A flexible regulation control method for power grid based on artificial intelligence

Abstract

The present invention relates to the field of artificial intelligence technology and power grid regulation and control, and in particular to a power grid flexible regulation and control method based on artificial intelligence; the method comprises the following steps: dividing the power grid into different sub-areas; the sub-areas monitor the power grid parameters in the area and establish an exclusive power grid flexible regulation model; the sub-areas calculate the parameter data of the power grid flexible regulation model according to the latest power grid parameters, obtain the parameter data belonging to the sub-areas, and transmit the parameter data to a central server; the central server aggregates the parameter data belonging to each sub-area to form a global regulation model, and transmits the global regulation model to each sub-area; the sub-areas apply the global regulation model to the power grid flexible regulation, and autonomously and dynamically adjust the power grid flexible regulation parameters; finally, the power grid flexible regulation and control method is executed regularly; the present invention can more comprehensively grasp the operating status of the entire power grid, timely discover potential risks and take preventive measures.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology and power grid regulation and control, and in particular to a power grid flexible regulation and control method based on artificial intelligence. Background Art

[0002] With the rapid development of power systems and the large-scale access of new energy sources, the power grid is facing increasingly complex operating environments and regulation requirements. Traditional power grid regulation methods mainly rely on pre-set rules and parameters, which are difficult to adapt to the dynamic changes and diversified needs of the power grid.

[0003] In recent years, artificial intelligence technology has made significant progress in various fields, providing new ideas and methods for power grid regulation. However, existing artificial intelligence-based power grid regulation has some limitations, such as insufficient adaptability to complex scenarios, insufficient real-time performance, and poor interpretability. In particular, when faced with large-scale distributed energy access and drastic fluctuations in power load, existing methods are difficult to achieve fast, accurate, and flexible regulation and control.

[0004] Therefore, there is an urgent need to develop an artificial intelligence-based flexible power grid regulation control method to overcome the shortcomings of the existing technology. Summary of the invention

[0005] In order to solve the problems that the existing technology is difficult to deal with in large-scale distributed energy access and severe power load, the present invention provides a power grid flexible regulation control method based on artificial intelligence, comprising the following steps:

[0006] S100: Divide the power grid into different sub-areas, and the sub-areas independently perform flexible regulation control of the power grid within the area;

[0007] S200: The sub-area monitors the grid parameters in the area and establishes a dedicated grid flexible regulation model;

[0008] S300: The sub-region calculates parameter data of the grid flexible regulation model according to the latest grid parameters, obtains parameter data belonging to the sub-region, and transmits the parameter data belonging to the sub-region to the central server;

[0009] S400: The central server aggregates the parameter data belonging to each of the sub-regions to form a global adjustment model, and transmits the global adjustment model to each of the sub-regions;

[0010] S500: The sub-area applies the obtained global regulation model to the grid flexible regulation, and autonomously and dynamically adjusts the grid flexible regulation parameters.

[0011] Furthermore, the dividing of the power grid into different sub-areas specifically includes:

[0012] According to the penetration rate of renewable energy in the grid, load characteristics and grid topology, multi-dimensional cluster analysis is used to determine the optimal clustering level;

[0013] dividing the power grid into different sub-areas according to the optimal clustering level;

[0014] Use historical operation data to verify and optimize the division results and form a sub-area division plan;

[0015] When new renewable energy facilities are connected, the grid topology changes, or load characteristics change, the sub-area division scheme is adjusted in real time.

[0016] Furthermore, the sub-region monitors the grid parameters in the region and establishes a dedicated grid flexible regulation model, specifically including:

[0017] The divided sub-areas monitor the grid parameters in the area in real time, wherein the grid parameters include renewable energy generation data, grid topology data, load data and electrical parameter data;

[0018] Performing data cleaning on the power grid parameters obtained through monitoring;

[0019] Extract key features from the power grid parameters after data cleaning, wherein the key features include load peak and valley values, renewable energy generation values, node voltage and current, and power grid topology characteristics;

[0020] Based on the grid parameters, a grid flexible regulation model is constructed using a machine learning algorithm, and the grid flexible regulation model is trained to predict load changes, grid state changes, and regulation requirements in the future.

[0021] Furthermore, the grid flexible regulation model is constructed based on the LSTM algorithm, and the specific calculation formula is:

[0022] ΔL t+1 ,ΔS t+1 ,ΔD t+1 =LSTM(L t ,R t ,V t ,I t ,T t );

[0023] Among them, L t is the load data at time t, R t is the renewable energy generation data at time t, V t is the node voltage data at time t, I t is the node current data at time t, T t is the grid topology data at time t, ΔLt+1 is the load change from current time t to future time t+1, ΔS t+1 is the change in the power grid state from the current time t to the future time t+1, ΔD t+1 is the grid regulation demand from current time t to future time t+1.

[0024] Furthermore, before transmitting the parameter data to the central server, the sub-region adds noise to the parameter data to form encrypted parameter data, thereby ensuring that the encrypted parameter data sent will not expose the original parameter data of the sub-region.

[0025] Furthermore, the specific steps of forming the global adjustment model are:

[0026] S401: Acquire encrypted parameter data transmitted by each sub-region, and pre-process the encrypted parameter data, wherein the pre-processing includes filtering noise added to the encrypted parameter data to obtain parameter data of each sub-region before the noise is added;

[0027] S402: performing spatiotemporal consistency analysis on the acquired parameter data corresponding to each of the sub-regions according to the geographic information, power grid topology and time synchronization information corresponding to each of the sub-regions, to ensure the accuracy and timeliness of the parameter data;

[0028] S403: classify the analyzed parameter data, aggregate them according to the classification type, and calculate the weighted average value;

[0029] S404: clustering the analyzed parameter data of all sub-regions with similar power grid characteristics using a multi-dimensional clustering algorithm based on K-means;

[0030] S405: Based on the aggregated parameter data and clustering results, using a deep learning algorithm, constructing a global adjustment model, and training and verifying the global adjustment model;

[0031] S406: Distribute the trained global adjustment model to each of the sub-regions.

[0032] Furthermore, the sub-region applies the obtained global regulation model to the grid flexible regulation, and the specific steps of autonomously and dynamically adjusting the grid flexible regulation parameters are:

[0033] S501: After receiving the global regulation model from the central server, the sub-region parses the received global regulation model to extract the required power grid flexible regulation control strategy and power grid flexible regulation parameters;

[0034] S502: The sub-region performs local adaptive adjustment on the global regulation model according to the regional power grid characteristics and the power grid parameters obtained by real-time monitoring, so that the global regulation model conforms to the actual situation of the power grid in the sub-region;

[0035] S503: The sub-region integrates the real-time monitored power grid parameters with the prediction data of the global regulation model to generate a comprehensive regulation plan;

[0036] S504: Based on the comprehensive regulation scheme, the sub-region performs flexible regulation of the power grid within the management scope, including load regulation, power generation scheduling and reactive power compensation.

[0037] Furthermore, the S501 also includes evaluating the flexible regulation control strategy of the power grid, specifically including:

[0038] S511: After the sub-region applies the global regulation model to the grid flexible regulation process, real-time monitoring and collection of key grid parameters in the region;

[0039] S512: Compare and analyze the key grid parameters collected in real time with the predicted data of the global regulation model, calculate the deviation value, response time, operation stability and regulation accuracy; and use multi-dimensional evaluation indicators to evaluate the grid flexible regulation control strategy;

[0040] S513: performing correlation analysis on the analysis and evaluation results in S512 to identify the main factors affecting the flexible regulation effect of the power grid;

[0041] S514: If during the evaluation process, the execution effect of the power grid flexible regulation control strategy does not meet expectations, feedback information is generated in real time, the power grid flexible regulation parameters are adjusted, and the power grid flexible regulation control strategy is optimized;

[0042] S515: After the grid flexible regulation control strategy is optimized and executed, continue to monitor the grid operation status, conduct a secondary evaluation of the grid operation status and the optimized grid flexible regulation effect. When the evaluation result is not satisfactory, continue to optimize the grid flexible regulation control strategy and re-evaluate until the grid flexible regulation control strategy achieves the specified execution effect or the number of optimization iterations reaches the specified threshold.

[0043] Furthermore, the multi-dimensional evaluation indicators include load regulation effect, renewable energy consumption rate, reactive power compensation effect and power quality improvement.

[0044] Furthermore, it also includes setting a periodic schedule, and regularly executing the power grid flexible regulation control method according to the time set in the periodic schedule;

[0045] When it is detected that the state of the power grid changes significantly or the performance of the global regulation model in any sub-area drops to a preset threshold, the power grid flexible regulation control method is immediately executed.

[0046] Beneficial effects:

[0047] 1. The present invention divides the power grid into multiple sub-areas, each of which can be adjusted autonomously according to its own situation and global information, thereby improving the power grid's ability to resist local faults and disturbances; at the same time, the central server can more comprehensively grasp the operating status of the entire power grid by collecting and analyzing data from each sub-area, timely discover potential risks and take preventive measures;

[0048] 2. The present invention uses the penetration rate of renewable energy as an important basis for sub-regional division, and uses renewable energy power generation data as a key monitoring parameter. Through real-time analysis and prediction of these parameter data, the present invention can better manage and schedule the access of renewable energy and improve its absorption rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 is a flow chart of the method of the present invention;

[0050] Figure 2 A flow chart of forming a global regulation model of the present invention;

[0051] Figure 3 A flow chart of applying the global regulation model to the flexible regulation of the power grid for the sub-regions of the present invention. DETAILED DESCRIPTION

[0052] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0053] like Figure 1 As shown, an embodiment of the present invention provides an artificial intelligence-based power grid flexible regulation control method, including:

[0054] S100: Divide the power grid into different sub-areas, and the sub-areas independently perform flexible regulation control of the power grid within the area.

[0055] When dividing the power grid into different sub-areas, various types of data are first collected, including but not limited to renewable energy generation data, load data, power grid topology data, etc., and then principal component analysis is used to select important features, where important features include but are not limited to load peak and valley values, node voltage and current, and power grid topology characteristics.

[0056] After assigning weights to the features of each dimension through principal component analysis, the Euclidean distance is used to calculate the comprehensive similarity between different regions. Based on the calculated comprehensive similarity matrix, multi-dimensional clustering analysis is used to determine the optimal clustering level, divide the power grid into different sub-regions, and ensure that the renewable energy penetration rate, load characteristics and power grid topology structure in the same sub-region are as similar as possible.

[0057] The division results are verified and optimized using historical operating data to form a sub-area division plan. When renewable energy facilities are connected, the grid topology changes, or load characteristics change, the sub-area division plan is adjusted in real time or periodically to adapt to dynamic changes in the grid operating environment.

[0058] S200: The sub-area monitors the grid parameters within the area and establishes an exclusive grid flexible regulation model.

[0059] In this embodiment, the sub-areas after division monitor the grid parameters in the corresponding area in real time, and the monitored grid parameters include but are not limited to renewable energy power generation data, grid topology data, load data and electrical parameter data; then the collected grid parameter data is cleaned and preprocessed to remove noise and abnormal values, and extract key feature values ​​including load peak and valley values, renewable energy power generation values, node voltage and current, and grid topology characteristics;

[0060] Based on the grid parameters, the LSTM algorithm is used to build a grid flexible regulation model, and the grid flexible regulation model is trained to predict load changes, grid state changes and regulation needs in the future.

[0061] Among them, the grid flexible regulation model is built based on the LSTM algorithm, and the specific calculation formula is:

[0062] ΔL t+1 ,ΔS t+1 ,ΔD t+1 =LSTM(L t ,R t ,V t ,I t ,T t );

[0063] Among them, L t is the load data at time t, R t is the renewable energy generation data at time t, V t is the node voltage data at time t, I t is the node current data at time t, T t is the grid topology data at time t, ΔL t+1 is the load change from current time t to future time t+1, ΔSt+1 is the change in the power grid state from the current time t to the future time t+1, ΔD t+1 is the grid regulation demand from current time t to future time t+1.

[0064] S300: The sub-area calculates parameter data of the grid flexible regulation model according to the latest grid parameters, obtains parameter data belonging to the sub-area, and transmits the parameter data to the central server.

[0065] In an embodiment, when a sub-region transmits parameter data to a central server, noise is added to the parameter data to form encrypted parameter data, ensuring that the encrypted parameter data sent does not expose the original parameter data of the sub-region; at the same time, each sub-region uploads the parameter data of the power grid flexible regulation model at a different time to reduce the communication delay of the central server.

[0066] By adding noise to the parameter data transmission process of the sub-area, the privacy and security of the original parameter data of the sub-area can be effectively protected, so that the power grid can achieve centralized management and optimized control of the sub-area without exposing sensitive data.

[0067] S400: The central server aggregates the parameter data belonging to all sub-regions to form a global adjustment model, and transmits the global adjustment model to all sub-regions.

[0068] In this embodiment, the global regulation model forms an overall grid flexible regulation control strategy by summarizing the grid parameters obtained by real-time monitoring of each sub-region and the parameters of the grid flexible regulation model, and coordinates the regulation actions of each sub-region to achieve the global optimal control target; wherein the steps of forming the global regulation model are as follows: Figure 2 As shown, including:

[0069] S401: Acquire encrypted parameter data transmitted by each sub-region, and pre-process the encrypted parameter data, wherein the pre-processing includes filtering the noise added to the encrypted parameter data to obtain the parameter data of each sub-region before the noise is added; wherein the parameter data includes but is not limited to the weight and bias item of the flexible regulation model of the sub-region power grid, the renewable energy generation forecast data of the sub-region, the load forecast data of the sub-region, and the power grid topology information of the sub-region;

[0070] S402: Performing spatiotemporal consistency analysis on the acquired parameter data corresponding to each sub-region based on the geographic information, power grid topology, time synchronization information and other data corresponding to each sub-region to ensure the accuracy and timeliness of the parameter data;

[0071] S403: classifying the analyzed parameter data from different sub-regions, aggregating the parameter data such as load data and renewable energy generation data according to the type, and calculating the weighted average thereof;

[0072] S404: clustering the parameter data of sub-regions with similar power grid characteristics after analysis using a K-means multi-dimensional clustering algorithm, analyzing the characteristics of the clustered data, and optimizing the power grid flexible regulation control strategy;

[0073] S405: Based on the aggregated parameter data and clustering results, a global adjustment model is constructed using a deep learning algorithm, and the global adjustment model is trained and verified to improve prediction accuracy and adjustment effect;

[0074] S406: Distribute the trained global regulation model to all sub-areas to ensure that each sub-area uses a unified and optimized global regulation model during the power grid regulation process.

[0075] In this embodiment, the global regulation model comprehensively considers the status information of each sub-region and the mutual influence from the perspective of the entire power grid, so as to formulate a more reasonable power grid flexible regulation control strategy than the independent decision-making of each sub-region; at the same time, the global regulation model is sent to each sub-region to guide the sub-region to collaboratively complete the global goals including but not limited to load balancing and renewable energy consumption on the premise of ensuring its own stable operation.

[0076] S500: The sub-area applies the global regulation model to the grid flexible regulation, and autonomously and dynamically adjusts the grid flexible regulation parameters.

[0077] In this embodiment, after the sub-region receives the global regulation model sent by the central server, the operation steps of autonomously and dynamically adjusting the grid flexible regulation parameters are as follows: Figure 3 As shown, including:

[0078] S501: After receiving the global regulation model from the central server, the sub-region parses the received global regulation model to extract the required power grid flexible regulation control strategy and power grid flexible regulation parameters;

[0079] S502: The sub-region performs local adaptive adjustment on the global regulation model according to the characteristics of the power grid in the region and the power grid parameters obtained by real-time monitoring, and calibrates the key model parameters in the global regulation model so that the global regulation model conforms to the actual situation of the power grid in the sub-region;

[0080] S503: The sub-region continuously collects the grid parameters in the region, wherein the grid parameters reflect the grid operation status in the sub-region in real time; then the grid parameters are used as real-time and detailed input data, and the grid parameters are fused with the prediction data of the global regulation model using a data fusion algorithm to generate a comprehensive regulation plan;

[0081] S504: Based on the comprehensive regulation plan, the sub-region performs flexible regulation of the power grid within the management scope, including but not limited to load regulation, power generation scheduling and reactive power compensation.

[0082] In this embodiment, step S501 also includes evaluating the power grid flexible regulation control strategy, specifically including:

[0083] S511: After the sub-region applies the global regulation model to the flexible regulation process of the power grid, real-time monitoring and collection of key power grid parameters in the region during the regulation process;

[0084] S512: Compare and analyze the key grid parameters collected in real time with the predicted data of the global regulation model, and calculate the evaluation indicators, including deviation value, response time, operation stability and regulation accuracy; and evaluate the grid flexible regulation control strategy using multi-dimensional evaluation indicators; the multi-dimensional rating indicators include but are not limited to load regulation effect, renewable energy consumption rate, reactive power compensation effect and power quality improvement;

[0085] S513: performing correlation analysis on the comprehensive evaluation results in S512 to identify the main factors that have a significant impact on the flexible regulation effect of the power grid;

[0086] S514: If during the evaluation process, the execution effect of the power grid flexible regulation control strategy does not meet expectations, feedback information is generated in real time, the power grid flexible regulation parameters are adjusted, and the power grid flexible regulation control strategy is optimized;

[0087] S515: After the grid flexible regulation control strategy is optimized and executed, continue to monitor the grid operation status, conduct a second evaluation of the grid operation status and the optimized grid flexible regulation effect. When the execution effect obtained by the evaluation does not meet the expected effect, re-optimize the grid flexible regulation control strategy based on the grid parameters and re-evaluate. In order to prevent the execution effect from not meeting expectations, the maximum number of optimization iterations is set to solve the problem. When the number of optimization iterations of the grid flexible regulation control strategy reaches the specified threshold, the grid flexible regulation control method is re-executed as a whole, and the grid flexible regulation parameters are recalculated.

[0088] In this embodiment, the sub-region performs specific regulation actions based on the flexible regulation control strategy of the power grid proposed by the global regulation model and the real-time status within its own region, and feeds back the actual data of the regulation process to the central server for dynamic correction and optimization of the global regulation model.

[0089] In this embodiment, a periodic schedule is also provided for periodically executing the power grid flexible regulation control method described in the above embodiment according to the periodic schedule. At the same time, when a significant change in the power grid state is detected, or when the performance of the global regulation model is degraded in any one or more sub-areas, the power grid flexible regulation control method can be executed immediately without considering the next execution time of the power grid flexible regulation control method set in the periodic schedule, and after this execution, the periodic schedule continues from the current execution time.

[0090] The above-mentioned embodiments only express the preferred implementation modes of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present invention. It should be pointed out that, for a person skilled in the art, several modifications, improvements and substitutions can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the present invention patent shall be subject to the attached claims.

Claims

1. A power grid flexible regulation control method based on artificial intelligence, characterized in that: The following steps are involved: S100: Divide the power grid into different sub-areas, and the sub-areas independently perform flexible regulation control of the power grid within the area; S200: The sub-area monitors the grid parameters in the area and establishes a dedicated grid flexible regulation model; S300: The sub-region calculates parameter data of the grid flexible regulation model according to the latest grid parameters, obtains parameter data belonging to the sub-region, and transmits the parameter data belonging to the sub-region to the central server; S400: The central server aggregates the parameter data belonging to each of the sub-regions to form a global adjustment model, and transmits the global adjustment model to each of the sub-regions; S500: The sub-area applies the obtained global regulation model to the grid flexible regulation, and autonomously and dynamically adjusts the grid flexible regulation parameters; The power grid is divided into different sub-areas, specifically including: According to the penetration rate of renewable energy in the grid, load characteristics and grid topology, multi-dimensional cluster analysis is used to determine the optimal clustering level; dividing the power grid into different sub-areas according to the optimal clustering level; Use historical operation data to verify and optimize the division results and form a sub-area division plan; When new renewable energy facilities are connected, the grid topology changes, or load characteristics change, the sub-area division scheme is adjusted in real time; The sub-region applies the obtained global regulation model to the grid flexible regulation, and the specific steps of autonomously and dynamically adjusting the grid flexible regulation parameters are as follows: S501: After receiving the global regulation model from the central server, the sub-region parses the received global regulation model to extract the required power grid flexible regulation control strategy and power grid flexible regulation parameters; S502: The sub-region performs local adaptive adjustment on the global regulation model according to the regional power grid characteristics and the power grid parameters obtained by real-time monitoring, so that the global regulation model conforms to the actual situation of the power grid in the sub-region; S503: The sub-region integrates the real-time monitored power grid parameters with the prediction data of the global regulation model to generate a comprehensive regulation plan; S504: Based on the comprehensive regulation scheme, the sub-region performs flexible regulation of the power grid within the management scope, including load regulation, power generation scheduling and reactive power compensation.

2. The power grid flexible regulation control method based on artificial intelligence according to claim 1 is characterized in that: The sub-area monitors the grid parameters within the area and establishes a dedicated grid flexible regulation model, specifically including: The divided sub-areas monitor the grid parameters in the area in real time, wherein the grid parameters include renewable energy generation data, grid topology data, load data and electrical parameter data; Performing data cleaning on the power grid parameters obtained through monitoring; Extract key features from the power grid parameters after data cleaning, wherein the key features include load peak and valley values, renewable energy generation values, node voltage and current, and power grid topology characteristics; Based on the grid parameters, a grid flexible regulation model is constructed using a machine learning algorithm, and the grid flexible regulation model is trained to predict load changes, grid state changes, and regulation requirements in the future.

3. The method for flexible regulation and control of power grid based on artificial intelligence according to claim 2 is characterized in that: The grid flexible regulation model is constructed based on the LSTM algorithm, and the specific calculation formula is: ； in, for Load data at the moment, for Renewable energy generation data at all times, for The node voltage data at the moment, for The node current data at the moment, for The grid topology data at the moment, From the current time To the future time The load variation, From the current time To the future time The grid status changes, From the current time To the future time grid regulation needs.

4. The method for flexible regulation and control of power grid based on artificial intelligence according to claim 3 is characterized in that: Before transmitting the parameter data to the central server, the sub-region adds noise to the parameter data to form encrypted parameter data, ensuring that the encrypted parameter data sent will not expose the original parameter data of the sub-region.

5. The method for flexible regulation and control of power grid based on artificial intelligence according to claim 4 is characterized in that: The specific steps of forming the global adjustment model are: S401: Acquire encrypted parameter data transmitted by each sub-region, and pre-process the encrypted parameter data, wherein the pre-processing includes filtering noise added to the encrypted parameter data to obtain parameter data of each sub-region before the noise is added; S402: performing spatiotemporal consistency analysis on the acquired parameter data corresponding to each of the sub-regions according to the geographic information, power grid topology and time synchronization information corresponding to each of the sub-regions, to ensure the accuracy and timeliness of the parameter data; S403: classify the analyzed parameter data, aggregate them according to the classification type, and calculate the weighted average value; S404: clustering the analyzed parameter data of all sub-regions with similar power grid characteristics using a multi-dimensional clustering algorithm based on K-means; S405: Based on the aggregated parameter data and clustering results, using a deep learning algorithm, constructing a global adjustment model, and training and verifying the global adjustment model; S406: Distribute the trained global adjustment model to each of the sub-regions.

6. The method for flexible regulation and control of power grid based on artificial intelligence according to claim 5 is characterized in that: The S501 also includes evaluating the flexible regulation control strategy of the power grid, specifically including: S511: After the sub-region applies the global regulation model to the grid flexible regulation process, real-time monitoring and collection of key grid parameters in the region; S512: Compare and analyze the key grid parameters collected in real time with the predicted data of the global regulation model, calculate the deviation value, response time, operation stability and regulation accuracy; and use multi-dimensional evaluation indicators to evaluate the grid flexible regulation control strategy; S513: performing correlation analysis on the analysis and evaluation results in S512 to identify the main factors affecting the flexible regulation effect of the power grid; S514: If during the evaluation process, the execution effect of the power grid flexible regulation control strategy does not meet expectations, feedback information is generated in real time, the power grid flexible regulation parameters are adjusted, and the power grid flexible regulation control strategy is optimized; S515: After the grid flexible regulation control strategy is optimized and executed, continue to monitor the grid operation status, conduct a secondary evaluation of the grid operation status and the optimized grid flexible regulation effect. When the evaluation result is not satisfactory, continue to optimize the grid flexible regulation control strategy and re-evaluate until the grid flexible regulation control strategy achieves the specified execution effect or the number of optimization iterations reaches the specified threshold.

7. The artificial intelligence-based flexible regulation control method for power grid according to claim 6 is characterized in that: The multi-dimensional evaluation indicators include load regulation effect, renewable energy consumption rate, reactive power compensation effect and power quality improvement.

8. The method for flexible regulation and control of power grid based on artificial intelligence according to claim 7, characterized in that: It also includes setting a periodic schedule, and regularly executing the power grid flexible regulation control method according to the time set in the periodic schedule; When it is detected that the state of the power grid changes significantly or the performance of the global regulation model in any sub-area drops to a preset threshold, the power grid flexible regulation control method is immediately executed.

Images