Power grid management and control method based on artificial intelligence and related equipment

Abstract

The embodiment of the invention relates to the technical field of intelligent power grids, and discloses a power grid management and control method based on artificial intelligence and related equipment, and the method comprises the steps: collecting the real-time power grid data of power grid equipment at the current time period; inputting the real-time power grid data into a power grid state prediction model, and predicting to obtain a current power grid operation state; inputting the current power grid operation state and the real-time power grid data into a power grid strategy dynamic optimization module to generate a new power distribution strategy; setting power grid control parameters according to the new power distribution strategy and the real-time power grid data, and executing the new power distribution strategy; and when a preset optimization period is reached, performing power grid resource dynamic optimization by adopting a power grid resource dynamic optimization algorithm according to the current power grid operation state and the real-time resource data to obtain an optimized power grid resource optimization scheme. According to the embodiment of the invention, the real-time monitoring of the power grid state and the optimal distribution of power resources are realized.

Description

Technical Field

[0001] Embodiments of the present invention relate to the technical field of smart grids, and particularly to a power grid control method and related devices based on artificial intelligence. Background Art

[0002] At present, with the growth of global energy demand and the rapid development of new energy technologies, smart grids have become the key to the modernization of power systems. Smart grids utilize means such as information and communication technologies, advanced measurement technologies, automation technologies, and artificial intelligence to achieve the efficient, reliable, and environmentally friendly operation of power grids. However, despite the significant progress made in smart grid technologies, the existing systems still face the following technical problems in practical applications: 1. Insufficient data processing capabilities: Existing power grid management systems often rely on traditional data processing methods when dealing with large-scale and high-frequency energy data, and these methods are difficult to meet the requirements of real-time optimization in terms of processing speed and accuracy.

[0003] 2. Limited prediction accuracy: Power grid load prediction is the key to optimizing power distribution, but existing prediction models are usually based on historical data and simple statistical methods, and it is difficult to adapt to the complex and changing power grid environment, resulting in inaccurate prediction results.

[0004] 3. Lack of adaptability: Existing power grid management software often lacks sufficient adaptability and cannot automatically adjust management strategies according to changes in the operating state of the power grid, which limits the response ability of the system in the face of emergencies.

[0005] 4. Poor user interaction experience: The user interface design is not intuitive enough and the operation is complex, making it difficult for power grid management personnel to quickly understand and control the system state, which affects the management efficiency.

[0006] 5. Difficulty in integrating new energy: With the access of new energy such as solar energy and wind energy, existing power grid management systems face challenges in dealing with these intermittent and unpredictable energy sources and lack effective integration and management strategies. Summary of the Invention

[0007] In view of the above problems, embodiments of the present invention provide a power grid control method and related devices based on artificial intelligence, which are used to solve the problems of insufficient data processing capabilities and low prediction accuracy existing in the prior art.

[0008] According to one aspect of the embodiments of the present invention, a power grid control method based on artificial intelligence is provided, and the method includes: Collecting real-time power grid data of the current period of power grid equipment; the real-time power grid data includes power production, load demand, voltage fluctuation data, and weather condition data; Input the real-time grid data into the grid state prediction model to predict the current grid operating state; the current grid operating state includes peak load periods, anomaly detection information, and grid fault location information; wherein, the grid state prediction model is pre-trained based on grid state sample data for a deep learning network, and the grid state sample data includes historical grid operation data and corresponding sample labels; Input the current grid operating state and the real-time grid data into the grid strategy dynamic optimization module to generate a new power distribution strategy; Set grid control parameters according to the new power distribution strategy and the real-time grid data, and execute the new power distribution strategy; When the preset optimization period is reached, perform dynamic optimization of grid resources according to the current grid operating state and real-time resource data using a grid resource dynamic optimization algorithm to obtain an optimized grid resource optimization plan; wherein, the grid resource dynamic optimization algorithm includes a hybrid algorithm combining model predictive control and reinforcement learning, and the real-time resource data includes load demand information, new energy power generation, remaining capacity of the energy storage system, grid topology, and external factor data.

[0009] In an alternative approach, before inputting the real-time grid data into the grid state prediction model to predict the current grid operating state, the method further includes: Collect historical grid operation data; the historical grid operation data includes historical power production, historical load demand, historical voltage fluctuation data, and historical weather condition data; Add corresponding sample labels to the historical grid operation data to obtain grid state sample data; the sample labels include the corresponding historical grid operating states, and the historical grid operating state labels include peak load period labels, anomaly detection information labels, and grid fault location information labels; Input the grid state sample data into the deep learning network for iterative training, and adjust the parameters of the deep learning network to obtain a trained grid state prediction model.

[0010] In an alternative approach, the deep learning network includes a time series prediction module, a spatial correlation prediction module, and an anomaly detection module; wherein, the time series prediction module includes one of a self-attention mechanism and a CNN-LSTM hybrid model; the spatial correlation prediction module includes a GNN algorithm; the anomaly detection module includes a variational autoencoder; The step of inputting the grid state sample data into the deep learning network for iterative training and adjusting the parameters of the deep learning network to obtain a trained grid state prediction model includes: Input the power grid state sample data into the self-attention mechanism or the CNN-LSTM hybrid model for training to obtain a time series prediction module for predicting the load peak period prediction results; Input the power grid state sample data into the GNN algorithm for training to obtain a spatial association prediction module for locating power grid faults; Input the power grid state sample data into the variational autoencoder for training to obtain an anomaly detection module for detecting abnormal events.

[0011] In an alternative approach, the real-time power grid data of the current period of the power grid equipment is collected, including: Respectively collect the power generation information of each power generation equipment in the power grid to obtain power generation data; the power generation equipment includes thermal power plants, hydropower stations, wind power stations and solar photovoltaic systems; Obtain real-time load data from the load monitoring equipment of the substation and the distribution network to obtain the load demand; Monitor and record voltage fluctuations through voltage sensors installed at key nodes of the power grid to collect voltage fluctuation data; Obtain weather data affecting energy production and consumption through communication with the meteorological service center.

[0012] In an alternative approach, input the current power grid operating state and the real-time power grid data into the power grid strategy dynamic optimization module to generate a new power distribution strategy, including: Evaluate the effectiveness of the current power distribution strategy based on the current power grid operating state and the real-time power grid data, and identify improvement points; Generate a new power distribution strategy according to the improvement points and the current power grid operating state according to a preset adjustment strategy; Conduct a simulation test on the new power distribution strategy; When the test passes, apply the new power distribution strategy to power grid management.

[0013] In an alternative approach, when the preset optimization period is reached, perform dynamic optimization of the power grid resources according to the current power grid operating state and real-time resource data using the power grid resource dynamic optimization algorithm to obtain an optimized power grid resource optimization plan, including: Determine a multi-objective function based on traditional power generation fuel costs, energy storage loss costs, new energy curtailment penalties and voltage deviations; among them, set power balance constraints, energy storage dynamic constraints and power grid equipment limit constraints; When the preset optimization period is reached, perform target optimization on the multi-objective function according to the current power grid operating state and real-time resource data using the reinforcement learning adaptive strategy to obtain an optimized power grid resource optimization plan.

[0014] In an alternative manner, the method further includes: displaying the current power grid operation state and the optimized power grid resource optimization plan in a user interface.

[0015] According to another aspect of the embodiments of the present invention, there is provided an artificial intelligence-based power grid control system, including: An acquisition module, configured to acquire real-time power grid data of power grid equipment in the current period; the real-time power grid data includes power generation, load demand, voltage fluctuation data, and weather condition data; A prediction module, configured to input the real-time power grid data into a power grid state prediction model to predict the current power grid operation state; the current power grid operation state includes whether it is a peak load period and whether the voltage is abnormal; wherein, the power grid state prediction model is pre-trained based on power grid state sample data for a deep learning network, and the power grid state sample data includes power grid historical operation data and corresponding sample labels; A power strategy module, configured to input the current power grid operation state and the real-time power grid data into a power grid strategy dynamic optimization module to generate a new power distribution strategy; An execution module, configured to set power grid control parameters according to the new power distribution strategy and the real-time power grid data, and execute the new power distribution strategy; A resource optimization module, configured to, when a preset optimization period arrives, perform dynamic optimization of power grid resources by using a power grid resource dynamic optimization algorithm according to the current power grid operation state and real-time resource data to obtain an optimized power grid resource optimization plan; wherein, the power grid resource dynamic optimization algorithm includes a hybrid algorithm combining model predictive control and reinforcement learning, and the real-time resource data includes load demand information, new energy power generation, remaining capacity of an energy storage system, power grid topology, and external factor data.

[0016] According to another aspect of the embodiments of the present invention, there is provided a computer device, including: a processor, a memory, a communication interface, and a communication bus, and the processor, the memory, and the communication interface complete communication with each other through the communication bus; The memory is used to store at least one executable instruction, and the executable instruction causes the processor to execute the operations of the artificial intelligence-based power grid control method.

[0017] According to yet another aspect of the embodiments of the present invention, there is provided a computer-readable storage medium, in which at least one executable instruction is stored, and when the executable instruction runs on a computer device, it causes the computer device to execute the operations of the artificial intelligence-based power grid control method.

[0018] In an embodiment of the present invention, real-time grid data of grid equipment in the current period is collected; the real-time grid data is input into a grid state prediction model to predict the current grid operation state; the current grid operation state and the real-time grid data are input into a grid strategy dynamic optimization module to generate a new power distribution strategy; according to the new power distribution strategy and the real-time grid data, grid control parameters are set and the new power distribution strategy is executed; when a preset optimization period is reached, according to the current grid operation state and real-time resource data, a grid resource dynamic optimization algorithm is used for grid resource dynamic optimization to obtain an optimized grid resource optimization plan, which can improve the data processing ability, prediction accuracy and adaptive ability of the grid management system, while optimizing the user interface, enhancing the integration and management of new energy, realizing real-time monitoring of the grid operation state, load prediction, fault diagnosis and optimal allocation of power resources, thereby reducing the operation cost, improving the energy utilization rate, and supporting the development of sustainable energy.

[0019] The above description is only an overview of the technical solution of the embodiment of the present invention. In order to be able to understand the technical means of the embodiment of the present invention more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of the embodiment of the present invention more obvious and understandable, the following specifically describes the specific implementation manners of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The drawings are only used to illustrate the embodiments and are not considered as a limitation of the present invention. And throughout the drawings, the same reference numerals are used to represent the same components. In the drawings: Figure 1 A flowchart showing the method for grid control and management based on artificial intelligence provided by an embodiment of the present invention is shown; Figure 2 A flowchart showing the method for grid control and management based on artificial intelligence provided by an embodiment of the present invention is shown; Figure 3 A structural diagram showing the grid control and management system based on artificial intelligence provided by an embodiment of the present invention is shown; Figure 4 A structural diagram showing the computer device provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] The following will describe the exemplary embodiments of the present invention in more detail with reference to the drawings. Although the exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments described herein.

[0022] Figure 1The flowchart of the grid control method based on artificial intelligence provided by the embodiments of the present invention is shown. This method is executed by a computer device. The computer device can be an electronic device deployed with a grid control system based on artificial intelligence, and can be a computer, a distributed device, an edge processing device, etc. that communicate with each grid device. The embodiments of the present invention do not make specific limitations. As Figure 1 shown, the method includes the following steps: Step 110: Collect real-time grid data of grid devices in the current period.

[0023] Among them, the real-time grid data includes power generation output, load demand, voltage fluctuation data, and weather condition data. In the embodiments of the present invention, the real-time grid data is obtained through the following methods: Respectively collect the power generation information of each power generation device in the power grid to obtain power generation data. Among them, the power generation devices include thermal power plants, hydropower stations, wind power stations, and solar photovoltaic systems. In the embodiments of the present invention, the computer device is integrated with each power generation device, so that the power generation information of each power generation device can be collected. Obtain real-time load data from the load monitoring devices of the substation and the distribution network. These data reflect the power demand in different time periods, so as to obtain the load demand.

[0024] Monitor and record voltage fluctuation and stability information through voltage sensors installed at key nodes of the power grid, so as to collect voltage fluctuation data.

[0025] Through communication with the meteorological service center, obtain weather data that affects energy production and consumption. Specifically, it can be connected to the API interface of the meteorological service center to obtain weather data that affects energy production and consumption, such as temperature, wind force, precipitation, etc.

[0026] Step 120: Input the real-time grid data into the grid state prediction model to predict the current grid operation state.

[0027] Among them, the current grid operation state includes information on peak load periods, abnormal detection information such as whether the voltage is abnormal, and grid fault location information. Among them, the grid state prediction model is obtained by training a deep learning network in advance according to grid state sample data, and the grid state sample data includes grid historical operation data and corresponding sample labels.

[0028] In the embodiments of the present invention, before inputting the real-time grid data into the grid state prediction model, the grid state prediction model is also trained in advance. The specific training process includes: Step 001: Collect historical operation data of the power grid; the historical operation data of the power grid includes historical power generation, historical load demand, historical voltage fluctuation data, and historical weather condition data. Among them, after obtaining the historical operation data of the power grid, preprocess the historical operation data of the power grid, and perform data cleaning and standardization processing to eliminate noise and outliers.

[0029] Step 002: Add corresponding sample labels to the historical operation data of the power grid to obtain power grid state sample data. Among them, the sample labels include the corresponding historical operation states of the power grid, and the historical operation state labels of the power grid include load peak period labels, anomaly detection information labels, and power grid fault location information labels.

[0030] Step 003: Input the power grid state sample data into a deep learning network for iterative training, and adjust the parameters of the deep learning network to obtain a trained power grid state prediction model.

[0031] In the embodiments of the present invention, according to the characteristics of power grid data and the requirements of the prediction task, the architecture of the deep learning network is designed, and appropriate network types, number of layers, number of neurons, activation functions, and loss functions are selected. Specifically, the architecture of the deep learning network includes a time series prediction module, a spatial correlation prediction module, and an anomaly detection module. Among them, the time series prediction module includes one of a self-attention mechanism and a CNN-LSTM hybrid model; the spatial correlation prediction module includes a GNN algorithm; the anomaly detection module includes a variational autoencoder. The time series prediction task, such as load prediction, is performed through the self-attention mechanism or the CNN-LSTM hybrid model. Therefore, the process of inputting the power grid state sample data into the deep learning network for iterative training is as follows: Step 030: Input the power grid state sample data into the self-attention mechanism or the CNN-LSTM hybrid model for training to obtain a time series prediction module for predicting the load peak period prediction result. Among them, Transformer captures global dependencies through the attention mechanism and is suitable for long sequence prediction. In the CNN-LSTM hybrid model, CNN is used to extract spatial features, and LSTM is used to process time series features.

[0032] Step 031: Input the power grid state sample data into the GNN algorithm for training to obtain a spatial correlation prediction module for locating power grid faults. Among them, GNN (Graph Neural Networks) processes the power grid as a graph structure, where nodes represent each power grid device and edges represent connection relationships. The GNN algorithm is used to process tasks with strong spatial correlation, such as power grid fault location, etc.

[0033] Step 032: Input the power grid state sample data into a variational autoencoder for training to obtain an anomaly detection module for detecting abnormal events.

[0034] In the embodiment of the present invention, a loss function is selected according to the task type. For a regression task: MSE (mean squared error) is adopted. This loss function is sensitive to outliers and is suitable for stable data. For a classification task, cross-entropy loss is adopted. During training, the model is initialized: the network parameters such as weights and biases are initialized. The initial values of these parameters have an important impact on the training effect of the model. Training execution: A part of the data set is used as the training set to execute model training. The predicted output is calculated through forward propagation, and the network parameters are updated according to the gradient calculated by the loss function through the backpropagation algorithm. Hyperparameter tuning: Hyperparameters such as the learning rate, batch size, and number of training epochs are adjusted to improve the convergence speed and prediction performance of the model. Model validation: The performance of the model is evaluated using the validation set, cross-validation is performed to avoid overfitting, and the model structure or hyperparameters are adjusted according to the validation results. Model testing: The final performance of the model is evaluated on an independent test set to ensure that the model has good generalization ability. Model iteration: The model is iteratively optimized according to the test results until the preset performance level is reached.

[0035] After training the trained power grid state prediction model, the real-time power grid data is preprocessed and then input into the power grid state prediction model to obtain the current power grid operation state; the current power grid operation state includes peak load periods, anomaly detection information, and power grid fault location information.

[0036] Among them, the embodiment of the present invention also extracts knowledge from the trained power grid state prediction model, such as typical operation modes of the power grid, key influencing factors, etc., to provide decision-making support for the intelligent management of the power grid.

[0037] Step 130: Input the current power grid operation state and the real-time power grid data into the power grid strategy dynamic optimization module to generate a new power distribution strategy.

[0038] Among them, the power grid strategy dynamic optimization module constructs an objective function according to the key factors of power distribution and optimizes it to obtain a power distribution strategy.

[0039] Step 140: Set power grid control parameters according to the new power distribution strategy and the real-time power grid data, and execute the new power distribution strategy.

[0040] Among them, according to the current power grid operation status and the real-time power grid data, evaluate the effectiveness of the current power distribution strategy, and identify improvement points; according to the improvement points and the current power grid operation status, generate a new power distribution strategy according to a preset adjustment strategy; conduct a simulation test on the new power distribution strategy; when the test passes, apply the new power distribution strategy to grid management.

[0041] Step 150: When the preset optimization period is reached, according to the current power grid operation status and real-time resource data, adopt a power grid resource dynamic optimization algorithm to perform dynamic optimization of power grid resources, and obtain an optimized power grid resource optimization plan.

[0042] Among them, the power grid resource dynamic optimization algorithm includes a hybrid algorithm combining model predictive control and reinforcement learning, and the real-time resource data includes load demand information, new energy power generation, remaining capacity of the energy storage system, power grid topology structure, and external factor data.

[0043] When the preset optimization period is reached, according to the current power grid operation status and real-time resource data, adopt a reinforcement learning adaptive strategy to optimize the multi-objective function, and obtain an optimized power grid resource optimization plan.

[0044] Specifically, in the embodiments of the present invention, a multi-objective function is determined according to the traditional power generation fuel cost, energy storage loss cost, and voltage deviation; among them, power balance constraints, energy storage dynamic constraints, and power grid equipment limit constraints are set.

[0045] Among them, in one embodiment of the present invention, the objective function can be expressed as: Among them, is the traditional power generation fuel cost, \((t)\) is the energy storage loss cost, is the voltage deviation. , and are weight coefficients, which are dynamically adjusted during the process of reinforcement learning.

[0046] Among them, according to the relationship between the stored power, the obtained power, the updated power, the load power, and the loss power, a power balance constraint equation is set. Energy storage dynamic constraints are set, such as the thermal power ramp rate and the energy storage SOC range, and power grid equipment limit constraints. The power grid equipment limit constraint is to ensure that the voltage stability is within the ±5% threshold range.

[0047] In the embodiments of the present invention, the process of learning and strengthening is: ; among them, represents the current state, such as load demand and new energy output. a represents an action, such as an energy storage charge and discharge instruction. represents the immediate reward, such as the cost reduction value. is the learning rate, represents the discount factor.

[0048] Set to update the optimization window every 5 minutes, and recalculate the optimal solution based on the latest data. Among them, the state space (s) includes discretized variables and continuous variables. The discretized variables include load levels (high / medium / low), new energy penetration rate, and energy storage SOC intervals. The continuous variables include voltage deviation values and real-time electricity prices.

[0049] Set the action space (a) to include discrete actions and continuous actions. The discrete actions include starting and stopping hydropower, and forced curtailment of light / wind instructions; the continuous actions include adjusting the percentage of thermal power output and the energy storage charge and discharge power.

[0050] Through the Q-learning algorithm, dynamically adjust the optimization weights according to the historical operation feedback. Among them, the historical operation feedback includes the decrease in the energy storage charge and discharge efficiency, etc.

[0051] Through the above process, adopt a hybrid algorithm that combines model predictive control (MPC) and reinforcement learning (RL) to achieve the dynamic optimal allocation of grid resources.

[0052] In an embodiment of the present invention, it further includes: displaying the current grid operation state and the optimized grid resource optimization plan in the user interface. Through the intuitive user interface, grid managers can easily monitor the system state, view real-time data and prediction results, and execute control commands. The interface supports multi-platform access, including desktops, mobile devices, and tablets.

[0053] In the embodiment of the present invention, by collecting the real-time grid data of grid equipment in the current period; inputting the real-time grid data into the grid state prediction model to predict the current grid operation state; inputting the current grid operation state and the real-time grid data into the grid strategy dynamic optimization module to generate a new power distribution strategy; setting grid control parameters according to the new power distribution strategy and the real-time grid data, and executing the new power distribution strategy; when the preset optimization period is reached, according to the current grid operation state and real-time resource data, adopt the grid resource dynamic optimization algorithm to perform grid resource dynamic optimization to obtain the optimized grid resource optimization plan, which can improve the data processing ability, prediction accuracy and adaptive ability of the grid management system, while optimizing the user interface, enhancing the integration and management of new energy, and realizing the real-time monitoring of the grid operation state, load prediction, fault diagnosis, and the optimal allocation of power resources, thereby reducing the operation cost, improving the energy utilization rate, and supporting the development of sustainable energy.

[0054] Figure 2 Shows the structural schematic diagram of the grid control system based on artificial intelligence provided by the embodiment of the present invention. AsFigure 2 As shown in Figure 2 , the system 200 includes: A data acquisition module 210, configured to acquire real-time power grid data of power grid equipment in the current period; the real-time power grid data includes power production, load demand, voltage fluctuation data, and weather condition data; A prediction module 220, configured to input the real-time power grid data into a power grid state prediction model to predict the current power grid operation state; the current power grid operation state includes whether it is a peak load period and whether the voltage is abnormal; wherein, the power grid state prediction model is obtained by pre-training a deep learning network according to power grid state sample data, and the power grid state sample data includes power grid historical operation data and corresponding sample labels; A power strategy module 230, configured to input the current power grid operation state and the real-time power grid data into a power grid strategy dynamic optimization module to generate a new power distribution strategy; An execution module 240, configured to set power grid control parameters according to the new power distribution strategy and the real-time power grid data, and execute the new power distribution strategy; A resource optimization module 250, configured to, when a preset optimization period arrives, perform dynamic optimization of power grid resources according to the current power grid operation state and real-time resource data by using a power grid resource dynamic optimization algorithm to obtain an optimized power grid resource optimization plan; wherein, the power grid resource dynamic optimization algorithm includes a hybrid algorithm combining model predictive control and reinforcement learning, and the real-time resource data includes load demand information, new energy power generation, remaining capacity of the energy storage system, power grid topology structure, and external factor data.

[0055] In an optional manner, the device further includes: a collection module, configured to collect power grid historical operation data; the power grid historical operation data includes historical power production, historical load demand, historical voltage fluctuation data, and historical weather condition data; A label annotation module, configured to add corresponding sample labels to the power grid historical operation data to obtain power grid state sample data; the sample labels include corresponding power grid historical operation states, and the power grid historical operation state labels include peak load period labels and voltage anomaly labels; An iterative training module, configured to input the power grid state sample data into a deep learning network for iterative training, and adjust parameters of the deep learning network to obtain a trained power grid state prediction model.

[0056] In an optional manner, the deep learning network includes a time series prediction module, a spatial correlation prediction module, and an anomaly detection module; wherein, the time series prediction module includes one of a self-attention mechanism and a CNN-LSTM hybrid model; the spatial correlation prediction module includes a GNN algorithm; the anomaly detection module includes a variational autoencoder; Inputting the power grid state sample data into a deep learning network for iterative training and adjusting the parameters of the deep learning network to obtain a trained power grid state prediction model, including: Inputting the power grid state sample data into a self-attention mechanism or a CNN-LSTM hybrid model for training to obtain a time series prediction module for predicting the prediction results during peak load periods; Inputting the power grid state sample data into a GNN algorithm for training to obtain a spatial association prediction module for locating power grid faults; Inputting the power grid state sample data into a variational autoencoder for training to obtain an anomaly detection module for detecting abnormal events.

[0057] In an alternative manner, the collecting of the real-time power grid data of the current period of the power grid equipment includes: Respectively collecting the power generation amount information of each power generation equipment in the power grid to obtain power generation amount data; the power generation equipment includes thermal power plants, hydropower stations, wind power stations and solar photovoltaic systems; Obtaining real-time load data from the load monitoring equipment of the substation and the distribution network to obtain the load demand; Monitoring and recording voltage fluctuations through voltage sensors installed at key nodes of the power grid to collect voltage fluctuation data; Obtaining weather data affecting energy production and consumption through communication with a meteorological service center.

[0058] In an alternative manner, the inputting of the current power grid operation state and the real-time power grid data into the power grid strategy dynamic optimization module to generate a new power distribution strategy includes: Evaluating the effectiveness of the current power distribution strategy according to the current power grid operation state and the real-time power grid data and identifying improvement points; Generating a new power distribution strategy according to the improvement points and the current power grid operation state according to a preset adjustment strategy; Conducting a simulation test on the new power distribution strategy; When the test passes, applying the new power distribution strategy to power grid management.

[0059] In an alternative manner, when a preset optimization period is reached, according to the current power grid operation state and real-time resource data, a power grid resource dynamic optimization algorithm is used for power grid resource dynamic optimization to obtain an optimized power grid resource optimization plan, including: Determining a multi-objective function according to traditional power generation fuel costs, energy storage loss costs, new energy curtailment penalties and voltage deviations; wherein, power balance constraints, energy storage dynamic constraints and power grid equipment limit constraints are set; When the preset optimization period is reached, according to the current power grid operation state and real-time resource data, an enhanced learning adaptive strategy is adopted to optimize the multi-objective function, and an optimized power grid resource optimization plan is obtained.

[0060] In an alternative embodiment, the method further includes: displaying the current power grid operation state and the optimized power grid resource optimization plan in a user interface.

[0061] As Figure 3 shown, in the intelligent power grid optimization management system based on artificial intelligence provided by the embodiments of the present invention, by integrating with various power grid devices, the collection of real-time power grid data is realized, and relying on the DPM (Data Processor) of the data processing module, the data is processed, so as to effectively improve the data processing speed and data processing accuracy.

[0062] The embodiments of the present invention collect real-time power grid data of power grid devices in the current period; input the real-time power grid data into a power grid state prediction model to predict the current power grid operation state; input the current power grid operation state and the real-time power grid data into a power grid strategy dynamic optimization module to generate a new power distribution strategy; set power grid control parameters according to the new power distribution strategy and the real-time power grid data, and execute the new power distribution strategy; when the preset optimization period is reached, according to the current power grid operation state and real-time resource data, a power grid resource dynamic optimization algorithm is adopted to perform power grid resource dynamic optimization, and an optimized power grid resource optimization plan is obtained, which can improve the data processing ability, prediction accuracy and adaptive ability of the power grid management system, while optimizing the user interaction interface, enhancing the integration and management of new energy, realizing real-time monitoring of the power grid operation state, load prediction, fault diagnosis and optimal allocation of power resources, thereby reducing the operation cost, improving the energy utilization rate, and supporting the development of sustainable energy.

[0063] Figure 4 The structural schematic diagram of the computer device provided by the embodiments of the present invention is shown. The specific implementation of the computer device is not limited in the specific embodiments of the present invention.

[0064] As Figure 4 shown, the computer device may include: a processor 402, a communications interface 404, a memory 406, and a communication bus 408.

[0065] Wherein: the processor 402, the communication interface 404, and the memory 406 communicate with each other through the communication bus 408. The communication interface 404 is used to communicate with network elements of other devices such as clients or other servers. The processor 402 is used to execute the program 410, and specifically can execute the relevant steps in the above embodiments of the power grid control method based on artificial intelligence.

[0066] Specifically, the program 410 may include program code, and the program code includes computer-executable instructions.

[0067] The processor 402 may be a central processing unit CPU, or a specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present invention. One or more processors included in the computer device may be of the same type of processor, such as one or more CPUs; or may be of different types of processors, such as one or more CPUs and one or more ASICs.

[0068] The memory 406 is used to store the program 410. The memory 406 may include a high-speed RAM memory, and may also include a non-volatile memory, such as at least one disk memory.

[0069] The program 410 can specifically be called by the processor 402 to cause the computer device to perform the following operations: Collect real-time power grid data of the power grid equipment in the current period; the real-time power grid data includes power generation, load demand, voltage fluctuation data, and weather condition data; Input the real-time power grid data into the power grid state prediction model to predict the current power grid operation state; the current power grid operation state includes whether it is a peak load period and whether the voltage is abnormal; wherein, the power grid state prediction model is obtained by training a deep learning network in advance according to power grid state sample data, and the power grid state sample data includes power grid historical operation data and corresponding sample labels; Input the current power grid operation state and the real-time power grid data into the power grid policy dynamic optimization module to generate a new power distribution policy; Set power grid control parameters according to the new power distribution policy and the real-time power grid data, and execute the new power distribution policy; When the preset optimization period is reached, according to the current power grid operation state and real-time resource data, a dynamic optimization algorithm for power grid resources is used to perform dynamic optimization of power grid resources, and an optimized power grid resource optimization plan is obtained; wherein, the dynamic optimization algorithm for power grid resources includes a hybrid algorithm combining model predictive control and reinforcement learning, and the real-time resource data includes load demand information, new energy power generation, remaining capacity of the energy storage system, power grid topology structure, and external factor data.

[0070] In an alternative manner, before predicting the current power grid operation state by inputting the real-time power grid data into the power grid state prediction model, the method further includes: Collect power grid historical operation data; the power grid historical operation data includes historical power generation, historical load demand, historical voltage fluctuation data, and historical weather condition data; add corresponding sample labels to the power grid historical operation data to obtain power grid state sample data; the sample labels include the corresponding power grid historical operation states, and the power grid historical operation state labels include load peak period labels and voltage anomaly labels; Input the power grid state sample data into a deep learning network for iterative training, and adjust the parameters of the deep learning network to obtain a trained power grid state prediction model.

[0071] In an alternative manner, the deep learning network includes a time series prediction module, a spatial correlation prediction module, and an anomaly detection module; wherein, the time series prediction module includes one of a self-attention mechanism and a CNN-LSTM hybrid model; the spatial correlation prediction module includes a GNN algorithm; the anomaly detection module includes a variational autoencoder; The step of inputting the power grid state sample data into a deep learning network for iterative training and adjusting the parameters of the deep learning network to obtain a trained power grid state prediction model includes: Input the power grid state sample data into the self-attention mechanism or the CNN-LSTM hybrid model for training to obtain a time series prediction module for predicting the load peak period prediction result; Input the power grid state sample data into the GNN algorithm for training to obtain a spatial correlation prediction module for locating power grid faults; Input the power grid state sample data into the variational autoencoder for training to obtain an anomaly detection module for detecting abnormal events.

[0072] In an alternative manner, the step of collecting the real-time power grid data of the current period of the power grid equipment includes: Respectively collect the power generation information of each power generation equipment in the power grid to obtain power generation data; the power generation equipment includes thermal power plants, hydropower stations, wind power stations, and solar photovoltaic systems; Obtain real-time load data from the load monitoring devices of the substation and the distribution network, and obtain the load demand; Monitor and record voltage fluctuations through voltage sensors installed at key nodes of the power grid, and collect voltage fluctuation data; Obtain weather data affecting energy production and consumption through communication with the meteorological service center.

[0073] In an alternative manner, inputting the current power grid operating state and the real-time power grid data into the power grid strategy dynamic optimization module to generate a new power distribution strategy includes: Evaluate the effectiveness of the current power distribution strategy based on the current power grid operating state and the real-time power grid data, and identify improvement points; Generate a new power distribution strategy according to the improvement points and the current power grid operating state according to a preset adjustment strategy; Conduct a simulation test on the new power distribution strategy; When the test passes, apply the new power distribution strategy to power grid management.

[0074] In an alternative manner, when a preset optimization period is reached, perform dynamic optimization of the power grid resources using a power grid resource dynamic optimization algorithm based on the current power grid operating state and real-time resource data to obtain an optimized power grid resource optimization plan, including: Determine a multi-objective function based on traditional power generation fuel costs, energy storage loss costs, new energy curtailment penalties, and voltage deviations; among them, set power balance constraints, energy storage dynamic constraints, and power grid equipment limit constraints; When a preset optimization period is reached, perform objective optimization on the multi-objective function using a reinforcement learning adaptive strategy based on the current power grid operating state and real-time resource data to obtain an optimized power grid resource optimization plan.

[0075] In an alternative manner, the method further includes: displaying the current power grid operating state and the optimized power grid resource optimization plan in the user interface.

[0076] In an embodiment of the present invention, real-time grid data of grid equipment in the current period is collected; the real-time grid data is input into a grid state prediction model to predict the current grid operation state; the current grid operation state and the real-time grid data are input into a grid strategy dynamic optimization module to generate a new power distribution strategy; according to the new power distribution strategy and the real-time grid data, grid control parameters are set and the new power distribution strategy is executed; when a preset optimization period is reached, according to the current grid operation state and real-time resource data, a grid resource dynamic optimization algorithm is used for grid resource dynamic optimization to obtain an optimized grid resource optimization plan, which can improve the data processing ability, prediction accuracy and adaptive ability of the grid management system, while optimizing the user interface, enhancing the integration and management of new energy, and realizing real-time monitoring of the grid operation state, load prediction, fault diagnosis and optimal allocation of power resources, thereby reducing the operation cost, improving the energy utilization rate and supporting the development of sustainable energy.

[0077] An embodiment of the present invention provides a computer-readable storage medium, which stores at least one executable instruction. When the executable instruction runs on a computer device, the computer device is enabled to execute the grid control method based on artificial intelligence in any of the above method embodiments.

[0078] The executable instruction can specifically be used to enable the computer device to perform the following operations: Collect real-time grid data of grid equipment in the current period; the real-time grid data includes power production, load demand, voltage fluctuation data and weather condition data; Input the real-time grid data into a grid state prediction model to predict the current grid operation state; the current grid operation state includes whether it is a peak load period and whether the voltage is abnormal; wherein, the grid state prediction model is obtained by training a deep learning network in advance according to grid state sample data, and the grid state sample data includes grid historical operation data and corresponding sample labels; Input the current grid operation state and the real-time grid data into a grid strategy dynamic optimization module to generate a new power distribution strategy; According to the new power distribution strategy and the real-time grid data, set grid control parameters and execute the new power distribution strategy; When a preset optimization period is reached, according to the current grid operation state and real-time resource data, a grid resource dynamic optimization algorithm is used for grid resource dynamic optimization to obtain an optimized grid resource optimization plan; wherein, the grid resource dynamic optimization algorithm includes a hybrid algorithm combining model predictive control and reinforcement learning, and the real-time resource data includes load demand information, new energy power generation, remaining capacity of the energy storage system, grid topology structure and external factor data.

[0079] In an alternative manner, before inputting the real-time grid data into the grid state prediction model to predict the current grid operating state, the method further includes: Collecting historical grid operation data; the historical grid operation data includes historical power generation, historical load demand, historical voltage fluctuation data, and historical weather condition data; Adding corresponding sample labels to the historical grid operation data to obtain grid state sample data; the sample labels include the corresponding historical grid operating states, and the historical grid operating state labels include peak load period labels and voltage anomaly labels; Inputting the grid state sample data into a deep learning network for iterative training, and adjusting the parameters of the deep learning network to obtain a trained grid state prediction model.

[0080] In an alternative manner, the deep learning network includes a time series prediction module, a spatial correlation prediction module, and an anomaly detection module; wherein, the time series prediction module includes one of a self-attention mechanism and a CNN-LSTM hybrid model; the spatial correlation prediction module includes a GNN algorithm; the anomaly detection module includes a variational autoencoder; The inputting the grid state sample data into a deep learning network for iterative training, and adjusting the parameters of the deep learning network to obtain a trained grid state prediction model includes: Inputting the grid state sample data into a self-attention mechanism or a CNN-LSTM hybrid model for training to obtain a time series prediction module for predicting peak load period prediction results; Inputting the grid state sample data into a GNN algorithm for training to obtain a spatial correlation prediction module for locating grid faults; Inputting the grid state sample data into a variational autoencoder for training to obtain an anomaly detection module for detecting abnormal events.

[0081] In an alternative manner, the collecting of the real-time grid data of the grid equipment in the current period includes: Respectively collecting the power generation information of each power generation equipment in the grid to obtain power generation data; the power generation equipment includes thermal power plants, hydropower stations, wind power stations, and solar photovoltaic systems; Obtaining real-time load data from load monitoring devices of substations and distribution networks to obtain the load demand; Monitoring and recording voltage fluctuations through voltage sensors installed at key grid nodes to collect voltage fluctuation data; Obtaining weather data affecting energy production and consumption through communication with a meteorological service center.

[0082] In an alternative approach, inputting the current power grid operating state and the real-time power grid data into the power grid strategy dynamic optimization module to generate a new power distribution strategy includes: Evaluating the effectiveness of the current power distribution strategy based on the current power grid operating state and the real-time power grid data to identify improvement points; Generating a new power distribution strategy according to the improvement points and the current power grid operating state in accordance with a preset adjustment strategy; Performing a simulation test on the new power distribution strategy; When the test passes, applying the new power distribution strategy to power grid management.

[0083] In an alternative approach, when a preset optimization period is reached, performing dynamic optimization of power grid resources using a power grid resource dynamic optimization algorithm based on the current power grid operating state and real-time resource data to obtain an optimized power grid resource optimization plan, including: Determining a multi-objective function based on traditional power generation fuel costs, energy storage loss costs, new energy curtailment penalties, and voltage deviation; wherein, setting power balance constraints, energy storage dynamic constraints, and power grid equipment limit constraints; When a preset optimization period is reached, performing target optimization on the multi-objective function using a reinforcement learning adaptive strategy based on the current power grid operating state and real-time resource data to obtain an optimized power grid resource optimization plan.

[0084] In an alternative approach, the method further includes: displaying the current power grid operating state and the optimized power grid resource optimization plan in a user interface.

[0085] In an embodiment of the present invention, by collecting real-time power grid data of power grid equipment in the current period; inputting the real-time power grid data into a power grid state prediction model to predict the current power grid operating state; inputting the current power grid operating state and the real-time power grid data into a power grid strategy dynamic optimization module to generate a new power distribution strategy; setting power grid control parameters according to the new power distribution strategy and the real-time power grid data, and executing the new power distribution strategy; when a preset optimization period is reached, performing dynamic optimization of power grid resources using a power grid resource dynamic optimization algorithm based on the current power grid operating state and real-time resource data to obtain an optimized power grid resource optimization plan, it is possible to improve the data processing ability, prediction accuracy, and adaptive ability of the power grid management system, while optimizing the user interaction interface, enhancing the integration and management of new energy, realizing real-time monitoring of the power grid operating state, load forecasting, fault diagnosis, and optimal allocation of power resources, thereby reducing operating costs, improving energy utilization efficiency, and supporting the development of sustainable energy.

[0086] An embodiment of the present invention provides a power grid control device based on artificial intelligence, which is used to execute the above-mentioned power grid control method based on artificial intelligence.

[0087] An embodiment of the present invention provides a computer program, which can be called by a processor to make a computer device execute the power grid control method based on artificial intelligence in any of the above method embodiments.

[0088] An embodiment of the present invention provides a computer program product. The computer program product includes a computer program stored on a computer-readable storage medium. The computer program includes program instructions. When the program instructions run on a computer, the computer is made to execute the power grid control method based on artificial intelligence in any of the above method embodiments.

[0089] The algorithms or displays provided herein are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used in conjunction with the teachings provided herein. The structure required to construct such systems will be apparent from the above description. In addition, the embodiments of the present invention are not directed to any specific programming language. It should be understood that the content of the present invention described herein can be implemented using various programming languages, and the description of the specific language above is to disclose the best mode of the present invention.

[0090] In the specification provided herein, a large number of specific details are set forth. However, it can be understood that the embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures, and technologies have not been shown in detail so as not to obscure the understanding of this specification.

[0091] Similarly, it should be understood that, in order to streamline the present invention and assist in understanding one or more of the various inventive aspects, in the above description of the exemplary embodiments of the present invention, the various features of the embodiments of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, the disclosed method should not be construed as reflecting the intention that the claimed invention requires more features than are expressly recited in each claim.

[0092] Those skilled in the art can understand that the modules in the devices in the embodiments can be adaptively changed and arranged in one or more devices different from those of the embodiments. The modules or units or components in the embodiments can be combined into one module or unit or component, and can also be divided into multiple sub-modules or sub-units or sub-components. Except that at least some of such features and / or processes or units are mutually exclusive, any combination can be adopted to combine all the features disclosed in this specification (including the accompanying claims, abstract and drawings), and all the processes or units of any method or device so disclosed. Unless otherwise clearly stated, each feature disclosed in this specification (including the accompanying claims, abstract and drawings) can be replaced by an alternative feature that provides the same, equivalent or similar purpose.

[0093] It should be noted that the above embodiments illustrate the present invention rather than limit the present invention, and those skilled in the art can design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in the claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention can be implemented by means of hardware including several different elements and by means of a suitably programmed computer. In the unit claims listing several devices, several of these devices can be embodied by the same item of hardware. The use of the words first, second, and third, etc. does not denote any order. These words can be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be construed as limiting the order of execution.

Claims

1. An artificial intelligence-based power grid control method, characterized in that The method includes: Collecting real-time power grid data of power grid equipment in the current period; the real-time power grid data includes power generation, load demand, voltage fluctuation data, and weather condition data; Inputting the real-time power grid data into a power grid state prediction model to predict the current power grid operation state; the current power grid operation state includes peak load periods, anomaly detection information, and power grid fault location information; wherein, the power grid state prediction model is obtained by pre-training a deep learning network according to power grid state sample data, and the power grid state sample data includes historical power grid operation data and corresponding sample labels; Inputting the current power grid operation state and the real-time power grid data into a power grid strategy dynamic optimization module to generate a new power distribution strategy; Setting power grid control parameters according to the new power distribution strategy and the real-time power grid data, and implementing the new power distribution strategy; When a preset optimization period is reached, performing dynamic optimization of power grid resources according to the current power grid operation state and real-time resource data by using a power grid resource dynamic optimization algorithm to obtain an optimized power grid resource optimization plan; wherein, the power grid resource dynamic optimization algorithm includes a hybrid algorithm combining model predictive control and reinforcement learning, and the real-time resource data includes load demand information, new energy power generation, remaining capacity of the energy storage system, power grid topology structure, and external factor data.

2. The method according to claim 1, wherein Before inputting the real-time power grid data into the power grid state prediction model to predict the current power grid operation state, the method further includes: Collecting historical power grid operation data; the historical power grid operation data includes historical power generation, historical load demand, historical voltage fluctuation data, and historical weather condition data; Adding corresponding sample labels to the historical power grid operation data to obtain power grid state sample data; the sample labels include corresponding historical power grid operation states, and the historical power grid operation state labels include peak load period labels, anomaly detection information labels, and power grid fault location information labels; Inputting the power grid state sample data into a deep learning network for iterative training, and adjusting the parameters of the deep learning network to obtain a trained power grid state prediction model.

3. The method according to claim 1, wherein The deep learning network includes a time series prediction module, a spatial correlation prediction module, and an anomaly detection module; wherein, the time series prediction module includes one of a self-attention mechanism and a CNN-LSTM hybrid model; the spatial correlation prediction module includes a GNN algorithm; the anomaly detection module includes a variational autoencoder; The inputting the power grid state sample data into a deep learning network for iterative training, and adjusting the parameters of the deep learning network to obtain a trained power grid state prediction model includes: Inputting the power grid state sample data into the self-attention mechanism or the CNN-LSTM hybrid model for training to obtain a time series prediction module for predicting peak load period prediction results; Inputting the power grid state sample data into the GNN algorithm for training to obtain a spatial correlation prediction module for locating power grid faults; Input the power grid state sample data into a variational autoencoder for training to obtain an anomaly detection module for detecting abnormal events.

4. The method according to any one of claims 1 to 3, characterized in that, Collect the real-time power grid data of the current period of the power grid equipment, including: Respectively collect the power generation information of each power generation equipment in the power grid to obtain power generation data; the power generation equipment includes thermal power plants, hydropower stations, wind power stations, and solar photovoltaic systems; Obtain real-time load data from the load monitoring equipment of the substation and the distribution network to obtain the load demand; Monitor and record voltage fluctuations through voltage sensors installed at key nodes of the power grid to collect voltage fluctuation data; Obtain weather data affecting energy production and consumption through communication with the meteorological service center.

5. The method according to any one of claims 1-3, characterized in that, Input the current power grid operation state and the real-time power grid data into the power grid strategy dynamic optimization module to generate a new power distribution strategy, including: Evaluate the effectiveness of the current power distribution strategy according to the current power grid operation state and the real-time power grid data, and identify improvement points; Generate a new power distribution strategy according to the improvement points and the current power grid operation state according to a preset adjustment strategy; Conduct a simulation test on the new power distribution strategy; When the test passes, apply the new power distribution strategy to power grid management.

6. The method according to any one of claims 1 to 3, characterized in that When the preset optimization period arrives, perform dynamic optimization of the power grid resources using the power grid resource dynamic optimization algorithm according to the current power grid operation state and real-time resource data to obtain an optimized power grid resource optimization plan, including: Determine a multi-objective function according to the traditional power generation fuel cost, energy storage loss cost, and voltage deviation; among them, set power balance constraints, energy storage dynamic constraints, and power grid equipment limit constraints; When the preset optimization period arrives, perform target optimization on the multi-objective function using a reinforcement learning adaptive strategy according to the current power grid operation state and real-time resource data to obtain an optimized power grid resource optimization plan.

7. The method according to any one of claims 1-3, characterized in that, The method further includes: Display the current power grid operation state and the optimized power grid resource optimization plan in the user interface.

8. An artificial intelligence-based power grid control system, characterized in that, The system includes: A collection module for collecting the real-time power grid data of the current period of the power grid equipment; the real-time power grid data includes power production, load demand, voltage fluctuation data, and weather condition data; A prediction module for inputting the real-time power grid data into a power grid state prediction model to predict the current power grid operation state; the current power grid operation state includes peak load periods, anomaly detection information, and power grid fault location information; among them, the power grid state prediction model is obtained by pre-training a deep learning network according to power grid state sample data, and the power grid state sample data includes power grid historical operation data and corresponding sample labels; A power strategy module for inputting the current power grid operation state and the real-time power grid data into a power grid strategy dynamic optimization module to generate a new power distribution strategy; An execution module for setting power grid control parameters according to the new power distribution strategy and the real-time power grid data and executing the new power distribution strategy; A resource optimization module, which is used to perform dynamic optimization of grid resources by using a dynamic grid resource optimization algorithm according to the current grid operation state and real-time resource data when a preset optimization period arrives, so as to obtain an optimized grid resource optimization plan; wherein, the dynamic grid resource optimization algorithm includes a hybrid algorithm combining model predictive control and reinforcement learning, and the real-time resource data includes load demand information, new energy power generation, remaining capacity of the energy storage system, grid topology structure, and external factor data.

9. A computer device, characterized in that, It includes: A processor, a memory, a communication interface, and a communication bus. The processor, the memory, and the communication interface complete communication with each other through the communication bus; The memory is used to store at least one executable instruction, and the executable instruction causes the processor to execute the operations of the artificial intelligence-based grid control method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, At least one executable instruction is stored in the storage medium. When the executable instruction runs on a computer device, the computer device is caused to execute the operations of the artificial intelligence-based grid control method according to any one of claims 1-7.

Images