Power grid stability optimization system and method based on intelligent control

Abstract

The invention relates to the technical field of power grid optimization, in particular to a power grid stability optimization system and method based on intelligent control, and the system comprises a multi-dimensional prediction module which obtains data of environment variables, time dimensions and equipment states, and then carries out the collection of the data to obtain multi-dimensional data; when the system is used, various data sources such as environment variables, time dimensions and equipment states are fused, so that complex fluctuation characteristics of new energy power generation can be accurately captured, and particularly, the sensitivity of short-time fluctuation is greatly improved; meanwhile, more accurate load prediction data is provided for a power grid management system, so that the load fluctuation of the power grid is effectively reduced by adjusting user requirements in time, the stability of the power grid is improved, power storage is facilitated when the generated power is excessive, power is supplemented through an energy storage system when the generated power is insufficient, the phenomena of power gap and excess are avoided, and the power utilization rate is improved. And the load management of the power grid is optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid optimization, and in particular to a power grid stability optimization system and method based on intelligent control. Background Art

[0002] The power grid is an electricity transmission and distribution system consisting of power plants, substations, transmission lines, distribution networks and users. Its core function is to transmit the generated electric energy to different areas through high-voltage transmission lines, and then distribute it to end users through low-voltage distribution networks after reducing the voltage at substations.

[0003] The patent application number is 202210329469.4, which states in the specification that "the present invention belongs to the field of electric power, and in particular to a power grid optimization method and system, including obtaining a first power generation power at a power generation end and a first power consumption at a load end, constructing a first prediction model with power generation power as input, power consumption power as output, power limit of a power grid transmission line as a condition, and the overall line loss rate as an optimization condition; analyzing the line loss rate and the overall line loss rate of the transmission line path from the power generation end to the load end; configuring the transmission line path to train and converge the first prediction model, and obtaining a second prediction model for optimizing the power grid. By obtaining the first prediction The model can coordinate the relationship between the power generation end and the power transmission load end of the power grid, and reduce the overall line loss rate of the power grid; by configuring the transmission line path to train and converge the first prediction model, a second prediction model for optimizing the power grid is obtained, which can improve the accuracy of the prediction model and help to further reduce the overall line loss rate of the power grid. Although the above technology has the advantages of improving the accuracy of the prediction model and reducing the overall line loss rate of the power grid, it is generally not flexible when facing large fluctuations in power generation, especially when new energy sources such as wind energy and solar energy have large fluctuations in power generation, which will lead to uneven load distribution in the power grid.

[0004] In summary, developing a power grid stability optimization system and method based on intelligent control is still a key issue that needs to be urgently solved in the field of power grid optimization technology. Summary of the invention

[0005] The purpose of the present invention is to solve the problem that although the above-mentioned technology in the prior art has the advantages of improving the accuracy of the prediction model and reducing the overall line loss rate of the power grid, it has general flexibility when facing large fluctuations in power generation, especially new energy sources such as wind energy and solar energy, which have large fluctuations in power generation during power generation, resulting in uneven load distribution in the power grid. The present invention provides a power grid stability optimization system and method based on intelligent control.

[0006] To achieve the above object, the present invention provides the following technical solutions: The present invention provides a power grid stability optimization system based on intelligent control, comprising: a multi-dimensional prediction module, which obtains data of environmental variables, time dimensions and equipment status respectively and then aggregates them to obtain multi-dimensional data; The automatic adjustment module is connected with the multi-dimensional prediction module and automatically formulates charging and discharging strategies based on the demand response mechanism, energy storage system and virtual power plant; The monitoring feedback module is connected with the multi-dimensional prediction module and the automatic adjustment module, and uses edge computing and digital twin technology to establish a global simulation model of power grid operation; The learning optimization module is connected with the monitoring feedback module to introduce autonomous learning capabilities for continuous optimization of power grid operation strategies.

[0007] Furthermore, the multidimensional prediction module includes: Based on the hybrid deep learning model, the time dependency and spatial correlation of the power generation fluctuation of the power grid are captured according to the multi-dimensional data, and the power fluctuation trend of the power grid is predicted, thereby obtaining prediction data; The automatic adjustment module includes: Based on the demand response mechanism, the user electricity demand and the grid operation parameters are adjusted in real time according to the predicted data, and then the balance between the load side and the power generation side is dynamically adjusted. By combining the energy storage system and the virtual power plant, power storage is performed during the peak period of grid power generation, and power is released during the trough period of grid power generation. The predicted data provides data support for the energy storage system and the virtual power plant to formulate charging and discharging strategies.

[0008] Furthermore, the monitoring feedback module includes: The global simulation model of power grid operation is used to capture the correlation between power grid generation and load fluctuation in real time to obtain operation result data. When the operation result data deviates from the predicted data, the operation strategy of the automatic adjustment module is adjusted based on the intelligent feedback mechanism; The learning optimization module includes: A closed-loop learning model is designed based on a reinforcement learning algorithm. According to the operation result data, the optimal coordination data between power generation, power consumption and load fluctuations is dynamically explored through a reward and punishment mechanism. After the optimal coordination data is obtained, the power grid operation strategy is optimized.

[0009] Furthermore, the workflow of the multidimensional prediction module is: Through sensors and historical data platforms, we collect environmental variables such as wind speed, solar radiation and temperature, time dimensions such as peak, valley and cycle characteristics of power demand, and equipment status such as energy storage system load and virtual power plant output status in real time, and perform denoising, standardization and normalization on the collected data to form a multi-dimensional data set. ,in is an environment variable, is the time dimension, is the device state, and the convolutional neural network is Perform feature extraction and generate local feature matrix , ,in are the weight matrices and biases, Represents multidimensional data, is the activation function, which inputs the extracted features into the long short-term memory network to capture the temporal dependency. ,in represents a recurrent unit with weights, is the hidden state, Represents multidimensional data, is a discrete index, and then the variational autoencoder is used to generate the time-space joint prediction distribution. ,in It means that given the observed data Under the condition of hidden variables The distribution of is the likelihood function that represents the probability of a given latent variable Observe the data under the condition The probability distribution of It is Observation Dimension The mean of It is Observation Dimension The variance of is a standard Gaussian distribution representing a single observation The mean is The variance is The normal distribution of is a hidden variable The expected value of It represents its covariance structure in high-dimensional space, and finally predicts the future power generation fluctuation trend based on the time distribution. ,in is the power set representing future predictions, It means at the time The predicted power value, is the total number of time steps representing the forecast.

[0010] Furthermore, the workflow of the automatic adjustment module is: Obtain the prediction data provided by the multidimensional prediction module , use the demand response mechanism to compare the predicted power generation data with the current load power, adjust the load, and balance the load by performing peak-shaving and valley-filling operations through the energy storage charging and discharging formula. The load adjustment formula is: ,in It is at the moment load changes or power loss, It is a coefficient used to adjust the impact of the difference between the generated power and the load power on the load change. It's time The power generation capacity, It's time Load power, energy storage charging and discharging formula: ,in It's time The change in energy storage, is the charging efficiency, It's time The excess power, is the discharge efficiency, is the load power, is the power generation, That is, when the power generation is greater than the load power, the excess power will be stored and the energy storage system will work in charging mode. That is, when the power generation is less than the load power, the energy storage system needs to discharge to supplement the missing power.

[0011] Furthermore, the workflow of the monitoring feedback module is: The global simulation model of power grid operation is used to capture the relationship between power grid generation and load fluctuation in real time to obtain operation result data. When the operation result data deviates from the predicted data, the operation strategy of the automatic adjustment module is adjusted based on the intelligent feedback mechanism. The deviation capture formula is: ,in is the error between the actual value and the predicted result, is the actual value, is the predicted value, and the intelligent feedback correction formula is: ,in is the value after error correction, is the original parameter value before adjustment. is the adjustment factor, is the difference between the actual value and the predicted value.

[0012] Furthermore, the workflow of the learning optimization module is: The closed-loop learning model accepts the operation result data and prediction bias , through the reward and punishment mechanism, dynamically explore the optimal coordinated data between power generation, power consumption and load fluctuations, and optimize the power grid operation strategy after obtaining the optimal coordinated data. The reward function formula is: ,in is the value obtained by the reinforcement learning algorithm, is the adjustment factor, is a summation symbol, is the power change of a generating unit or load, Is to take the absolute value and learn the update formula: ,in is the value function expressed in state Next Action The expected return, is the update operator, Is the current state indicating that the system is The state of the moment, is the current action, is the learning rate, It’s an instant reward. is the discount factor, is the maximum value of all possible actions in the next state, is the next state.

[0013] On the other hand, the present invention also provides an intelligently controlled inertial support grid method, which comprises the following steps: S1, respectively obtain the data of environmental variables, time dimension and device status and then aggregate them to obtain multi-dimensional data; S2, connected with the multi-dimensional prediction module, and automatically formulates charging and discharging strategies based on demand response mechanism, energy storage system and virtual power plant; S3, connected with the multi-dimensional prediction module and the automatic adjustment module, using edge computing and digital twin technology to establish a global simulation model of power grid operation; S4, connected with the monitoring feedback module, introduces autonomous learning capability for continuous optimization of power grid operation strategy.

[0014] Furthermore, in step S1, the method of respectively acquiring the data of the environmental variables, the time dimension and the device status and then aggregating them to obtain multi-dimensional data is as follows: Based on the hybrid deep learning model, the time dependency and spatial correlation of the power generation fluctuation of the power grid are captured according to the multi-dimensional data, and the power fluctuation trend of the power grid is predicted, thereby obtaining prediction data; In step S2, the method of automatically formulating a charging and discharging strategy based on the demand response mechanism, energy storage system and virtual power plant in connection with the multi-dimensional prediction module is as follows: Based on the demand response mechanism, the user electricity demand and the grid operation parameters are adjusted in real time according to the predicted data, and then the balance between the load side and the power generation side is dynamically adjusted. By combining the energy storage system and the virtual power plant, power storage is performed during the peak period of grid power generation, and power is released during the trough period of grid power generation. The predicted data provides data support for the energy storage system and the virtual power plant to formulate charging and discharging strategies.

[0015] Furthermore, in step S3, the method of establishing a global simulation model of power grid operation by connecting with the multidimensional prediction module and the automatic adjustment module and utilizing edge computing and digital twin technology is as follows: The global simulation model of power grid operation is used to capture the correlation between power grid generation and load fluctuation in real time to obtain operation result data. When the operation result data deviates from the predicted data, the operation strategy of the automatic adjustment module is adjusted based on the intelligent feedback mechanism; In step S4, the monitoring feedback module is connected to introduce autonomous learning capabilities, and the method for continuously optimizing the power grid operation strategy is: A closed-loop learning model is designed based on a reinforcement learning algorithm. According to the operation result data, the optimal coordination data between power generation, power consumption and load fluctuations is dynamically explored through a reward and punishment mechanism. After the optimal coordination data is obtained, the power grid operation strategy is optimized.

[0016] Beneficial Effects Compared with the known public technology, the technical solution provided by the present invention has the following beneficial effects: When the present invention is in use, it is beneficial to accurately capture the complex fluctuation characteristics of renewable energy power generation by integrating multiple data sources such as environmental variables, time dimensions and equipment status, especially greatly improving the sensitivity of short-term fluctuations. At the same time, it provides more accurate load forecast data for the power grid management system, which is beneficial to timely adjust user needs, effectively reduce the load fluctuation of the power grid, improve the stability of the power grid, and facilitate power storage when the power generation power is in excess, and supplement power through the energy storage system when the power generation is insufficient, avoiding power gaps and surpluses, and optimizing the load management of the power grid.

[0017] When in use, the present invention is conducive to quickly responding to emergencies in the operation of the power grid, ensuring the stability of the power grid, flexibly adjusting the balance between power generation and load, improving the adaptability and response ability of the power grid, reducing human intervention, enhancing the automatic adjustment ability of the system, reducing interference from human factors, and improving adjustment efficiency. The learning process is guided by a reward and punishment mechanism, so that the power grid can cope with complex and dynamically changing environments, and the operation strategy of the power grid is always in the optimal state. It can better cope with uncertain factors such as fluctuations in new energy power generation and load changes, which is conducive to improving the flexibility and stability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A system diagram of a power grid stability optimization system based on intelligent control according to the present invention; Figure 2 The present invention is a flow chart of an intelligently controlled inertial support grid method. DETAILED DESCRIPTION

[0019] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0020] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but includes other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0021] The present invention is further described in detail below in conjunction with the accompanying drawings: Embodiment 1: like Figure 1 As shown, the present invention provides a power grid stability optimization system based on intelligent control, including: a multi-dimensional prediction module, which obtains data of environmental variables, time dimensions and equipment status respectively and then aggregates them to obtain multi-dimensional data; The multi-dimensional prediction module includes: Based on the hybrid deep learning model, which is a type of model that combines multiple deep learning architectures with traditional machine learning algorithms, optimization algorithms, time series and dynamic processing algorithms, the hybrid deep learning model captures the time dependency and spatial correlation of power generation fluctuations of the power grid according to the multi-dimensional data, and predicts the power fluctuation trend of the power grid, thereby obtaining prediction data; The workflow of the multidimensional prediction module is: Through sensors and historical data platforms, we collect environmental variables such as wind speed, solar radiation and temperature, time dimensions such as peak, valley and cycle characteristics of power demand, and equipment status such as energy storage system load and virtual power plant output status in real time, and perform denoising, standardization and normalization on the collected data to form a multi-dimensional data set. ,in is an environment variable, is the time dimension, is the device state, and the convolutional neural network is Perform feature extraction and generate local feature matrix , ,in are the weight matrices and biases, Represents multidimensional data, is the activation function, which inputs the extracted features into the long short-term memory network to capture the temporal dependency. ,in represents a recurrent unit with weights, is the hidden state, Represents multidimensional data, is a discrete index, and then uses the variational autoencoder to generate the time-space joint prediction distribution. The variational autoencoder is a generative deep learning framework based on a probabilistic model, which is used to learn the low-dimensional potential representation of data and generate new data. ,in It means that given the observed data Under the condition of hidden variables The distribution of is the likelihood function that represents the probability of a given latent variable Observe the data under the condition The probability distribution of It is Observation Dimension The mean of It is Observation Dimension The variance of is a standard Gaussian distribution representing a single observation The mean is The variance is The normal distribution of is a hidden variable The expected value of It represents its covariance structure in high-dimensional space, and finally predicts the future power generation fluctuation trend based on the time distribution. ,in is the power set representing future predictions, It means at the time The predicted power value, is the total number of time steps representing the prediction; Specifically, through the deployed sensors and historical data platforms, multi-dimensional data required for power grid operation are collected in real time, including environmental variables such as wind speed, solar radiation, and temperature, time dimensions such as peak and trough periods of the power grid, and the cyclical characteristics of electricity demand, and equipment status such as energy storage system load and virtual power plant output status. The collected multi-dimensional data are denoised, standardized, and normalized to remove outliers and noise and ensure the quality of the input data.

[0022] The automatic adjustment module is connected with the multi-dimensional prediction module and automatically formulates charging and discharging strategies based on the demand response mechanism, energy storage system and virtual power plant; The automatic adjustment module includes: Based on the demand response mechanism, the user's electricity demand and the grid operation parameters are adjusted in real time according to the forecast data, thereby dynamically adjusting the balance between the load side and the power generation side. By combining the energy storage system and the virtual power plant, power storage is performed during the peak period of grid power generation, and power is released during the trough period of grid power generation. According to the forecast data, data support is provided for the energy storage system and the virtual power plant to formulate charging and discharging strategies; The workflow of the automatic adjustment module is: Obtain the prediction data provided by the multidimensional prediction module , use the demand response mechanism to compare the predicted power generation data with the current load power, perform load regulation, and balance the load by performing peak shaving and valley filling operations through the energy storage charging and discharging formula. Peak shaving and valley filling operations optimize power use and balance supply and demand fluctuations by reducing the peak load of the power system (peak shaving) and increasing the load during the valley period (valley filling). The load regulation formula is: ,in It is at the moment load changes or power loss, It is a coefficient used to adjust the impact of the difference between the generated power and the load power on the load change. It's time The power generation capacity, It's time Load power, energy storage charging and discharging formula: ,in It's time The change in energy storage, is the charging efficiency, It's time The excess power, is the discharge efficiency, is the load power, is the power generation, That is, when the power generation is greater than the load power, the excess power will be stored and the energy storage system will work in charging mode. That is, when the power generation is less than the load power, the energy storage system needs to discharge to supplement the missing power; Specifically, the automatic adjustment module first obtains real-time forecast data from the multidimensional forecast module. The forecast data mainly includes the power generation and load power in the next 2-4 hours. It uses the demand response mechanism to automatically adjust the user's electricity demand according to the difference between the predicted power generation and load power to balance the difference between the power generation side and the load side. Specifically, if the predicted power generation is greater than the load power, the system will send an adjustment signal to the user to reduce the power demand, such as reducing industrial loads and adjusting the operation of household appliances; if the predicted load power is greater than the power generation, the system will encourage the user to increase electricity consumption, or start the energy storage system for supplementation, which is conducive to timely adjustment of user demand, effectively reducing the load fluctuation of the power grid, improving the stability of the power grid, and facilitating power storage when the power generation is in excess, and supplementing electricity through the energy storage system when the power generation is insufficient, avoiding power shortages and surpluses, and optimizing the load management of the power grid.

[0023] The monitoring feedback module is connected with the multi-dimensional prediction module and the automatic adjustment module, and uses edge computing and digital twin technology to establish a global simulation model of power grid operation; The monitoring feedback module includes: The global simulation model is a mathematical model that integrates multiple subsystems and variables and is used to capture the relationship between power generation and load fluctuations in real time to obtain operation result data. When the operation result data deviates from the predicted data, the operation strategy of the automatic adjustment module is adjusted based on the intelligent feedback mechanism. The workflow of the monitoring feedback module is: The global simulation model of power grid operation is used to capture the relationship between power grid generation and load fluctuation in real time to obtain operation result data. When the operation result data deviates from the predicted data, the operation strategy of the automatic adjustment module is adjusted based on the intelligent feedback mechanism. The deviation capture formula is: ,in is the error between the actual value and the predicted result, is the actual value, is the predicted value, and the intelligent feedback correction formula is: ,in is the value after error correction, is the original parameter value before adjustment. is the adjustment factor, is the difference between the actual value and the predicted value; Specifically, the monitoring and feedback module compares the real-time data of the power grid operation with the predicted data, and uses the global simulation model to capture the relationship between the various parts of the power grid based on the real-time data, and simulates the operation status of the power grid under different working conditions. The deviation capture is mainly used to judge the error between the actual operation data and the predicted data. When the captured deviation exceeds a certain threshold, the intelligent feedback mechanism will correct the automatic adjustment module. After the deviation correction is completed, the adjusted parameters will be fed back to the automatic adjustment module. The automatic adjustment module will re-perform load adjustment, energy storage charging and discharging and other operations according to the corrected values ​​to balance the difference between load and power generation, which is conducive to quickly responding to emergencies in the operation of the power grid, ensuring the stability of the power grid, and flexibly adjusting the balance between power generation and load, thereby improving the adaptability and response capabilities of the power grid, reducing human intervention, enhancing the system's automatic adjustment capabilities, reducing interference from human factors, and improving adjustment efficiency.

[0024] The learning optimization module is connected with the monitoring feedback module to introduce autonomous learning capabilities for continuous optimization of power grid operation strategies; The learning optimization module includes: Design a closed-loop learning model based on the reinforcement learning algorithm. The reinforcement learning algorithm is a type of algorithm that learns how to make decisions through interaction with the environment. The closed-loop learning model is a learning system that continuously obtains real-time feedback and adjusts behavior based on the feedback. Based on the operation result data, the optimal coordination data between power generation, power consumption and load fluctuations is dynamically explored through a reward and punishment mechanism. After obtaining the optimal coordination data, the power grid operation strategy is optimized; The workflow of the learning optimization module is: The closed-loop learning model accepts the operation result data and prediction bias , through the reward and punishment mechanism, dynamically explore the optimal coordinated data between power generation, power consumption and load fluctuations, and optimize the power grid operation strategy after obtaining the optimal coordinated data. The reward function formula is: ,in is the value obtained by the reinforcement learning algorithm, is the adjustment factor, is a summation symbol, is the power change of a generating unit or load, Is to take the absolute value and learn the update formula: ,in is the value function expressed in state Next Action The expected return, is the update operator, Is the current state indicating that the system is The state of the moment, is the current action, is the learning rate, It’s an instant reward. is the discount factor, is the maximum value of all possible actions in the next state, is the next state; Specifically, the learning optimization module first receives the operation result data and prediction deviation provided by the monitoring feedback module. The operation result data includes the state data of power generation, load, energy storage, etc. in the power grid; the prediction deviation is the difference between the actual operation state of the power grid and the prediction result. Based on these data, the closed-loop learning model can evaluate the current operation state of the power grid and adjust the strategy according to historical data, and guide the learning process through the reward and punishment mechanism. Specifically, when the power grid strategy successfully reduces the deviation and maintains stable operation, the system will give rewards, and when the system has an imbalance between power generation and load or the power grid is unstable, it will be punished, so that the power grid can cope with complex and dynamically changing environments, so that the operation strategy of the power grid is always in the optimal state, and can better cope with uncertain factors such as fluctuations in new energy power generation and load changes, which is conducive to improving the flexibility and stability of the system.

[0025] Embodiment 2: like Figure 2 As shown, embodiment 2 provides an intelligently controlled inertial support grid method, which includes the following steps: S1, respectively obtain the data of environmental variables, time dimension and device status and then aggregate them to obtain multi-dimensional data; S2, connected with the multi-dimensional prediction module, and automatically formulates charging and discharging strategies based on demand response mechanism, energy storage system and virtual power plant; S3, connected with the multi-dimensional prediction module and the automatic adjustment module, using edge computing and digital twin technology to establish a global simulation model of power grid operation; S4, connected with the monitoring feedback module, introduces autonomous learning capability for continuous optimization of power grid operation strategy.

[0026] Furthermore, in step S1, the method of respectively acquiring the data of the environmental variables, the time dimension and the device status and then aggregating them to obtain multi-dimensional data is as follows: Based on the hybrid deep learning model, the time dependency and spatial correlation of the power generation fluctuation of the power grid are captured according to the multi-dimensional data, and the power fluctuation trend of the power grid is predicted, thereby obtaining prediction data; In step S2, the method of automatically formulating a charging and discharging strategy based on the demand response mechanism, energy storage system and virtual power plant in connection with the multi-dimensional prediction module is as follows: Based on the demand response mechanism, the user electricity demand and the grid operation parameters are adjusted in real time according to the predicted data, and then the balance between the load side and the power generation side is dynamically adjusted. By combining the energy storage system and the virtual power plant, power storage is performed during the peak period of grid power generation, and power is released during the trough period of grid power generation. The predicted data provides data support for the energy storage system and the virtual power plant to formulate charging and discharging strategies.

[0027] Furthermore, in step S3, the method of establishing a global simulation model of power grid operation by connecting with the multidimensional prediction module and the automatic adjustment module and utilizing edge computing and digital twin technology is as follows: The global simulation model of power grid operation is used to capture the correlation between power grid generation and load fluctuation in real time to obtain operation result data. When the operation result data deviates from the predicted data, the operation strategy of the automatic adjustment module is adjusted based on the intelligent feedback mechanism; In step S4, the monitoring feedback module is connected to introduce autonomous learning capabilities, and the method for continuously optimizing the power grid operation strategy is: A closed-loop learning model is designed based on a reinforcement learning algorithm. According to the operation result data, the optimal coordination data between power generation, power consumption and load fluctuations is dynamically explored through a reward and punishment mechanism. After the optimal coordination data is obtained, the power grid operation strategy is optimized.

[0028] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A power grid stability optimization system based on intelligent control, characterized in that: include: The multi-dimensional prediction module obtains the data of environmental variables, time dimensions and equipment status respectively and then aggregates them to obtain multi-dimensional data; The automatic adjustment module is connected with the multi-dimensional prediction module and automatically formulates charging and discharging strategies based on the demand response mechanism, energy storage system and virtual power plant; The monitoring feedback module is connected with the multi-dimensional prediction module and the automatic adjustment module, and uses edge computing and digital twin technology to establish a global simulation model of power grid operation; The learning optimization module is connected with the monitoring feedback module to introduce autonomous learning capabilities for continuous optimization of power grid operation strategies.

2. The power grid stability optimization system based on intelligent control according to claim 1, characterized in that: The multi-dimensional prediction module includes: Based on the hybrid deep learning model, the time dependency and spatial correlation of the power generation fluctuation of the power grid are captured according to the multi-dimensional data, and the power fluctuation trend of the power grid is predicted, thereby obtaining prediction data; The automatic adjustment module includes: Based on the demand response mechanism, the user electricity demand and the grid operation parameters are adjusted in real time according to the predicted data, and then the balance between the load side and the power generation side is dynamically adjusted. By combining the energy storage system and the virtual power plant, power storage is performed during the peak period of grid power generation, and power is released during the trough period of grid power generation. The predicted data provides data support for the energy storage system and the virtual power plant to formulate charging and discharging strategies.

3. The power grid stability optimization system based on intelligent control according to claim 2 is characterized in that: The monitoring feedback module includes: The global simulation model of power grid operation is used to capture the correlation between power grid generation and load fluctuation in real time to obtain operation result data. When the operation result data deviates from the predicted data, the operation strategy of the automatic adjustment module is adjusted based on the intelligent feedback mechanism; The learning optimization module includes: A closed-loop learning model is designed based on a reinforcement learning algorithm. According to the operation result data, the optimal coordination data between power generation, power consumption and load fluctuations is dynamically explored through a reward and punishment mechanism. After the optimal coordination data is obtained, the power grid operation strategy is optimized.

4. The power grid stability optimization system based on intelligent control according to claim 3 is characterized in that: The workflow of the multidimensional prediction module is: Through sensors and historical data platforms, we collect environmental variables such as wind speed, solar radiation and temperature, time dimensions such as peak, valley and cycle characteristics of power demand, and equipment status such as energy storage system load and virtual power plant output status in real time, and perform denoising, standardization and normalization on the collected data to form a multi-dimensional data set. ,in is an environment variable, is the time dimension, is the device state, the convolutional neural network is Perform feature extraction and generate local feature matrix , ,in are the weight matrices and biases, Represents multidimensional data, is the activation function, which inputs the extracted features into the long short-term memory network to capture the temporal dependency. ,in represents a recurrent unit with weights, is the hidden state, Represents multidimensional data, is a discrete index, and then the variational autoencoder is used to generate the time-space joint prediction distribution. ,in It means that given the observed data Under the condition of hidden variables The distribution of is the likelihood function that represents the probability of a given latent variable Observe the data under the condition The probability distribution of It is Observation Dimension The mean of It is Observation Dimension The variance of is a standard Gaussian distribution representing a single observation The mean is The variance is The normal distribution of is a hidden variable The expected value of It represents its covariance structure in high-dimensional space, and finally predicts the future power generation fluctuation trend based on the time distribution. ,in is the power set representing future predictions, It means at the time The predicted power value, is the total number of time steps representing the forecast.

5. The power grid stability optimization system based on intelligent control according to claim 4 is characterized in that: The workflow of the automatic adjustment module is: Obtain the prediction data provided by the multidimensional prediction module , use the demand response mechanism to compare the predicted power generation data with the current load power, adjust the load, and balance the load by performing peak-shaving and valley-filling operations through the energy storage charging and discharging formula. The load adjustment formula is: ,in It is at the moment load changes or power loss, It is a coefficient used to adjust the impact of the difference between the generated power and the load power on the load change. It's time The power generation capacity, It's time Load power, energy storage charging and discharging formula: ,in It's time The change in energy storage, is the charging efficiency, It's time The excess power, is the discharge efficiency, is the load power, is the power generation, That is, when the power generation is greater than the load power, the excess power will be stored and the energy storage system will work in charging mode. That is, when the power generation is less than the load power, the energy storage system needs to discharge to supplement the missing power.

6. A power grid stability optimization system based on intelligent control according to claim 5, characterized in that: The workflow of the monitoring feedback module is: The global simulation model of power grid operation is used to capture the relationship between power grid generation and load fluctuation in real time to obtain operation result data. When the operation result data deviates from the predicted data, the operation strategy of the automatic adjustment module is adjusted based on the intelligent feedback mechanism. The deviation capture formula is: ,in is the error between the actual value and the predicted result, is the actual value, is the predicted value, and the intelligent feedback correction formula is: ,in is the value after error correction, is the original parameter value before adjustment. is the adjustment factor, is the difference between the actual value and the predicted value.

7. The power grid stability optimization system based on intelligent control according to claim 6, characterized in that: The workflow of the learning optimization module is: The closed-loop learning model accepts the operation result data and prediction bias , through the reward and punishment mechanism, dynamically explore the optimal coordinated data between power generation, power consumption and load fluctuations, and optimize the power grid operation strategy after obtaining the optimal coordinated data. The reward function formula is: ,in is the value obtained by the reinforcement learning algorithm, is the adjustment factor, is a summation symbol, is the power change of a generating unit or load, Is to take the absolute value and learn the update formula: ,in is the value function expressed in state Next Action The expected return, is the update operator, Is the current state indicating that the system is The state of the moment, is the current action, is the learning rate, It’s an instant reward. is the discount factor, is the maximum value of all possible actions in the next state, is the next state.

8. A method for optimizing power grid stability based on intelligent control, based on a system for optimizing power grid stability based on intelligent control according to any one of claims 1 to 7, characterized in that: The following steps are involved: S1, respectively obtain the data of environmental variables, time dimension and device status and then aggregate them to obtain multi-dimensional data; S2, connected with the multi-dimensional prediction module, and automatically formulates charging and discharging strategies based on demand response mechanism, energy storage system and virtual power plant; S3, connected with the multi-dimensional prediction module and the automatic adjustment module, using edge computing and digital twin technology to establish a global simulation model of power grid operation; S4, connected with the monitoring feedback module, introduces autonomous learning capability for continuous optimization of power grid operation strategy.

9. The method for optimizing power grid stability based on intelligent control according to claim 8, characterized in that: In step S1, the method of obtaining the data of the environment variable, the time dimension and the device status respectively and then combining them to obtain multi-dimensional data is as follows: Based on the hybrid deep learning model, the time dependency and spatial correlation of the power generation fluctuation of the power grid are captured according to the multi-dimensional data, and the power fluctuation trend of the power grid is predicted, thereby obtaining prediction data; In step S2, the method of automatically formulating a charging and discharging strategy based on the demand response mechanism, energy storage system and virtual power plant in connection with the multi-dimensional prediction module is as follows: Based on the demand response mechanism, the user electricity demand and the grid operation parameters are adjusted in real time according to the predicted data, and then the balance between the load side and the power generation side is dynamically adjusted. By combining the energy storage system and the virtual power plant, power storage is performed during the peak period of grid power generation, and power is released during the trough period of grid power generation. The predicted data provides data support for the energy storage system and the virtual power plant to formulate charging and discharging strategies.

10. A method for optimizing power grid stability based on intelligent control according to claim 9, characterized in that: In step S3, the method of establishing a global simulation model of power grid operation by connecting with the multidimensional prediction module and the automatic adjustment module and utilizing edge computing and digital twin technology is as follows: The global simulation model of power grid operation is used to capture the correlation between power grid generation and load fluctuation in real time to obtain operation result data. When the operation result data deviates from the predicted data, the operation strategy of the automatic adjustment module is adjusted based on the intelligent feedback mechanism; In step S4, the monitoring feedback module is connected to introduce autonomous learning capabilities, and the method for continuously optimizing the power grid operation strategy is: A closed-loop learning model is designed based on a reinforcement learning algorithm. According to the operation result data, the optimal coordination data between power generation, power consumption and load fluctuations is dynamically explored through a reward and punishment mechanism. After the optimal coordination data is obtained, the power grid operation strategy is optimized.

Images