Digital twinning-based power grid support performance evaluation system and method

Abstract

The invention relates to the technical field of power systems, and discloses a digital twinning-based power grid support performance evaluation system and method. Comprising a data acquisition and processing module, a digital twin model construction and integration module, a multi-scene simulation and prediction evaluation module, an intelligent scheduling and fault prediction module, a multi-source data fusion and big data analysis module, a performance evaluation and optimization decision module, and a visualization and man-machine interaction module. The data acquisition module is used for acquiring real-time states, load data and external environment data of equipment in a physical power grid. Through an intelligent scheduling algorithm, a digital twinborn model and a fault prediction technology, the system can monitor and optimize a new energy access process in real time, solves the challenge of new energy power generation fluctuation on power grid stability, and the algorithm can dynamically adjust power grid load distribution and predict potential faults in advance under the condition of centralized access or fluctuation of new energy. And the stability of power transmission and the safety of equipment are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular to a power grid support performance evaluation system and method based on digital twins. Background Art

[0002] The transformation of the global energy structure and the access to new energy sources have increased the challenges faced by modern power grids. Traditional power systems are usually based on fixed rules or experience-driven models in terms of power generation and transmission and distribution planning, dispatching and management. In particular, renewable energy such as wind and solar energy, due to their intermittent and volatile nature, puts pressure on the stability and reliability of the power grid. When faced with fluctuations, traditional power grid dispatching and management systems have delayed responses and low dispatching efficiency, making it difficult to meet the operating needs of modern power grids.

[0003] As an emerging technical means, digital twin technology collects the status data of physical systems in real time, creates a digital model corresponding to the physical system in virtual space, and realizes the integration of virtual and real. This technology can provide global and refined simulation and optimization tools for power grid operation, and plays an important role in power grid management. By combining intelligent scheduling algorithms and big data analysis technologies, digital twins can simulate, predict and optimize the operating status of power grids in real time, and have significant advantages in handling large-scale new energy access, fault prediction and optimized scheduling.

[0004] At present, digital twin technology has been initially applied to certain aspects of the power grid, but the existing system lacks a global and intelligent scheduling optimization mechanism. Especially when faced with complex working conditions and multi-dimensional operating goals, it is still unable to achieve real-time global optimization and fault prediction. In addition, the complexity of data such as power load, power generation, and meteorology has increased. How to effectively use massive data for efficient analysis and prediction has become an urgent problem to be solved.

[0005] Therefore, those skilled in the art provide a power grid support performance evaluation system and method based on digital twins to solve the problems raised in the above background technology. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention provides a power grid support performance evaluation system and method based on digital twins to solve the problems raised in the above background technology.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a power grid support performance evaluation system based on digital twins, including a data acquisition and processing module, a digital twin model construction and integration module, a multi-scenario simulation and prediction evaluation module, an intelligent scheduling and fault prediction module, a multi-source data fusion and big data analysis module, a performance evaluation and optimization decision module, and a visualization and human-computer interaction module: The data acquisition and processing module is used to collect the real-time status of the equipment in the physical power grid, load data and external environment data, and clean and standardize the collected data; The digital twin model construction and integration module includes: Micro-level models are used to build twin models of key grid equipment; Meso-level models for twin simulation of local grid power flow and voltage stability; The macro-level model is used for twin simulation of load balancing and power generation scheduling of the entire network. The optimization model is calculated through the following power balance equation: , in, is the power output of generator i, is the power demand of load node j; The multi-scenario simulation and prediction evaluation module is used to generate a simulation situation of power grid operation under complex scenarios and evaluate power grid performance based on the simulation results; The intelligent dispatching and fault prediction module is used to optimize power dispatching through a deep reinforcement learning algorithm and to perform fault prediction through a long short-term memory network; The multi-source data fusion and big data analysis module fuses multi-source data through extended Kalman filtering; The performance evaluation and optimization decision module is used to make optimization decisions based on voltage stability, power loss and operating cost; The visualization and human-computer interaction module is used to display the real-time operation status and dispatching decisions of the power grid, and supports voice interaction through natural language processing technology.

[0008] Preferably, the data acquisition and processing module includes: Sensor network unit, used to collect the operating status of power grid equipment in real time; Edge computing node units, used for local processing and preprocessing of collected data; The cloud data synchronization unit is used to upload the pre-processed data to the cloud for centralized analysis.

[0009] Preferably, the digital twin model constructs a microscopic model of the integrated module to simulate the operating state of the equipment through the following state equation: , in, is the state vector of device i at time t, , are the state matrix and control input matrix, For external interference.

[0010] Preferably, the meso-level model describes the voltage and power flow of the local power grid by the following power flow equation: , in, and is the active power and reactive power of node i, is the node voltage, is the admittance between nodes.

[0011] Preferably, the macro-layer model determines the power output of the generators in the entire grid by the following optimization scheduling formula: , in, is the system operating cost, is the power output of generator i, , , is the cost coefficient of the generator.

[0012] Preferably, the multi-scenario simulation and prediction evaluation module uses the Monte Carlo method to generate a scenario set, and the scenario set Defined as: , in, is the load demand, , is the power of wind and photovoltaic power generation, For meteorological conditions, is the number of simulation scenarios.

[0013] Preferably, the intelligent scheduling module performs scheduling optimization based on a deep reinforcement learning algorithm, and its objective function is: , in, To accumulate rewards, is the immediate reward at time t, is the discount factor.

[0014] Preferably, the fault prediction module performs fault prediction based on a long short-term memory network model, and the hidden state update equation of the long short-term memory network model is: , in, is the hidden state, For input data, , is the weight matrix, For bias.

[0015] Preferably, the multi-source data fusion and big data analysis module performs state fusion of multi-source data through an extended Kalman filter algorithm, and its state update equation is: , in, is the state transfer matrix, is the control input matrix, is the observation matrix, is the Kalman gain.

[0016] A method for evaluating power grid support performance based on digital twins includes the following steps: S1, collect real-time status data of power grid equipment and external environment data; S2. Based on the collected data, build digital twin models at the device layer, local power grid layer and whole network layer; S3, using the Monte Carlo method to generate a variety of complex scenarios for simulation and performance evaluation; S4. Optimize power dispatch strategy based on deep reinforcement learning algorithm; S5. Use the extended Kalman filter algorithm to fuse multi-source data and conduct real-time evaluation of the power grid status; S6. Provide optimized grid operation solutions and display real-time operation status and dispatching decisions through a visualization platform.

[0017] The present invention provides a power grid support performance evaluation system and method based on digital twins, which has the following beneficial effects: 1. Through intelligent scheduling algorithms, digital twin models and fault prediction technology, the system of the present invention can monitor and optimize the new energy access process in real time, solve the challenges of new energy power generation fluctuations to the stability of the power grid. Its algorithm can dynamically adjust the load distribution of the power grid in the case of concentrated access or fluctuations of new energy, and predict potential faults in advance to ensure the stability of power transmission and the safety of equipment.

[0018] 2. The present invention is based on multi-objective optimization algorithms and big data analysis. The system can optimize the power transmission path, reduce power loss, and achieve efficient allocation of power resources. At the same time, the intelligent scheduling mechanism combined with real-time data analysis can minimize the power generation cost and operating loss, and improve the overall economic benefits of the system. The intelligent planning algorithm can reduce operating costs while ensuring the reliability of the power grid, so as to increase the economic benefits of new energy power transmission.

[0019] 3. The present invention is based on multi-objective optimization algorithms and big data analysis. The system can optimize the power transmission path, reduce power loss, and achieve efficient allocation of power resources. At the same time, the intelligent scheduling mechanism combined with real-time data analysis can minimize the power generation cost and operating loss, and improve the overall economic benefits of the system. The intelligent planning algorithm can reduce operating costs while ensuring the reliability of the power grid, so as to increase the economic benefits of new energy power transmission. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a system framework diagram of the present invention; Figure 2 It is a schematic diagram of the data acquisition and processing module of the present invention; Figure 3 Schematic diagram of the digital twin model construction and integration module of the present invention. DETAILED DESCRIPTION

[0021] In order to make the technical personnel in the technical field understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in combination with the drawings in the embodiment of the present invention. Obviously, the described embodiment is a partial embodiment of the present invention, not a complete embodiment. Based on the embodiment of the present invention, other embodiments obtained by ordinary technicians in the field without creative work should fall within the scope of protection of the present invention.

[0022] The present invention is described in detail below in conjunction with the accompanying drawings: Example: Please refer to the attached Figure 1 -Attached Figure 3 , an embodiment of the present invention provides a power grid support performance evaluation system based on digital twins, including a data acquisition and processing module, a digital twin model construction and integration module, a multi-scenario simulation and prediction evaluation module, an intelligent scheduling and fault prediction module, a multi-source data fusion and big data analysis module, a performance evaluation and optimization decision module, and a visualization and human-computer interaction module: The data acquisition and processing module is used to collect the real-time status of the equipment in the physical power grid, load data and external environment data, and clean and standardize the collected data; Digital twin model building and integration modules, including: Micro-level models are used to build twin models of key grid equipment; Meso-level models for twin simulation of local grid power flow and voltage stability; The macro-level model is used for twin simulation of load balancing and power generation scheduling of the entire network. The optimization model is calculated through the following power balance equation: , in, is the power output of generator i, is the power demand of load node j; Multi-scenario simulation and prediction evaluation module, used to generate power grid operation simulation status under complex scenarios and evaluate power grid performance based on simulation results; Intelligent dispatch and fault prediction module, which is used to optimize power dispatch through deep reinforcement learning algorithm and perform fault prediction through long short-term memory network; Multi-source data fusion and big data analysis module, which fuses multi-source data through extended Kalman filtering; Performance evaluation and optimization decision module, used to make optimization decisions based on voltage stability, power loss and operating costs; The visualization and human-computer interaction module is used to display the real-time operating status and dispatch decisions of the power grid, and supports voice interaction through natural language processing technology.

[0023] The benefits of the data acquisition and processing module are to provide high-quality data support for the subsequent digital twin model construction, optimized scheduling and fault prediction of the system, effectively reduce the impact of noise and abnormal data, and improve the accuracy and efficiency of the overall system; Benefits of digital twin model construction and integration modules: Comprehensive dynamic monitoring of the power grid from the device level to the whole network level is realized. The models at all levels work together to accurately simulate the real-time status and changes of the power grid, provide global optimization and monitoring capabilities that combine virtual and real, help cope with complex working conditions of the power grid, and ensure efficient dispatching and transmission of electricity; The benefits of the multi-scenario simulation and predictive assessment module help power grid operators formulate response plans in advance, improve the system's emergency response and fault recovery capabilities, and enhance the stability and security of the power grid; The benefits of the intelligent dispatching and fault prediction module are that it can adaptively adjust power distribution according to real-time and historical data of the power grid, reduce power loss and optimize power generation costs. At the same time, fault prediction can identify potential equipment failures in advance, perform predictive maintenance, reduce power outage time, and ensure the reliability of the power grid; Benefits of multi-source data fusion and big data analysis module: Through big data technology, the system can mine hidden patterns from massive data, conduct grid operation trend forecasting and decision support, and further enhance the intelligent level of grid management and data-driven decision-making capabilities; Benefits of the performance evaluation and optimization decision module Balance the safety, economy and efficiency of power grid operation, effectively reduce the operating cost of the power grid, and improve the economic benefits of new energy power transmission; The benefits of the visualization and human-computer interaction module are that the real-time status, dispatching decisions and fault warnings of the power grid can be intuitively understood through the visualization interface. At the same time, combined with natural language processing technology, the system supports voice interaction, which improves user experience and operational efficiency, and helps to improve the intelligence level and responsiveness of power grid management.

[0024] The data acquisition and processing modules include: Sensor network unit, used to collect the operating status of power grid equipment in real time; Edge computing node units, used for local processing and preprocessing of collected data; The cloud data synchronization unit is used to upload the pre-processed data to the cloud for centralized analysis.

[0025] Benefits of sensor network units Through sensors deployed on different devices, the system can accurately grasp the real-time operation status of the power grid and support subsequent modeling, scheduling and fault prediction operations. In addition, real-time data collection improves the accuracy and timeliness of power grid management and contributes to the intelligent monitoring and operation and maintenance of the power grid; Benefits of edge computing node units Data processing at the edge can significantly reduce data transmission delays and improve system response speed. Especially when facing large amounts of data, edge computing can share data processing pressure, reduce dependence on the central cloud, and improve the overall operating efficiency and stability of the system. The advantage of the cloud-based data synchronization unit is that it can integrate, analyze and further process data from multiple edge nodes, effectively supporting digital twin modeling, simulation prediction, optimization scheduling and other functions across the entire network, realizing cross-regional and global power grid data sharing and intelligent management, and improving the coordination of power grid management and data availability.

[0026] The micro-level model of the digital twin model construction and integration module simulates the operating state of the equipment through the following state equations: , in, is the state vector of device i at time t, , are the state matrix and control input matrix, For external interference.

[0027] The function of the state equation formula of the micro-level model is to describe the time evolution process of the equipment. According to the current state and external control variables, the state of the equipment in the next time step is predicted, so that the system can track the operating status of the equipment in real time, evaluate the performance and health of the equipment, and determine whether there are abnormalities or potential failures. The system can dynamically adjust the operating status of the equipment to keep it in the optimal operating state. According to the operation requirements of the power grid, the equipment can respond to external signals for real-time adjustment to optimize the overall performance of the power grid. By combining the difference between the expected state and the law of state evolution, the system can detect potential abnormalities of the equipment in advance, predict failures, and provide support for preventive maintenance of the equipment; The advantage of the state equation formula of the micro-level model is that it can describe the operating status of the equipment in detail, and the system can Accurately determine whether it is in normal working condition and identify potential performance degradation or abnormal behavior; control input of state equation The system can adjust and optimize the equipment status according to the real-time needs of the power grid, which helps the equipment maintain efficient operation under different load conditions and avoid excessive energy consumption or overload. By synchronously calculating and monitoring the state equations of multiple devices, the system can achieve collaborative optimization between different devices, enabling the power grid to achieve more efficient power dispatch and resource allocation at the equipment level, enhancing the flexibility and adaptability of the system.

[0028] The meso-level model describes the voltage and power flow of the local power grid through the following power flow equation: , in, and is the active power and reactive power of node i, is the node voltage, is the admittance between nodes.

[0029] The meso-level model power flow equation describes the power flow in the power grid, and according to the system admittance matrix , calculate the voltage changes between nodes, which helps to analyze the voltage differences, power flow direction and power loss between different nodes; by calculating the voltage of the node The system can analyze the voltage stability in the power grid in real time and determine whether each node is within a reasonable voltage range; the power flow equation can provide a basis for power dispatching. By analyzing the power flow and voltage distribution, the system can adjust the power generation and load distribution according to demand, ensuring that the power system can meet the load demand while operating efficiently; Benefits of the meso-level model flow equation: Real-time calculation of the power flow and voltage level of each node, helping grid managers to understand the operating status of the grid. The system can quickly identify abnormal nodes and take timely measures to avoid large-scale failures. By analyzing the voltage deviation of the nodes, the system can predict nodes with unstable voltage and adjust load distribution or perform voltage regulation in advance. It can keep the voltage of each node within a safe range to prevent equipment damage or system collapse caused by unstable voltage. It provides power flow information for the entire network to help the system optimize power dispatching and power distribution. By optimizing the power flow between different nodes, power transmission can be more efficient, reduce power loss, and improve the energy utilization efficiency of the grid. In complex scenarios such as extreme weather and sudden load changes, the flow equation can simulate the changes in voltage and power flow at different nodes, helping grid operators to respond to load fluctuations in advance and maintain the stability and safety of the grid. Through flow analysis, the system can identify loss points in power flow, and reduce unnecessary losses by adjusting the power transmission path and optimizing power generation, thereby improving the economic benefits of the grid.

[0030] The macro-level model determines the power output of the entire grid generator through the following optimization scheduling formula: , in, is the system operating cost, is the power output of generator i, , , is the cost coefficient of the generator.

[0031] The role of the macro-level model optimization scheduling formula is to adjust the power output of the generator to ensure that the total operating cost of the generator set is minimized while meeting the load demand of the entire network; by optimizing power scheduling, the power output of the generator is ensured to be consistent with its operating cost. For generators with lower costs, the system will give priority to their output of higher power. Conversely, generators with higher costs will be arranged to output lower power, thereby achieving a balance between the economy and efficiency of the system; balancing the power output of the generator with the power demand of the system, and ensuring the balance between the total power generation power of the system and the load demand by optimizing the power distribution of the generator, so as to achieve a stable power supply; Benefits of the macro-level model optimization scheduling formula The system dynamically allocates power output according to the cost characteristics of different generators, minimizing the overall power generation cost, avoiding unnecessary high-cost operation, and significantly reducing the operating costs of the system; the system can quickly calculate the optimal power generation plan to achieve efficient use of power generation resources, and quickly respond to load fluctuations and changes in power demand according to real-time needs to ensure stable operation of the system; the system can quickly adjust the generator power output through this formula to ensure a dynamic balance between power generation and power demand, avoid situations where the load exceeds the power supply capacity, and improve the grid's scheduling flexibility and response capabilities during peak loads; achieve efficient coordination between traditional and new energy, and by reducing the power output of traditional generators, when new energy generation is sufficient, the system can operate at the optimal economy and improve the proportion of clean energy and sustainability of the overall power grid.

[0032] The multi-scenario simulation and prediction evaluation module uses the Monte Carlo method to generate scenario sets. Defined as: , in, is the load demand, , is the power of wind and photovoltaic power generation, For meteorological conditions, is the number of simulation scenarios.

[0033] The role of the scenario set formula is to generate operating scenarios through sampling. The scenarios integrate various changing conditions faced by the power grid in actual operation, including different load demands, wind and photovoltaic power generation, and various meteorological conditions, so that the system can fully simulate various operating conditions, covering conventional operation, extreme weather, new energy power generation fluctuations and other scenarios; by generating multiple scenarios, the system can perform power flow analysis, voltage stability assessment and dispatch simulation for each scenario, and assessing the performance of the power grid under various complex conditions helps to identify potential risks and ensure that the power grid can operate stably under different conditions; through scenario simulation, the operating risks of the power grid under different conditions can be quantified, and by counting the key operating indicators under each scenario, the system can more accurately predict system risks and uncertainties under different conditions; The benefits of the scenario set formula are that it helps grid operators predict and analyze grid performance in complex scenarios in advance, and enhances the system's ability to respond to extreme conditions; the system can detect problems such as voltage instability and power flow imbalance in different scenarios, and optimize scheduling in a targeted manner to ensure that the grid maintains stable and reliable operation under different operating conditions; the system can effectively evaluate the impact of new energy access, design corresponding response strategies, and improve the utilization efficiency of new energy power generation; the system can quantify the risk level of the grid under different operating conditions, help operators formulate more scientific scheduling plans and emergency plans, reduce the probability of grid failures, and improve operational safety; the system can adjust simulation scenarios based on factors, simulate the grid operation performance under different weather conditions, and ensure that the system can adapt to changes in meteorological conditions.

[0034] The intelligent scheduling module performs scheduling optimization based on the deep reinforcement learning algorithm, and its objective function is: , in, To accumulate rewards, is the immediate reward at time t, is the discount factor.

[0035] Benefits of deep reinforcement learning algorithm formula The system can autonomously learn and adjust the dispatching strategy according to the actual operating conditions, automatically adapt to the dynamic changes of power load and the volatility of renewable energy power generation, and improve the intelligence level of power grid dispatching; the system can make global optimization in multi-time dispatching decisions, improve the overall operating efficiency of the power grid, reduce power loss, and enhance load balancing capabilities; the system can weigh current and future benefits during dispatching, and avoid long-term instability caused by short-term benefits; the system optimizes dispatching when renewable energy power generation is sufficient to maximize the utilization rate of renewable energy and reduce the cost and emissions of traditional power generation; guide the system to reduce unnecessary power loss and power generation costs during the dispatching process; the system maximizes the dispatching strategy based on cumulative rewards, flexibly responds to various emergencies, and ensures the stability and efficient operation of the power grid.

[0036] The fault prediction module performs fault prediction based on the long short-term memory network model. The hidden state update equation of the long short-term memory network model is: , in, is the hidden state, For input data, , is the weight matrix, For bias.

[0037] The algorithmic formula of the long short-term memory network works through the hidden state Updates that capture the current moment in the time series Unlike traditional neural networks, LSTM networks are particularly suitable for processing time series data and can memorize the long-term historical status of equipment operation. LSTM networks can continuously update the health status of the equipment and generate new hidden states. , allowing the system to dynamically track the health of the equipment while monitoring its operating status in real time, and predict failures based on its historical performance; Benefits of the algorithm formula of the long short-term memory network: the system can identify abnormal trends in equipment operation in advance and issue early warnings in time to avoid sudden failures; long-term memory capabilities help reduce the occurrence of false alarms and missed alarms, making the system's predictions of equipment failures more accurate and improving the reliability and safety of power grid operation; long short-term memory networks can process complex, noisy time series data, helping the system to predict the health of equipment in complex environments; the system can continuously predict the future operating status and potential failures of the equipment during operation, thereby achieving real-time fault prediction; the system can arrange maintenance and repairs in advance to avoid downtime and damage caused by sudden equipment failures, extend the service life of equipment, reduce the maintenance costs of sudden failures and the economic losses caused by downtime, and improve the economic efficiency of power grid operation.

[0038] The multi-source data fusion and big data analysis module uses the extended Kalman filter algorithm to perform state fusion of multi-source data, and its state update equation is: , in, is the state transfer matrix, is the control input matrix, is the observation matrix, is the Kalman gain.

[0039] The extended Kalman filter can obtain the best estimate from multiple data sources by combining the predicted state and actual observations of the system, which helps the system integrate heterogeneous data from different sources and eliminate errors caused by noise and inaccuracy; the state update equation of the extended Kalman filter can update the state vector of the system in real time, and after obtaining new observation data, the system can quickly correct and optimize the current state estimate; through the calculation of the Kalman gain, the system can effectively reduce the impact of observation noise and measurement error on the state estimate. The Kalman gain dynamically adjusts the weights of the predicted value and the observed value, making the fused state estimate more accurate and reducing uncertainty; The extended Kalman filter can fuse multi-source heterogeneous data, allowing the system to use data from different sources to improve the perception and understanding of the power grid's operating status, helping the power grid to achieve more accurate state estimation and fault detection under complex conditions; real-time performance enables the system to quickly adapt to changes in the power grid's operating status and improve the grid's operating efficiency and response speed; dynamically adjust the weights of predicted values ​​and observed values ​​to reduce the impact of noise, sensor failure or inaccurate data, and improve the accuracy of the fused state estimation, making the system more reliable; it can effectively deal with the uncertainty and complexity of power grid operation, allowing the system to maintain efficient operation management and optimize decision-making capabilities in a dynamic environment, especially under load fluctuations and new energy fluctuations.

[0040] A method for evaluating power grid support performance based on digital twins includes the following steps: S1, collect real-time status data of power grid equipment and external environment data; S2. Based on the collected data, build digital twin models at the device layer, local power grid layer and whole network layer; S3, using the Monte Carlo method to generate a variety of complex scenarios for simulation and performance evaluation; S4. Optimize power dispatch strategy based on deep reinforcement learning algorithm; S5. Use the extended Kalman filter algorithm to fuse multi-source data and conduct real-time evaluation of the power grid status; S6. Provide optimized grid operation solutions and display real-time operation status and dispatching decisions through a visualization platform.

[0041] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A power grid support performance evaluation system based on digital twins, characterized in that: It includes data acquisition and processing module, digital twin model construction and integration module, multi-scenario simulation and prediction evaluation module, intelligent scheduling and fault prediction module, multi-source data fusion and big data analysis module, performance evaluation and optimization decision module, and visualization and human-computer interaction module: The data acquisition and processing module is used to collect the real-time status of the equipment in the physical power grid, load data and external environment data, and clean and standardize the collected data; The digital twin model construction and integration module includes: Micro-level models are used to build twin models of key grid equipment; Meso-level models for twin simulation of local grid power flow and voltage stability; The macro-level model is used for twin simulation of the load balance and power generation scheduling of the entire network. The optimization model is calculated through the following power balance equation: , in, is the power output of generator i, is the power demand of load node j; The multi-scenario simulation and prediction evaluation module is used to generate a simulation situation of power grid operation under complex scenarios and evaluate power grid performance based on the simulation results; The intelligent dispatching and fault prediction module is used to optimize power dispatching through a deep reinforcement learning algorithm and to perform fault prediction through a long short-term memory network; The multi-source data fusion and big data analysis module fuses multi-source data through extended Kalman filtering; The performance evaluation and optimization decision module is used to make optimization decisions based on voltage stability, power loss and operating cost; The visualization and human-computer interaction module is used to display the real-time operation status and dispatching decisions of the power grid, and supports voice interaction through natural language processing technology.

2. A power grid support performance evaluation system based on digital twin according to claim 1, characterized in that: The data acquisition and processing module includes: Sensor network unit, used to collect the operating status of power grid equipment in real time; Edge computing node units, used for local processing and preprocessing of collected data; The cloud data synchronization unit is used to upload the pre-processed data to the cloud for centralized analysis.

3. The power grid support performance evaluation system based on digital twin according to claim 1, characterized in that: The digital twin model constructs a micro-layer model of the integrated module to simulate the operating state of the equipment through the following state equation: ， in, is the state vector of device i at time t, , are the state matrix and control input matrix, For external interference.

4. The power grid support performance evaluation system based on digital twin according to claim 1, characterized in that: The meso-level model describes the voltage and power flow of the local power grid through the following power flow equation: , in, and is the active power and reactive power of node i, is the node voltage, is the admittance between nodes.

5. The power grid support performance evaluation system based on digital twin according to claim 1, characterized in that: The macro-level model determines the power output of the entire grid generator through the following optimization scheduling formula: , in, is the system operating cost, is the power output of generator i, , , is the cost coefficient of the generator.

6. The power grid support performance evaluation system based on digital twin according to claim 1, characterized in that: The multi-scenario simulation and prediction evaluation module uses the Monte Carlo method to generate a scenario set. Defined as: ， in, is the load demand, , is the power of wind and photovoltaic power generation, For meteorological conditions, is the number of simulation scenarios.

7. The power grid support performance evaluation system based on digital twin according to claim 1, characterized in that: The intelligent scheduling module performs scheduling optimization based on the deep reinforcement learning algorithm, and its objective function is: , in, To accumulate rewards, is the immediate reward at time t, is the discount factor.

8. The power grid support performance evaluation system based on digital twin according to claim 1, characterized in that: The fault prediction module performs fault prediction based on the long short-term memory network model. The hidden state update equation of the long short-term memory network model is: ， in, is the hidden state, For input data, , is the weight matrix, For bias.

9. The power grid support performance evaluation system based on digital twin according to claim 1, characterized in that: The multi-source data fusion and big data analysis module performs state fusion of multi-source data through the extended Kalman filter algorithm, and its state update equation is: ， in, is the state transfer matrix, is the control input matrix, is the observation matrix, is the Kalman gain.

10. A method for evaluating the performance of a power grid support based on digital twins, according to a system for evaluating the performance of a power grid support based on digital twins according to any one of claims 1 to 9, characterized in that: The method comprises the following steps: S1, collecting real-time status data of power grid equipment and external environment data; S2. Based on the collected data, build digital twin models at the device layer, local power grid layer and whole network layer; S3, using the Monte Carlo method to generate a variety of complex scenarios for simulation and performance evaluation; S4. Optimize power dispatch strategy based on deep reinforcement learning algorithm; S5. Use the extended Kalman filter algorithm to fuse multi-source data and conduct real-time evaluation of the power grid status; S6. Provide optimized grid operation solutions and display real-time operation status and dispatching decisions through a visualization platform.

Images