheme set; S44, using the maximum likelihood estimation algorithm in combination with the energy scheduling scheme set to calculate the conditional probability of each node in the Bayesian network, and using the conditional probability of the node as a parameter of the Bayesian network; S45, sampling the Bayesian network according to the conditional probability of the Bayesian network to generate a new energy scheduling scheme individual; S46, selecting the energy scheduling scheme individual with the highest fitness value to make a vaccine, inoculating the newly generated energy scheduling scheme individuals, and selecting the next generation of energy scheduling scheme sets according to the fitness of the energy scheduling scheme individuals; S47, determine whether the number of iterations reaches the maximum number of iterations or whether the fitness of the new energy scheduling scheme individual reaches the preset threshold. If so, output the current energy scheduling scheme as the optimal energy scheduling scheme, otherwise, return to S42 and continue iterating.

8. A multi-mode coordinated control method based on a wind, solar and energy storage new energy power station according to claim 7, characterized in that: The calculation formula of the fitness function is: y=ω1C1+ω2C2+ω3C3+ω4C4-ω5C5 ω1+ω2+ω3+ω4+ω5=1; In the formula, y represents the fitness value, C1 represents the standardized carbon dioxide emissions, ω1 represents the weight of the standardized carbon dioxide emissions, C2 represents the standardized power generation cost, ω2 represents the weight of the standardized power generation cost, C3 represents the standardized grid load matching degree, ω3 represents the weight of the standardized grid load matching degree, C4 represents the standardized grid demand satisfaction, ω4 represents the weight of the standardized grid demand satisfaction, C5 represents the standardized energy consumption, and ω5 represents the weight of the standardized energy consumption.

9. The multi-mode coordinated control method based on a wind, solar and energy storage new energy power station according to claim 7 is characterized in that: The process of selecting the energy scheduling scheme individual with the highest fitness value to prepare a vaccine, inoculating the newly generated energy scheduling scheme individual, and selecting the next generation energy scheduling scheme set according to the fitness of the energy scheduling scheme individual includes the following steps: S461, selecting an energy scheduling scheme individual with the highest fitness value, and randomly selecting a number of characteristic gene positions from the energy scheduling scheme individual, and preparing a vaccine using information of the extracted characteristic gene positions; S462, randomly selecting a number of individuals in the energy scheduling scheme for vaccination according to the vaccination probability, and mutating the information on the characteristic gene positions of the selected individuals in the energy scheduling scheme according to the information of the vaccine; S463. Test the energy scheduling plan individuals that have been vaccinated. If their affinity is less than that of the parent energy scheduling plan individuals, the energy scheduling plan individuals will be replaced by the parent energy scheduling plan individuals. Conversely, if the affinity of the energy scheduling plan individuals after vaccination is greater than that of the parent energy scheduling plan individuals, the energy scheduling plan individuals will replace the parent energy scheduling plan individuals and enter the next generation population.

10. A multi-mode coordinated control system based on a wind, solar and energy storage new energy power station, used to implement the steps of the multi-mode coordinated control method based on a wind, solar and energy storage new energy power station according to any one of claims 1 to 9, characterized in that: The system includes an operation data acquisition module, an information prediction module, an energy scheduling scheme set recommendation module, an energy scheduling scheme determination module and a multi-mode coordination control module; The operation data acquisition module is used to obtain the real-time operation data of wind power, photovoltaic power generation and energy storage systems through external protocol access; The information prediction module is used to use real-time operation data combined with weather data to predict the power grid load and wind power and photovoltaic power generation in the future period of time; The energy scheduling solution set recommendation module is used to recommend energy scheduling solutions according to the predicted data using a collaborative recommendation algorithm to obtain an energy scheduling solution set; The energy scheduling scheme determination module is used to determine the best energy scheduling scheme from the energy scheduling scheme set according to the power generation capacity, energy storage capacity and grid load demand of the power station by using an improved Bayesian optimization algorithm based on an immune mechanism; The multi-mode coordinated control module is used to perform corresponding control operations on wind energy, solar energy and energy storage equipment according to the obtained optimal energy scheduling plan, so as to achieve efficient coordinated work of various energy sources.

Images