A model prediction-based intelligent power grid energy optimization management method

Abstract

The application provides a model prediction-based intelligent power grid energy optimization efficient management method, which takes efficient energy management of a micro power grid as a research object and comprises the following steps: firstly, a mathematical model of each power grid node of a distributed power grid, including a power consumption end, a power supply end, an energy storage end and the whole micro power grid, is established; secondly, for the distributed grid-connected power generation model of unstable photovoltaic and wind clean energy, the future photovoltaic power generation and wind power generation are predicted through training and testing of an artificial neural network (ANN) so as to reduce the unstable factors brought by the clean power to the power grid; finally, the model prediction control technology is ingeniously utilized, at each sampling time, a model prediction optimization problem is established according to the minimum energy cost and carbon emission elements, each element execution strategy is solved, the energy cost paid by consumers is reduced in a rolling optimization optimal strategy mode, the peak pressure of power demand is reduced, the carbon emission is minimized, and the power consumption comfort of users is improved.

Description

A Model-Based Prediction-Based Smart Grid Energy Optimization Management Method Technical Field

[0001] This invention belongs to the field of smart grid energy management technology, and in particular relates to a model-based prediction-based method for efficient energy management in smart grids. Background Technology

[0002] Traditional power systems are inefficient due to their complete reliance on fossil fuels and the fact that centralized generation is located far from users. In such cases, power generation often requires long-distance transmission and distribution lines to reach users, incurring significant resources for system construction and maintenance, and resulting in substantial transmission losses. Consequently, centralized generation typically causes more environmental pollution than distributed generation technologies. Currently, emerging smart grids with distributed generation and consumption models are appearing. These smart grid energy management models, with their low transmission losses and ability to intelligently control electricity for consumers, are performing exceptionally well in electricity markets worldwide.

[0003] Distributed microgrids utilizing renewable energy sources reduce dependence on fossil fuels and mitigate high carbon emissions. Furthermore, microgrids can connect to smart homes through methods such as energy islanding and grid connection. In energy islanding mode, residential microgrids and commercial grids cannot initiate energy trading mechanisms. However, in grid-connected mode, microgrids can buy and sell electricity from external grids. Forecasting of unstable clean energy sources such as wind and solar power within microgrids is crucial; the selection of forecasting models typically depends on available data, the objectives of the model network mechanism, and energy planning operations. Machine learning models process large amounts of data, providing accurate predictive analysis. Artificial neural networks extract and model the correlations and characteristics of renewable energy sources, making accurate predictions of their future power generation.

[0004] Current approaches to energy management primarily utilize mathematical, cybernetics, and heuristic algorithms. Firstly, mathematical methods mainly employ mixed-integer linear programming to provide optimal solutions for tasks such as energy use and renewable resource management, achieving an optimal schedule under dynamic electrical constraints while maintaining user comfort. However, mathematical and deterministic methods suffer from system and computational complexity. Secondly, control methods have been explored. Some researchers have employed integrated and automated control of renewable energy sources, such as photovoltaic power plants and solid oxide fuel cells in smart grids. Energy management systems utilize proportional-integral and adaptive neuro-fuzzy inference technologies to effectively balance supply and demand in the grid. Other researchers have designed a model predictive control-based dynamic energy management system for connecting to grid-connected microgrids in residential communities, where data is collected from various electrical components via smart metering systems. However, control-based methods are computationally intensive, resulting in slow optimization. Finally, heuristic algorithms are employed, primarily utilizing a layered architecture of cloud computing and edge computing to provide intelligent, autonomous strategic decision-making based on massive amounts of information. In homes and the power grid, large-scale information collection, communication, processing, and control are achieved through cooperative energy management agents. Experimental results show that this agent-based solution has good application prospects in collaborative energy management.

[0005] While the aforementioned methods have achieved some success in energy management, researchers have not comprehensively considered key functionalities such as advanced metering infrastructure, forecasting, and two-way communication within smart grids. Furthermore, objectives such as electricity costs, carbon emissions, and grid peak ratios have not been addressed simultaneously in optimization. Therefore, an optimization technique is needed that considers user priorities, real-time electricity prices, and user comfort constraints to address the uncertainties of load and renewable energy generation. This technique aims to achieve efficient energy utilization, reduced energy costs, mitigate peak ratios, alleviate carbon emissions, and improve end-user satisfaction, thereby simultaneously satisfying both electricity suppliers and users. Summary of the Invention

[0006] To overcome the shortcomings and deficiencies of existing technologies, this invention introduces a highly efficient microgrid energy management model to systematically schedule loads and the charging and discharging of electric vehicles. The smart microgrid is equipped with controllable appliances, photovoltaic panels, wind turbine power generation, electrolyzers, and energy storage systems. The charging and discharging of electric vehicles reduces peak loads, peak-to-average power ratios, costs, energy costs, and carbon emissions from household appliances. The energy storage system is dispatched using real-time pricing. This achieves a balance between electricity demand and supply, and reduces energy costs, grid peak power ratios, carbon emissions, and user dissatisfaction, realizing a new low-carbon and environmentally friendly model.

[0007] The purpose of this invention is to establish a mathematical model of each grid node in a distributed power grid, including the power consumption end, the power supply end, the energy storage end, and the overall microgrid. Secondly, it clarifies the distributed grid-connected power generation model for unstable clean energy sources, such as photovoltaic power generation and wind power generation models, and uses artificial neural networks (ANNs) to predict future clean energy power generation. Finally, based on factors such as minimum energy cost and carbon emissions, a model prediction optimization problem is established, and the execution strategies for each factor are solved to reduce energy costs paid by consumers, alleviate peak electricity demand pressure, minimize carbon emissions, and improve user comfort, thereby achieving a new low-carbon and environmentally friendly model.

[0008] This research focuses on the efficient energy management of microgrids, and includes the following steps: First, establishing mathematical models of each grid node in the distributed grid, including the power consumption end, power supply end, energy storage end, and the overall microgrid; second, for the distributed grid-connected power generation model of unstable photovoltaic and wind power clean energy, predicting future photovoltaic and wind power generation through training and testing of artificial neural networks (ANNs) to reduce the instability factors brought to the grid by clean power sources; finally, cleverly utilizing model predictive control technology, at each sampling moment, establishing a model predictive optimization problem based on minimum energy cost and carbon emission factors, solving for the execution strategy of each factor, and reducing the energy cost paid by consumers, alleviating the peak pressure of electricity demand, minimizing carbon emissions, and improving the user's electricity comfort by continuously optimizing the optimal strategy.

[0009] The technical solution adopted by this invention to solve its technical problem is:

[0010] A model-based predictive method for efficient energy optimization and management of smart grids, characterized by the following steps:

[0011] Step S1: Establish a mathematical model of each grid node of the distributed power grid, including the power consumption end, the power supply end, the energy storage end, and the overall microgrid;

[0012] Step S2: Define the unstable distributed grid-connected power generation model of photovoltaic and wind power, and predict the future power generation of clean energy through artificial neural networks (ANN);

[0013] Step S3: Based on the minimum energy cost, the ratio of peak to average grid emissions, and carbon emission factors, establish a model to predict the optimization problem and solve for the execution strategy of each factor.

[0014] Furthermore, in step S1, a mathematical model is performed on the smart microgrid equipped with household appliances, photovoltaic arrays, electrolyzers, wind power, electric vehicles, and energy storage devices, and a model is established to predict and optimize the problem based on the minimum energy cost, the ratio of peak to average grid values, and carbon emission factors.

[0015] Furthermore, in step S2, the solar radiation and wind speed time series are used as model inputs to remove irrelevant and redundant features. The selected sample feature inputs are divided into training and test datasets. Then, the training and test sets are used to predict the solar radiation and wind speed over the next day. The error between the artificial neural network prediction and the actual observation is calculated. In the optimization stage, the enhanced differential evolution algorithm (EDE) is used to further minimize the error.

[0016] Furthermore, in step S3, an optimization problem is established considering the objective functions of energy cost, small grid fluctuations, and carbon emissions, as well as the constraints of the system. The optimal strategy is obtained by minimizing the objective function.

[0017] Furthermore, step S1 specifically includes the following steps:

[0018] Step 1.1: Adaptive scheduling method based on real-time pricing signals. When the energy management system receives a low-price signal from real-time pricing, it immediately schedules and controls electrical appliances. Consumers schedule smart appliances within specified time intervals to avoid high costs due to operation during peak hours. The user's activation probability function is defined as:

[0019] (1)

[0020] Where S represents intelligent electrical equipment. It represents a specific hour of the day. It indicates a day of the week. It refers to a certain day of the year. Indicates the time step for prediction calculation. The standard deviation represents the social random factor. Indicates the probability of use during a season. This represents the start probability factor per hour. Represents the probability of random social use. Indicates the scaling factor;

[0021] Step 1.2: Establish a microgrid model; a microgrid should include at least electrical appliances, photovoltaic panels, wind power, electrolyzers, electric vehicles, and energy storage devices; the net energy produced by each device in the microgrid at time t. The energy generated within the predicted time period T:

[0022] (2)

[0023] The mathematical model for the hourly power generation of wind power is expressed as:

[0024] (3)

[0025] in, This represents the amount of wind power generated in t hours. This represents the wind energy conversion rate value. This represents the area scanned by the turbine blades of a wind turbine generator. This represents the wind speed in hour t. Given air density, wind power generation satisfies the following constraints:

[0026]

[0027] in, and These represent the speed at which the wind enters the blade and the speed at which it exits the blade, respectively.

[0028] The mathematical model for photovoltaic power generation is expressed as follows:

[0029] (4)

[0030] in, This indicates the amount of photovoltaic power generated per hour. and These represent the area of ​​the photovoltaic panel that receives sunlight and its efficiency, respectively. and Let represent the external temperature of the photovoltaic panel and the amount of solar radiation at hour t, respectively.

[0031] The voltage relationship between the electrodes of an electrolytic cell is expressed as follows:

[0032] (5)

[0033] in, The ohm resistance parameter is represented by s, the overvoltage parameter by C, and the electrode contact area by C. It is an overvoltage on the electrode. Indicates electrode current, Standard voltage for reversible batteries;

[0034] The stored energy of a static energy storage system is expressed as:

[0035] (6)

[0036] in, This represents the amount of electricity that can be stored in hour t. It refers to the efficiency of battery energy storage. and These represent the charging and discharging states during the time period t. Indicates the duration of the interval;

[0037] The energy stored in a mobile energy storage system is expressed as follows:

[0038] (7)

[0039] in, Indicates the arrival time of the electric vehicle. Indicates the time of departure of the electric force. This represents the charge / discharge amount of the electric vehicle in hour t. It is in electric vehicle mode. It is in the charging state of the electric vehicle. It is in a discharge state. It is in an idle state.

[0040] Furthermore, step S2 specifically includes the following steps:

[0041] The prediction model consists of three parts: (1) feature selector; (2) predictor; (3) optimizer;

[0042] (1) Feature selection stage: The prediction model based on information interaction technology uses solar radiation and wind speed time series as model input; the input is sorted through information interaction technology, and the sorted unit features are passed to the redundancy filter to remove irrelevant and redundant features; then, the selected sample features are input into training and test datasets.

[0043] (2) In the prediction phase, training and test sets are used to predict solar radiation and wind speed over the next full day. The artificial neural network consists of three layers: an input layer, a hidden layer, and an output layer. Each layer of artificial neurons (ANs) uses the sigmoid function as the activation function, as shown in equation (8).

[0044] (8)

[0045] Where S is the solar radiation intensity and wind speed input signal, parameter β is the steepness of the activation function, and b represents the signal deviation value; the prediction model designed above can be trained to learn and accurately estimate the electricity generated in the future; in this embodiment, since the multivariate autoregressive rule has higher convergence than the benchmark learning rule, the designed prediction framework adopts a supervised learning method to learn from time series analysis;

[0046] The prediction framework is trained using a training set and validated using a test set. The accuracy of the predictions is then determined by comparing them with actual ground observations. The mean absolute percentage error (MAPE) is used as the validation ratio between the artificial neural network's predicted output and the actual observed values, and can be expressed as:

[0047] (9)

[0048] in, Represented as actual observed values, Let n represent the predicted solar radiation intensity and wind speed, where n represents the number of samples used. Indicates the number of days observed; the predicted values ​​from the artificial neural network predictor are provided to the optimization stage of the prediction model to further minimize the error;

[0049] (3) Optimization stage: Calculate the error between the predicted value and the actual observed value of the artificial neural network. In the optimization stage, the enhanced differential evolution algorithm (EDE) is used to further minimize the error.

[0050] Furthermore, step S3 specifically includes the following steps:

[0051] The energy management system receives a demand response signal and broadcasts it to users in advance. Users then send their electricity consumption patterns back to the energy management system. The system then plans consumer electricity consumption strategies and electric vehicle charging / discharging strategies to reduce energy costs, mitigate grid fluctuations, and decrease carbon emissions. Therefore, the objective function is modeled as an optimization function to minimize these objectives. Each optimization objective is mathematically expressed according to the formula for the overall energy management problem.

[0052] (1) Energy consumption cost

[0053] The energy consumption cost, without considering the real-time electricity price signal of the microgrid, is expressed as follows:

[0054] (10)

[0055] in, This indicates the hourly power consumption of the microgrid. This represents the average hourly electricity price. This indicates the forecast time period; each node of the microgrid is equipped with unstable renewable energy, battery energy storage, and electric vehicle energy storage. The required input electricity for each day is expressed as follows:

[0056] (11)

[0057] in, This indicates the amount of electricity generated by renewable energy sources. This indicates the amount of electricity released by the energy storage system of an electric vehicle. This indicates the amount of electricity released by the battery storage. This indicates the introduction of electricity, while This indicates that no electricity is imported; the hourly and daily electricity costs for users considering the microgrid are expressed as follows:

[0058] (12)

[0059] (13)

[0060] in, This indicates the daily electricity cost; This indicates the electricity consumption per hour. This indicates the average hourly electricity price;

[0061] (2) The ratio of peak power to average power grid value;

[0062] The ratio of peak power consumption to average power consumption is expressed as:

[0063] (14)

[0064] in, This indicates the highest hourly electricity consumption within a day. Hourly electricity demand;

[0065] (3) Carbon emissions

[0066] The mathematical model for carbon emissions is expressed as follows:

[0067] (15)

[0068] in, Indicates carbon emissions. This represents the average electricity price. This indicates the price per kilowatt-hour. Indicates the carbon factor of electricity emissions;

[0069] (4) Construction of the objective optimization function

[0070] The optimization problem, which considers minimum energy cost, carbon emissions, peak to average ratio, and user comfort, is as follows:

[0071] (16)

[0072]

[0073] in, This indicates the energy storage level of electric vehicles upon arrival at charging stations. Indicates the energy stored in electric vehicles. This indicates the maximum discharge capacity of the electric vehicle battery. This indicates the minimum discharge capacity of an electric vehicle battery. This indicates the amount of electricity generated by the electrolytic cell. This indicates the maximum amount of solar radiation that the solar panel can receive. and These represent the minimum and maximum capacities of the energy storage system, respectively.

[0074] The optimal strategy sequence for each node in the power grid is obtained by solving the problem. The first strategy element is taken and fed back to each node in the power grid for control, so as to complete the power grid strategy control operation.

[0075] Furthermore, a model-based predictive smart grid energy optimization and efficient management system is characterized by being based on a computer system, including a processor, memory, and a computer program stored in the memory, wherein when the processor executes the program instructions, it implements the method described above.

[0076] A computer-readable storage medium is characterized in that it stores computer program instructions thereon, which, when loaded and executed by a processor, are capable of implementing the method described above.

[0077] Compared with the prior art, the present invention and its preferred embodiments have the following beneficial effects:

[0078] This approach can reduce energy costs for consumers, alleviate peak electricity demand pressure, minimize carbon emissions, and improve user comfort. It establishes mathematical models for each node in a distributed power grid, including the consumer, supplier, energy storage, and microgrid as a whole. It clarifies distributed grid-connected generation models for unstable clean energy sources, such as photovoltaic and wind power generation models, and uses artificial neural networks (ANNs) to predict future clean energy generation. Finally, based on factors such as minimum energy cost and carbon emissions, a model prediction optimization problem is established, and the execution strategies for each factor are solved. This effectively achieves efficient energy management and provides accurate predictions of the power generation efficiency of unstable renewable energy sources. Attached Figure Description

[0079] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0080] The accompanying drawings, which form part of this application, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0081] Figure 1 is a flowchart of the rolling optimization execution process according to an embodiment of the present invention;

[0082] Figure 2 is a schematic diagram of the system microgrid model of the background technology of this invention;

[0083] Figure 3 is a flowchart of the enhanced differential algorithm according to an embodiment of the present invention;

[0084] Figure 4 is an operation planning diagram of various smart home appliances according to an embodiment of the present invention;

[0085] Figure 5 is a comparison chart of electricity costs for various scheduling methods according to embodiments of the present invention;

[0086] Figure 6 is a comparison chart of the error rates of the improved artificial neural network in an embodiment of the present invention. Detailed Implementation

[0087] To make the features and advantages of this patent more apparent and understandable, specific embodiments are provided below for detailed explanation:

[0088] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used in this specification have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0089] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0090] The following is a further detailed description of the embodiment with reference to the accompanying drawings:

[0091] As shown in Figure 1, this embodiment provides a model-based prediction-based method for efficient energy management of smart grids, including the following steps:

[0092] Step 1: Establish a mathematical model of each grid node in the distributed power grid, including the power consumption end, the power supply end, the energy storage end, and the overall microgrid.

[0093] Step 1.1: Modeling the electrical appliances operating in the smart home, as shown in Figure 2, where each device is arranged according to consumer demand. An adaptive scheduling method based on real-time pricing signals is used. When the energy management system receives a low-price signal from the real-time pricing system, it immediately schedules and controls the appliances. Consumers schedule smart appliances within specified time intervals to avoid high costs due to operation during peak hours. The user's activation probability function is then defined as:

[0094] (1)

[0095] Where S represents intelligent electrical equipment. It represents a specific hour of the day. It indicates a day of the week. It refers to a certain day of the year. Indicates the time step for prediction calculation. The standard deviation represents the social random factor. Indicates the probability of use during a season. This represents the start probability factor per hour. Represents the probability of random social use. This represents the scaling factor.

[0096] Step 1.2: Establish the microgrid model. The microgrid includes electrical appliances, photovoltaic panels, wind power, electrolyzers, electric vehicles, and energy storage devices. The net energy produced by each device in the microgrid at time t is... The energy generated within the predicted time period T:

[0097] (2)

[0098] Step 1.2.1: Wind Power Generation

[0099] The hourly power generation of wind power is mainly determined by wind speed, and its mathematical model is expressed as:

[0100] (3)

[0101] in, This represents the amount of wind power generated in t hours. This represents the wind energy conversion rate value. This represents the area scanned by the turbine blades of a wind turbine generator. This represents the wind speed in hour t. Let represent air density. Wind power generation is directly affected by wind speed; that is, the higher the wind speed, the higher the power generation, and vice versa. However, the above model satisfies the following constraints:

[0102]

[0103] in, and These represent the speed at which the wind enters the blade and the speed at which it exits the blade, respectively.

[0104] Step 1.2.2: Photovoltaic power generation

[0105] The power generation of photovoltaic panels is directly affected by the intensity of solar radiation, and its mathematical model can be expressed as:

[0106] (4)

[0107] in, This indicates the amount of photovoltaic power generated per hour. and These represent the area of ​​the photovoltaic panel that receives sunlight and its efficiency, respectively. and Let represent the external temperature of the photovoltaic panel and the amount of solar radiation at hour t, respectively.

[0108] Step 1.2.3: Electrolyzer power generation

[0109] This embodiment employs an electrochemical reaction process in the electrolytic cell model. Furthermore, a maximum temperature of 100°C is considered. The voltage relationship between the electrodes of the electrolytic cell is expressed as follows:

[0110] (5)

[0111] in, The ohm resistance parameter is represented by s, the overvoltage parameter by C, and the electrode contact area by C. It is an overvoltage on the electrode. Indicates electrode current, Standard voltage for reversible batteries;

[0112] Step 1.3: Establish an energy storage system model. Energy storage systems include static storage and mobile storage systems, which are equipped with unstable renewable energy sources in a microgrid. They store electricity during peak hours (as load) and release energy during periods of low prices (as power source).

[0113] (1) Static energy storage system: The static energy storage system considered in this embodiment is mainly a battery energy storage system with a storage capacity of 3 kWh. Its stored energy can be expressed as:

[0114] (6)

[0115] in, This represents the amount of electricity that can be stored in hour t. It refers to the efficiency of battery energy storage. and These represent the charging and discharging states during the time period t. Indicates the duration of the interval.

[0116] (2) Mobile Energy Storage System: The mobile energy storage system considered in this embodiment is an electric vehicle, and the energy storage capacity of an electric vehicle is expressed as:

[0117] (7)

[0118] in, Indicates the arrival time of the electric vehicle. Indicates the time of departure of the electric force. This represents the charge / discharge amount of the electric vehicle in hour t. It is in electric vehicle mode. It is in the charging state of the electric vehicle. It is in a discharge state. It is in an idle state.

[0119] Step 2: Define the distributed grid-connected power generation model for unstable clean energy sources, such as photovoltaic power generation model and wind power generation model, and predict the future power generation of clean energy sources through artificial neural networks (ANN).

[0120] First, a differential algorithm is used to optimize the artificial neural network framework to predict solar radiation and wind speed, as shown in Figure 3, in order to effectively estimate power generation. The proposed prediction model consists of three parts: (1) feature selector; (2) predictor; (3) optimizer.

[0121] (1) Feature selection stage: The prediction model based on information interaction technology uses solar radiation and wind speed time series as model inputs. The inputs are sorted through information interaction technology, and the sorted unit features are passed to the redundancy filter to remove irrelevant and redundant features. Then, the selected sample features are input into training and test datasets.

[0122] (2) The prediction stage is based on artificial neural networks. Specifically, training and test sets are used to predict solar radiation and wind speed over the next full day. The artificial neural network consists of three layers: an input layer, a hidden layer, and an output layer. Each layer's artificial neurons (ANs) use the sigmoid function as the activation function, as shown in equation (8):

[0123] (8)

[0124] Where S represents the solar radiation intensity and wind speed input signals, β is the steepness of the activation function, and b represents the signal deviation. The prediction model designed above can be trained to learn and accurately estimate future electricity generation. In this embodiment, because the multivariate autoregressive rule has higher convergence than the baseline learning rule, the designed prediction framework uses a supervised learning method to learn from time series analysis.

[0125] The prediction framework is trained using a training set and validated on a test set to predict future values. The accuracy of the predictions is then determined by comparing the obtained predictions with actual ground observations. The Mean Absolute Percentage Error (MAPE) is used as the degree of validation between the artificial neural network's predicted output and the actual observed values, and can be expressed as:

[0126] (9)

[0127] in, Represented as actual observed values, Let n represent the predicted solar radiation intensity and wind speed, where n represents the number of samples used. The number of days observed is used to provide the predicted values ​​(solar radiation and wind speed) from the artificial neural network predictor to the optimization stage of the prediction model, so as to further minimize the error.

[0128] (3) Optimization stage: Calculate the error between the predicted value and the actual observed value of the artificial neural network. In the optimization stage, the enhanced differential evolution (EDE) algorithm is used to further minimize the error.

[0129] Step 3: Based on the factors of minimum energy cost, carbon emissions, and minimal grid fluctuations, establish a model to predict and optimize the problem, solve for the execution strategies of each factor, reduce the energy costs paid by consumers, alleviate the peak pressure of electricity demand, minimize carbon emissions, and improve the user's electricity comfort.

[0130] In energy management, the main objectives are energy cost reduction, reduced grid fluctuations, and lower carbon emissions, achieved through proactive consumer / end-user participation in distributed generation and demand response projects. To this end, the energy management system receives a demand response signal and broadcasts it to users in advance. Users then send their electricity consumption patterns back to the energy management system. The system will then orchestrate consumer electricity consumption strategies and electric vehicle charging / discharging strategies to reduce energy costs, mitigate grid fluctuations, and decrease carbon emissions. Therefore, the proposed objective function is modeled as an optimization function to minimize these objectives. Each optimization objective will be mathematically expressed according to the formula for the overall energy management problem:

[0131] (1) Energy consumption cost

[0132] Energy cost refers to the fee a consumer pays a utility provider for the electricity consumed within a specific period. Energy cost is determined by real-time electricity price signals provided by the power grid. Without considering the real-time electricity price signals of microgrids, energy consumption cost is expressed as:

[0133] (10)

[0134] in, This indicates the hourly power consumption of the microgrid. This represents the average hourly electricity price. This represents the predicted time period. Each node in the microgrid is equipped with unstable renewable energy, battery storage, and electric vehicle storage. The required daily power input can be expressed as:

[0135] (11)

[0136] in, This indicates the amount of electricity generated by renewable energy sources. This indicates the amount of electricity released by the energy storage system of an electric vehicle. This indicates the amount of electricity released by the battery's energy storage. This indicates the introduction of electricity, while This indicates that no electricity is imported. The hourly and daily electricity costs for users considering a microgrid can be expressed as:

[0137] (12)

[0138] (13)

[0139] in, This indicates the daily electricity cost. This indicates the electricity consumption per hour. This indicates the average hourly electricity price.

[0140] (2) Ratio of peak power to average power grid

[0141] The ratio of peak power consumption to average power consumption can be expressed as:

[0142] (14)

[0143] in, This indicates the highest hourly electricity consumption within a day. Hourly electricity demand.

[0144] (3) Carbon emissions

[0145] The mathematical model for carbon emissions can be expressed as:

[0146] (15)

[0147] in, Indicates carbon emissions. This represents the average electricity price. This indicates the price per kilowatt-hour. This indicates the carbon factor emitted by electricity.

[0148] (4) Construction of the objective optimization function

[0149] To achieve our desired goals—minimum energy cost, carbon emissions, peak-to-average ratio, and user comfort—the optimization problem can be designed as follows:

[0150] (16)

[0151]

[0152] in, This indicates the energy storage level of electric vehicles upon arrival at charging stations. Indicates the energy stored in electric vehicles. This indicates the maximum discharge capacity of the electric vehicle battery. This indicates the minimum discharge capacity of an electric vehicle battery. This indicates the amount of electricity generated by the electrolytic cell. This indicates the maximum amount of solar radiation that the solar panel can receive. and These represent the minimum and maximum capacity of the energy storage system, respectively.

[0153] The optimal strategy sequence for each node in the power grid is obtained by solving the problem. The first strategy element is taken and fed back to each node in the power grid for control, thus completing the power grid strategy control operation.

[0154] Step 4: Simulation Comparison and Analysis

[0155] In the given simulation case, the model-based prediction-based smart grid energy optimization and efficient management method is compared with the traditional mixed-integer linear programming. Four performance indicators are used in the case: energy cost, peak to average ratio, carbon emissions, and waiting time / delay. The performance of the proposed model is evaluated compared with existing models. The proposed system model is developed for residential smart home design with three types of loads: electrical control appliances, temperature control appliances, and lighting appliances. Furthermore, a predictive power generation framework based on solar irradiance and wind speed is implemented to accurately estimate power generation for effective energy management. Simulations were conducted for the following scenarios: (I) energy management without a microgrid; (II) 24-hour energy management with a microgrid. The scheduling plan for each appliance after running the proposed model-based prediction-based smart grid energy optimization and efficient management method is shown in Figure 4; the electricity costs generated by various scheduling methods are shown in Figure 5; and the error rate of predicting unstable clean energy using an improved artificial neural network is shown in Figure 6.

[0156] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0157] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more blocks of the flowchart illustrations and / or one or more blocks of the block diagrams.

[0158] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more block diagrams.

[0159] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, such that the instructions, which execute on the computer or other programmable apparatus, provide steps for implementing the functions specified in one or more flowcharts and / or one or more block diagrams.

[0160] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

[0161] This patent is not limited to the above-described preferred embodiments. Anyone can derive other forms of model-based prediction-based smart grid energy optimization and efficient management methods based on this patent. All equivalent changes and modifications made within the scope of this patent application shall fall within the scope of this patent.

Claims

1. A smart grid energy optimization management method based on model prediction, characterized in that, The process includes the following steps: Step S1: Establishing a mathematical model of each grid node in the distributed power grid, including the power consumption end, power supply end, energy storage end, and the overall microgrid; Step S1 specifically includes the following steps: Step 1.1: An adaptive scheduling method based on real-time pricing signals. When the energy management system receives a low-price signal from real-time pricing, it will immediately schedule and control electrical appliances; consumers can schedule smart appliances within a specified time interval to avoid high costs caused by operation during peak hours; the user's activation probability function is defined as: (1) Wherein, S represents intelligent electrical equipment, It represents a specific hour of the day. It indicates a day of the week. It refers to a certain day of the year. Indicates the time step for prediction calculation. The standard deviation represents the social random factor. Indicates the probability of use during a season. This represents the start probability factor per hour. Represents the probability of random social use. The scaling factor is represented; Step 1.2: Establish the microgrid model; A microgrid includes at least electrical appliances, photovoltaic panels, wind power, electrolyzers, electric vehicles, and energy storage devices; The net energy produced by each device in the microgrid at time t. The energy generated within the predicted time period T: (2) The mathematical model for the hourly power generation of wind power is expressed as: (3) Among them, This represents the amount of wind power generated in t hours. This represents the wind energy conversion rate value. This represents the area scanned by the turbine blades of a wind turbine generator. This represents the wind speed in hour t. Given air density, wind power generation satisfies the following constraints: in, and These represent the velocities of the wind turbine entering and exiting the blades, respectively; the mathematical model for photovoltaic power generation is expressed as: (4) Among them, This indicates the amount of photovoltaic power generated per hour. and These represent the area of ​​the photovoltaic panel that receives sunlight and its efficiency, respectively. and Let represent the external temperature of the photovoltaic panel and the amount of solar radiation at hour t, respectively; the voltage relationship between the electrodes of the electrolytic cell is expressed as: (5) Among them, The ohm resistance parameter is represented by s, the overvoltage parameter by C, and the electrode contact area by C. It is an overvoltage on the electrode. Indicates electrode current, The standard voltage of a reversible battery; the energy stored in a static energy storage system is expressed as: (6) Among them, This represents the amount of electricity that can be stored in hour t. It refers to the efficiency of battery energy storage. and These represent the charging and discharging states during the time interval t. The duration of the interval is indicated; the stored energy of the mobile energy storage system is expressed as: (7) Among them, Indicates the arrival time of the electric vehicle. Indicates the time of departure of the electric force. This represents the charge / discharge amount of the electric vehicle in hour t. It is in electric vehicle mode. It is in the charging state of the electric vehicle. It is in a discharge state. It is in an idle state; Step S2: Clarify the unstable photovoltaic and wind power distributed grid-connected power generation model, and predict the future clean energy power generation through artificial neural network ANN; Step S3: Establish a model prediction optimization problem based on minimum energy cost, grid peak and average ratio, and carbon emission factors, and solve the execution strategy of each factor; Step S3 specifically includes the following steps: The energy management system receives a demand response signal and broadcasts it to the user in advance; The user sends its electricity consumption mode response to the energy management system; The energy management system arranges the consumer's electricity consumption strategy and the electric vehicle's charging and discharging strategy to reduce energy costs, mitigate grid fluctuations, and reduce carbon emissions. Therefore, the objective function is modeled as an optimization function to minimize the above objectives; Each optimization objective is mathematically expressed according to the formula of the entire energy management problem: (1) Energy consumption cost without considering the real-time electricity price signal of the microgrid is expressed as: (10) Among them, This indicates the hourly power consumption of the microgrid. This represents the average hourly electricity price. This indicates the forecast time period; each node of the microgrid is equipped with unstable renewable energy, battery energy storage, and electric vehicle energy storage. The required input electricity for each day is expressed as follows: (11) Among them, This indicates the amount of electricity generated by renewable energy sources. This indicates the amount of electricity released by the energy storage system of an electric vehicle. This indicates the amount of electricity released by the battery storage. This indicates the introduction of electricity, while This indicates that no electricity is imported; the hourly and daily electricity costs for users considering the microgrid are expressed as follows: （12） (13) Among them, This indicates the daily electricity cost; This indicates the electricity consumption per hour. (1) The average electricity price per hour; (2) The ratio of peak power consumption to average power consumption; The ratio of peak power consumption to average power consumption is expressed as: (14) Among them, This indicates the highest hourly electricity consumption within a day. (3) Hourly electricity demand; (4) Carbon emissions. The mathematical model for carbon emissions is as follows: (15) Among them, Indicates carbon emissions. This represents the average electricity price. This indicates the price per kilowatt-hour. (4) The objective optimization function is constructed to minimize energy cost, carbon emissions, peak to average ratio, and user comfort. The optimization problem is: （16） in, This indicates the energy storage level of electric vehicles upon arrival at charging stations. Indicates the energy stored in electric vehicles. This indicates the maximum discharge capacity of the electric vehicle battery. This indicates the minimum discharge capacity of an electric vehicle battery. This indicates the amount of electricity generated by the electrolytic cell. This indicates the maximum amount of solar radiation that the solar panel can receive. and Let represent the minimum and maximum capacity of the energy storage system, respectively. The optimal strategy sequence of each node in the power grid is obtained by solving the problem. The first strategy element is taken and fed back to each node in the power grid for control, so as to complete the power grid strategy control operation.

2. The smart grid energy optimization management method based on model prediction according to claim 1, characterized in that: In step S1, a mathematical model is performed on a smart microgrid equipped with household appliances, photovoltaic arrays, electrolyzers, wind power, electric vehicles, and energy storage devices. The model is then used to predict and optimize the problem based on the minimum energy cost, the ratio of peak to average grid values, and carbon emission factors.

3. The smart grid energy optimization management method based on model prediction according to claim 1, characterized in that: In step S2, the solar radiation and wind speed time series are used as model inputs to remove irrelevant and redundant features. The selected sample feature inputs are divided into training and test datasets. Then, the training and test sets are used to predict the solar radiation and wind speed over the next day. The error between the artificial neural network prediction and the actual observation is calculated. In the optimization stage, the enhanced differential evolution algorithm (EDE) is used to further minimize the error.

4. The smart grid energy optimization management method based on model prediction according to claim 1, characterized in that: In step S3, an optimization problem is established considering the objective functions of energy cost, small grid fluctuations, and carbon emissions, as well as the constraints of the system. The optimal strategy is obtained by minimizing the objective function.

5. The smart grid energy optimization management method based on model prediction according to claim 1, characterized in that: Step S2 specifically includes the following steps: The prediction model consists of three parts: (1) feature selector; (2) predictor; (3) optimizer; (1) Feature selection stage: The prediction model based on information interaction technology uses solar radiation and wind speed time series as model input; the input is sorted through information interaction technology, and the sorted unit features are passed to the redundancy filter to remove irrelevant and redundant features; then, the selected sample feature input is divided into training and test datasets; (2) Prediction stage: The training and test sets are used to predict solar radiation and wind speed within the time range of the next day; The artificial neural network consists of three layers, namely the input layer, hidden layer and output layer; Among them, the artificial neurons ANs of each layer are activated by the sigmoid function, as shown in the following formula (8): (8) Where S is the solar radiation intensity and wind speed input signal, parameter β is the steepness of the activation function, and b represents the signal deviation value; the above prediction model is trained to learn and accurately estimate the future electricity generation; since the multivariate autoregressive rule has higher convergence than the benchmark learning rule, the designed prediction framework adopts a supervised learning method to learn from time series analysis; the prediction framework is trained using a training set, and the prediction framework is validated using a test set to predict future values, and the accuracy is obtained by comparing the obtained prediction results with the actual ground observation results; the mean absolute percentage error (MAPE) is used as the degree of validation between the artificial neural network prediction output and the actual observation value, and is expressed as: (9) Among them, Represented as actual observed values, Let n represent the predicted solar radiation intensity and wind speed, where n represents the number of samples used. The number of days observed is indicated; the predicted value of the artificial neural network predictor is provided to the optimization stage of the prediction model so that the error is further minimized; (3) Optimization stage: calculate the error between the predicted value of the artificial neural network and the actual observed value, and use the enhanced differential evolution algorithm EDE to further minimize the error in the optimization stage.

6. A smart grid energy optimization management system based on model prediction, characterized in that, It is based on a computer system, including a processor, memory, and a computer program stored in the memory, which, when the processor runs the program, implements the method as described in any one of claims 1-5.

7. A computer-readable storage medium, characterized in that, It stores computer program instructions that, when loaded and executed by a processor, enable the implementation of the method as described in any one of claims 1-5.

Images