Power dispatching optimization system based on large language model
Through the power scheduling optimization system based on the large language model, dynamic analysis and adjustment of power output, the problems of slow response and insufficient monitoring of traditional systems under high loads are solved, the flexibility and stability of the power grid are achieved, and energy waste and operation are reduced. cost.
Patent Information
- Application Number
- CN202411523702.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-30
- Publication Date
- 2024-11-29
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional power scheduling systems are difficult to achieve efficient data analysis and immediate decision-making support under large-scale data and high load conditions, resulting in slow response of power grid scheduling, unable to adapt to demand peaks in time or respond to emergencies, and relying on outdated monitoring methods cannot meet modern times. The demand for power systems leads to insufficient power supply or waste of electricity, increasing operating costs and grid pressure.
Adopt the power scheduling optimization system based on the large language model, including the load trend analysis module, the power distribution optimization module, the emergency response scheduling module and the grid stability monitoring module. Through self-supervised learning and BERT model, the power output is dynamically adjusted. , optimize the supply and demand relationship, realize real-time response and grid stability monitoring.
It improves the flexibility of power grid scheduling and resource utilization, ensures the stable operation of the system when power demand changes dramatically, reduces energy waste, and quickly activates the emergency response mechanism to ensure power supply continuity and overall safety.
Smart Images

Figure CN119048283A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of electric power dispatching, and in particular to an electric power dispatching optimization system based on a large language model. Background Art
[0002] The field of power dispatching technology includes predicting power demand, optimizing the use of power generation and transmission facilities, and ensuring the stability and economy of the system. The application of technology covers everything from power station management to overall grid dispatching. With the increase of renewable energy and the intelligence of the grid, power dispatching technology is also constantly introducing advanced algorithms and data analysis methods to improve the accuracy and efficiency of dispatching.
[0003] Among them, the main purpose of the power dispatch optimization system of the large language model is to improve the efficiency and response speed of power dispatch through automated and intelligent decision support. By processing a large amount of historical data and real-time data, the system can predict changes in power demand, automatically adjust power output, optimize energy distribution, and reduce operating costs, providing a more efficient management and operation model for the power market.
[0004] Traditional systems have difficulty in achieving efficient data analysis and instant decision support when faced with large amounts of data and high complexity, resulting in slow response in grid dispatching and inability to adapt to demand peaks or respond to sudden grid events in a timely manner. The lack of effective real-time analysis and response strategies makes the system prone to power shortages or power waste under high load conditions, increasing operating costs and grid pressure. In addition, for the maintenance of grid stability, traditional technologies rely on outdated monitoring methods, which cannot meet the needs of modern power systems in terms of prediction accuracy and processing speed, especially when processing large-scale data and implementing refined management, and often face problems of low efficiency and accuracy. Summary of the invention
[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose a power dispatch optimization system based on a large language model.
[0006] In order to achieve the above object, the present invention adopts the following technical solution: a power dispatch optimization system based on a large language model, the system comprising: The load trend analysis module is based on the current load data extracted from multi-source power grid data. Through self-supervised learning, it compares the load characteristics of real-time and historical data, and adjusts the prediction parameters according to the load change trend to obtain load change prediction information; The power distribution optimization module uses the BERT model to analyze the supply and demand relationship between multiple power supply stations and multiple nodes based on the load change prediction information, reallocates the output power, adjusts the power output to the optimal state, and simulates and tests the power grid response information to obtain power supply and demand adjustment information; The emergency response dispatch module collects external signals and internal power grid abnormality monitoring data, evaluates the urgency and impact range of the event, analyzes the power shortage area, performs power redistribution according to the power supply and demand adjustment information, matches the demand peak, and obtains the emergency dispatch configuration; The grid stability monitoring module continuously detects the operating parameters of the power system, analyzes grid outage events and power fluctuations, optimizes the power dispatch process according to the emergency dispatch configuration, evaluates and adjusts the grid operation status, and obtains stability monitoring records.
[0007] The present invention is improved in that the step of comparing load characteristics is specifically as follows: Current load data extracted based on multi-source power grid data, including load data at the current time and corresponding historical time periods, including current, voltage and power parameters, to obtain real-time and historical load feature sets; The real-time and historical load feature sets are compared and analyzed using the formula: Get the load change trend, where It represents the average absolute deviation between real-time load data and historical load data. Representative Real-time measurement of load power, Representative The historical measured load power, is the adjustment coefficient, which is used to adjust the scale difference between real-time data and historical data. is the total number of measurements.
[0008] The present invention is improved in that the step of acquiring the load change prediction information is specifically as follows: According to the load change trend, extract key change features, including growth rate, consumption peak and periodic fluctuation, to obtain a key trend analysis data set; Based on the key trend analysis data set, the prediction parameters are adjusted through self-supervised learning, using the formula: Calculate new parameters , generating load change prediction information, where is the current prediction parameter, Indicates The change in the parameter, is the adjustment coefficient for each parameter, is the standard deviation of the parameter variation, is the learning rate, is the number of parameters.
[0009] The present invention is improved in that the analysis steps of the supply and demand relationship are specifically as follows: Based on the load change prediction information and in combination with real-time monitoring data, a supply and demand data set is created, the data set including time series information and load parameters of multiple nodes; The BERT model is used to analyze the dynamic information in the supply and demand dataset, using the formula: Evaluate the supply and demand balance between power supply stations and nodes to obtain the supply and demand relationship index ,in, is the first data points, including real-time data of power supply stations or nodes, is the weight of each data point, is the total number of data points.
[0010] The present invention is improved in that the step of obtaining the power supply and demand adjustment information is specifically as follows: Based on the supply and demand relationship, identify key areas with excess or insufficient electricity, and obtain supply and demand imbalance information, including key areas and corresponding power conditions; Based on the supply and demand imbalance information, the adjusted power output is calculated using the formula: Get the new power allocation configuration ,in, Represents the current power output, To determine the adjustment needs, is the response adjustment factor, is the sensitivity coefficient; The new power allocation configuration is simulated and tested to evaluate the impact of power adjustment on grid stability, verify the effectiveness of the adjustment, and obtain power supply and demand adjustment information.
[0011] The present invention is improved in that the step of analyzing the power shortage area is specifically as follows: Collect external signals and internal power grid abnormality monitoring data, integrate the data, and establish preliminary event monitoring records; The data in the preliminary event monitoring records are analyzed using the formula: Calculate the urgency and impact of the incident and generate power shortage area analysis results ,in, Represents external signal data, Represents internal power grid monitoring data, is the area of the affected region, is the number of monitoring points.
[0012] The present invention is improved in that the step of obtaining the emergency dispatch configuration is specifically as follows: Based on the power supply and demand adjustment information and the power shortage area, calculate the current power consumption and expected demand of multiple areas, and generate demand peak analysis data; According to the peak demand analysis data, the formula is used: Calculate the total amount of electricity required to be dispatched , get the emergency dispatch configuration, where Representative area The peak electricity demand Indicates area The current electricity supply, is the grid efficiency factor, is the total number of regions.
[0013] The present invention is improved in that the steps of obtaining the stability monitoring record are specifically as follows: Continuously monitor the operating parameters of the power system, including voltage, current, frequency and temperature, collect data and establish a real-time power operation log; Based on the real-time power operation log, analyzing power grid outage events and power fluctuations, identifying abnormal patterns, and generating power grid event and fluctuation analysis results; According to the power grid event and fluctuation analysis results and emergency dispatch configuration, the formula is used: Conduct power dispatch adjustment and evaluation to obtain stability monitoring records, including: represents the adjusted power output, represents the initial power, and Represent expected and current demand, respectively. is the time variable, is the response time threshold, is the response amplitude adjustment coefficient, It is the sensitivity coefficient that affects the response to changes in demand.
[0014] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, through self-supervised learning, in terms of real-time data analysis, the system is allowed to extract useful information from a large amount of unlabeled data, thereby enhancing the adaptability and accuracy of the prediction. By analyzing the supply and demand relationship through the BERT model, the system can dynamically optimize power distribution, improve the dispatching flexibility and resource utilization of the power grid, and ensure the stable operation of the system when the power demand changes sharply. At the same time, it reduces energy waste, and the rapid activation and configuration of the emergency response mechanism, coupled with the continuous optimization of power grid monitoring, ensures the continuity of power supply and overall safety. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] Figure 1 is a system flow chart of the present invention; Figure 2 is a flow chart of comparing load characteristics in the present invention; Figure 3 A flowchart for obtaining load change prediction information in the present invention; Figure 4 It is a flow chart for analyzing the supply and demand relationship in the present invention; Figure 5 A flowchart for obtaining power supply and demand adjustment information in the present invention; Figure 6 A flowchart for analyzing power shortage areas in the present invention; Figure 7 A flowchart for obtaining the emergency dispatch configuration in the present invention; Figure 8 The figure is a flow chart for obtaining stability monitoring records in the present invention. DETAILED DESCRIPTION
[0016] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.
[0017] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating positions or positional relationships, are based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined.
[0018] Example See also Figure 1 The present invention provides a technical solution: a power dispatch optimization system based on a large language model comprises: The load trend analysis module is based on the current load data extracted from multi-source power grid data. Through self-supervised learning, it compares the load characteristics of real-time and historical data, and adjusts the prediction parameters according to the load change trend to obtain load change prediction information; The power distribution optimization module uses the BERT model to analyze the supply and demand relationship between multiple power supply stations and multiple nodes based on load change prediction information, redistribute output power, adjust power output to the optimal state, and simulate and test grid response information to obtain power supply and demand adjustment information; The emergency response dispatch module collects external signals and internal grid anomaly monitoring data, evaluates the urgency and impact of the event, analyzes the power shortage area, performs power redistribution according to the power supply and demand adjustment information, matches the demand peak, and obtains the emergency dispatch configuration; The grid stability monitoring module continuously detects the operating parameters of the power system, analyzes grid outage events and power fluctuations, optimizes the power dispatch process according to the emergency dispatch configuration, evaluates and adjusts the grid operation status, and obtains stability monitoring records.
[0019] Load change prediction information includes load increase / decrease rate, prediction accuracy, and trend stability information. Power supply and demand adjustment information includes new power configuration, energy-saving optimization indicators, and emergency response capabilities. Emergency dispatch configuration includes emergency level, deployment efficiency, and resource call speed. Stability monitoring records include voltage stability index, frequency fluctuation range, and fault response time.
[0020] See also Figure 2 , the steps for comparing load characteristics are as follows: Current load data extracted based on multi-source power grid data, including load data at the current time and corresponding historical time periods, including current, voltage and power parameters, to obtain real-time and historical load feature sets; Current, voltage and power data are extracted from multi-source power grid systems and integrated into real-time and historical load feature sets. Key parameters are accurately extracted from a large number of sensors and recorders through data integration technology to ensure data integrity and reliability. When integrating data from multiple data sources, the time synchronization and data integrity of the data are verified to ensure accurate comparison of real-time and historical data sets, and load feature sets including current, voltage and power are obtained. The feature sets will provide a basis for subsequent data analysis.
[0021] Comparative analysis of real-time and historical load feature sets is performed using the formula: Get the load change trend, where It represents the average absolute deviation between real-time load data and historical load data. Representative Real-time measurement of load power, Representative The historical measured load power, is the adjustment coefficient, which is used to adjust the scale difference between real-time data and historical data. is the total number of measurements; The following data was collected: , adjustment coefficient , real-time power The data points are [50, 55, 60], and the historical power The data points are [48, 53, 58]. Substitute the data into the formula to calculate: Calculate the difference for each data point and apply an adjustment factor: For the first data point: For the second data point: For the third data point: Take the absolute value of each difference: Absolute difference of the first data point: 4.5 Absolute difference of the second data point: 4.75 Absolute difference of the 3rd data point: 5 Compute the average of the differences across all data points: Mean absolute difference: It is 4.75, indicating that in the selected data points, the average absolute deviation between the real-time measured load power and the historically measured load power is 4.75, reflecting the deviation between the real-time data and the historical data under the action of the adjustment coefficient under different conditions, and providing a quantitative basis for further analysis of power grid load changes.
[0022] See also Figure 3 ,The specific steps for obtaining load change prediction information are: According to the load change trend, key change features are extracted, including growth rate, consumption peak and periodic fluctuation, to obtain a key trend analysis data set; Extract the main change characteristics from the load change trend report, such as growth rate, consumption peak, and periodic fluctuations, synchronize and verify the integrity and accuracy of data from multiple sensors and historical records, and ensure that the analyzed data can reflect the actual operating status of the power grid, thereby providing accurate actual input for adjusting the parameters of the prediction model. The generated key trend analysis data set provides an in-depth analysis of the power grid load characteristics.
[0023] Based on the key trend analysis data set, the prediction parameters are adjusted through self-supervised learning, using the formula: Calculate new parameters , generating load change prediction information, where is the current prediction parameter, Indicates The change in the parameter, is the adjustment coefficient for each parameter, is the standard deviation of the parameter variation, is the learning rate, which controls the step size of the overall parameter update. is the number of parameters; There are the following data, , ; , ; , ; Calculate the weighted sum of parameter changes: Calculate new parameter values: The results show that based on the current load change trend, the new parameters The adjustment to 1.4375 reflects the trend of fine-tuning existing forecast parameters to better adapt to load changes.
[0024] See also Figure 4 , the analysis steps of supply and demand relationship are as follows: Based on the load change prediction information and combined with real-time monitoring data, a supply and demand data set is created. The data set includes time series information and load parameters of multiple nodes. Integrate load change forecast information and real-time monitoring data from multiple sources to create a supply and demand dataset. The dataset will include detailed time series and load parameters for each power supply station and node. The process involves data extraction, cleaning and synchronization to ensure the integrity and accuracy of the data. By using data processing technology, the data is formatted to suit the needs of the deep learning model, thereby providing fully prepared input data for the application of the BERT model.
[0025] The BERT model is used to analyze the dynamic information in the supply and demand dataset using the formula: Evaluate the supply and demand balance between power supply stations and nodes, where nodes refer to consumption nodes (residential parks, etc.) and distribution nodes (substations), and obtain the supply and demand relationship index ,in, is the first data points, including real-time data of power supply stations or nodes, is the weight of each data point, assigned according to its importance in the overall analysis, is the total number of data points; There are 3 data points with weights , the BERT model is used for each data point The output supply and demand relationship scores are 0.8, 0.6 and 0.9 respectively, then: The result shows that the weighted supply-demand relationship index is 0.72, which reflects the current supply and demand status and will be used to generate a supply-demand relationship report to provide a basis for further management and optimization decisions of the power grid.
[0026] See also Figure 5 ,The specific steps for obtaining power supply and demand adjustment information are as follows: Based on the supply and demand relationship, identify key areas with excess or insufficient electricity and obtain information on supply and demand imbalance, including key areas and corresponding power conditions; Analyze supply and demand data and identify key areas of power surplus or shortage so that they can be used directly in the subsequent decision-making process of power redistribution, ensuring the accuracy and real-time nature of the basic data of the power adjustment strategy. The identified key areas directly affect the formulation and implementation of the adjustment strategy, ensuring that the entire adjustment process can accurately respond to the actual supply and demand situation, thereby improving the operating efficiency and response speed of the power grid.
[0027] Based on the supply and demand imbalance information, the adjusted power output is calculated using the formula: Get the new power allocation configuration ,in, Represents the current power output, To determine the adjustment needs, is the response adjustment factor, is the sensitivity coefficient; In the initial case, the current power of a grid node 500MW, based on adjustment requirements determined by supply and demand analysis For 50MW, the response adjustment factor The sensitivity is 0.1. is 0.3, according to the formula: The results show that the adjusted power will increase from 500MW to 551.5MW, providing enough electricity to meet the increased demand while maintaining system stability.
[0028] Conduct simulation tests on new power allocation configurations, evaluate the impact of power adjustment on grid stability, verify the effectiveness of the adjustment, and obtain information on power supply and demand adjustment; In the simulation test of the new power distribution, relevant grid operation data are collected, including voltage, current, frequency, load and other indicators. Through real-time data collection, data cleaning technology is used to remove noise and outliers to ensure data accuracy and availability. Simulation tests are carried out based on precise data to simulate different power supply and load conditions. The specific impact of power adjustment on the grid, including voltage stability, frequency stability and load balance, is evaluated. The simulation test calculates the grid response and predicts grid imbalance, thereby evaluating the overall stability of the grid.
[0029] See also Figure 6 , the steps to analyze the power shortage area are as follows: Collect external signals and internal power grid abnormality monitoring data, integrate the data, and establish preliminary event monitoring records; Collecting external environmental signals and abnormal monitoring data of the internal power grid involves using sensor networks for real-time data capture. Data sources include load monitoring of power grid nodes and safety system alarms. Integrating data involves formatting information from different data sources to ensure consistency, such as standardizing timestamps and synchronizing data collection frequencies. Handling data loss and outliers is also part of the integration to ensure the accuracy of subsequent analysis and establish preliminary event monitoring records.
[0030] The data from the preliminary event monitoring records were analyzed using the formula: Calculate the urgency and impact of the incident and generate power shortage area analysis results ,in, Represents external signal data, Represents internal power grid monitoring data, is the area of the affected region, is the number of monitoring points; The following data were collected: external signal data =100, internal power grid monitoring data =150, the area of the affected area For an area of 2000 square meters, the number of monitoring points is 40, calculated as follows: The results show that the impact index of the incident is 12500, which reflects the severity of the emergency and the amount of resources required for further emergency response and resource allocation decisions.
[0031] See also Figure 7, the specific steps for obtaining emergency dispatch configuration are: Based on the power supply and demand adjustment information and power shortage areas, calculate the current power consumption and expected demand in multiple regions, and generate demand peak analysis data; Based on the information on power supply and demand adjustment, regional demand analysis is conducted to collect current power consumption data in each region, including power usage in industrial, commercial and residential areas. The data usually comes from the real-time power monitoring system and historical power consumption records provided by the regional power grid company. Then, the peak power demand in each region in the next period of time is predicted. The prediction process involves time series analysis and trend prediction models, which can accurately predict power shortages and formulate targeted power adjustment strategies accordingly. Finally, all analysis results are integrated to form a comprehensive data report containing current power consumption and expected demand in multiple regions. This report will be used to guide actual power dispatch decisions and ensure the stability and reliability of power supply in each region. Analyze the data based on peak demand using the formula: Calculate the total amount of electricity required to be dispatched , get the emergency dispatch configuration, where Representative area The peak electricity demand Indicates area The current electricity supply, The grid efficiency factor includes the allocation of necessary resources from areas with surplus electricity to fill the gap. is the total number of regions; There are three areas where electricity demand They are 1000MW, 800MW and 1200MW respectively, and the current supply 900MW, 800MW and 1100MW respectively, the grid efficiency factor is 0.95, calculated as follows: The results show that a total of 210.53MW of electricity needs to be dispatched from areas with surplus electricity to meet the power gap in these three areas and ensure the continuity and stability of power supply.
[0032] See also Figure 8 , the specific steps for obtaining stability monitoring records are: Continuously monitor the operating parameters of the power system, including voltage, current, frequency and temperature, collect data and establish a real-time power operation log; Monitor the operating parameters of the power system, including continuous monitoring of voltage, current, frequency and temperature, to ensure multi-dimensional data collection. The operating data of the power grid is collected in real time and automatically transmitted to the data processing center. The collected data is initially screened and processed to screen out abnormal data for further analysis. Data processing includes denoising, standardization and time series analysis to ensure the quality and reliability of the data. The processed data is recorded in the real-time power operation log.
[0033] Analyze grid outage events and power fluctuations based on real-time power operation logs, identify abnormal patterns, and generate grid event and fluctuation analysis results; Based on real-time power operation logs, data analysis and machine learning technologies are used to conduct in-depth analysis of power grid outages and power fluctuations, and historical data models are used to predict and detect anomalies in the power grid. The analysis process includes multi-dimensional data comparison and pattern recognition. The collected data is processed according to the preset abnormal pattern recognition algorithm, and the abnormal patterns are identified and classified. The results show the abnormal frequency, duration and potential causes. The analysis results are integrated into power grid event and fluctuation analysis results.
[0034] According to the results of power grid event and fluctuation analysis and emergency dispatch configuration, the formula is adopted: Conduct power dispatch adjustment and evaluation to obtain stability monitoring records, including: represents the adjusted power output, represents the initial power, and Represent expected and current demand, respectively. is the time variable, is the response time threshold, is the response amplitude adjustment coefficient, is the sensitivity coefficient affecting the response to demand changes; The following data are available: initial power 500MW, expected demand The current demand is 600MW. 570MW, time variable The response time threshold is 5 hours. For 3 hours, adjust the coefficient is 0.05, the response rate is 0.1, substitute the data and calculate: The result shows that the adjusted power is 501MW. After dynamic adjustment, it can better adapt to actual demand changes and enhance the stability and efficiency of the power grid.
[0035] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.
Claims
1. A power dispatch optimization system based on a large language model, characterized in that: The system comprises: The load trend analysis module is based on the current load data extracted from multi-source power grid data. Through self-supervised learning, it compares the load characteristics of real-time and historical data, and adjusts the prediction parameters according to the load change trend to obtain load change prediction information; The power distribution optimization module uses the BERT model to analyze the supply and demand relationship between multiple power supply stations and multiple nodes based on the load change prediction information, reallocates the output power, adjusts the power output to the optimal state, and simulates and tests the power grid response information to obtain power supply and demand adjustment information; The emergency response dispatch module collects external signals and internal power grid abnormality monitoring data, evaluates the urgency and impact range of the event, analyzes the power shortage area, performs power redistribution according to the power supply and demand adjustment information, matches the demand peak, and obtains the emergency dispatch configuration; The grid stability monitoring module continuously detects the operating parameters of the power system, analyzes grid outage events and power fluctuations, optimizes the power dispatch process according to the emergency dispatch configuration, evaluates and adjusts the grid operation status, and obtains stability monitoring records.
2. The power dispatch optimization system based on a large language model according to claim 1 is characterized in that: The step of comparing load characteristics is specifically as follows: Current load data extracted based on multi-source power grid data, including load data at the current time and corresponding historical time periods, including current, voltage and power parameters, to obtain real-time and historical load feature sets; The real-time and historical load feature sets are compared and analyzed using the formula: Get the load change trend, where It represents the average absolute deviation between real-time load data and historical load data. Representative Real-time measurement of load power, Representative The historical measured load power, is the adjustment coefficient, which is used to adjust the scale difference between real-time data and historical data. is the total number of measurements.
3. The power dispatch optimization system based on large language model according to claim 1 is characterized in that: The steps of acquiring the load change prediction information are specifically as follows: According to the load change trend, key change features are extracted, including growth rate, consumption peak and periodic fluctuation, to obtain a key trend analysis data set; Based on the key trend analysis data set, the prediction parameters are adjusted through self-supervised learning, using the formula: Calculate new parameters , generating load change prediction information, where is the current prediction parameter, Indicates The change in the parameter, is the adjustment coefficient for each parameter, is the standard deviation of the parameter variation, is the learning rate, is the number of parameters.
4. The power dispatch optimization system based on large language model according to claim 1 is characterized in that: The analysis steps of the supply and demand relationship are specifically as follows: Based on the load change prediction information and in combination with real-time monitoring data, a supply and demand data set is created, the data set including time series information and load parameters of multiple nodes; The BERT model is used to analyze the dynamic information in the supply and demand dataset, using the formula: Evaluate the supply and demand balance between the power supply station and the node, and obtain the supply and demand relationship index ,in, is the first data points, including real-time data of power supply stations or nodes, is the weight of each data point, is the total number of data points.
5. The power dispatch optimization system based on large language model according to claim 1 is characterized in that: The steps for obtaining the power supply and demand adjustment information are specifically as follows: Based on the supply and demand relationship, identify key areas with excess or insufficient electricity, and obtain supply and demand imbalance information, including key areas and corresponding power conditions; Based on the supply and demand imbalance information, the adjusted power output is calculated using the formula: Get the new power allocation configuration ,in, Represents the current power output, To determine the adjustment needs, is the response adjustment factor, is the sensitivity coefficient; The new power allocation configuration is simulated and tested to evaluate the impact of power adjustment on grid stability, verify the effectiveness of the adjustment, and obtain power supply and demand adjustment information.
6. The power dispatch optimization system based on large language model according to claim 1, characterized in that: The steps of analyzing the power shortage area are specifically as follows: Collect external signals and internal power grid abnormality monitoring data, integrate the data, and establish preliminary event monitoring records; The data in the preliminary event monitoring records are analyzed using the formula: Calculate the urgency and impact of the incident and generate power shortage area analysis results ,in, Represents external signal data, Represents internal power grid monitoring data, is the area of the affected region, is the number of monitoring points.
7. The power dispatch optimization system based on large language model according to claim 1 is characterized in that: The steps for obtaining the emergency dispatch configuration are specifically as follows: Based on the power supply and demand adjustment information and the power shortage area, calculate the current power consumption and expected demand of multiple areas, and generate demand peak analysis data; According to the peak demand analysis data, the formula is used: Calculate the total amount of electricity required to be dispatched , get the emergency dispatch configuration, where Representative area The peak electricity demand Indicates area The current electricity supply, is the grid efficiency factor, is the total number of regions.
8. The power dispatch optimization system based on large language model according to claim 1, characterized in that: The steps for obtaining the stability monitoring record are specifically as follows: Continuously monitor the operating parameters of the power system, including voltage, current, frequency and temperature, collect data and establish a real-time power operation log; Based on the real-time power operation log, analyzing power grid outage events and power fluctuations, identifying abnormal patterns, and generating power grid event and fluctuation analysis results; According to the power grid event and fluctuation analysis results and emergency dispatch configuration, the formula is used: Conduct power dispatch adjustment and evaluation to obtain stability monitoring records, including: represents the adjusted power output, represents the initial power, and Represent expected and current demand, respectively. is the time variable, is the response time threshold, is the response amplitude adjustment coefficient, It is the sensitivity coefficient that affects the response to changes in demand.
Citation Information
Cited By
Intelligent power monitoring management system
CN119582424A
Intelligent control system for extracting traditional Chinese medicinal materials
CN119620814A
Flexible power supply distribution method and system based on load prediction
CN119921405A
Large model application system for energy power planning and intelligent optimization
CN120235310A
A large-scale model application system for energy and power planning and intelligent optimization
CN120235310B
