Intelligent dynamic power optimization regulation and control method and system

Abstract

The invention particularly relates to an intelligent dynamic power optimization regulation and control method and system. According to the intelligent dynamic electric power optimization regulation and control method, after data acquisition and processing, mathematical models are established for various elements in an electric power system so as to describe electrical characteristics and operation behaviors of the elements; adopting a time sequence analysis and machine learning algorithm to predict power load and renewable energy output; according to the method, a target of optimization regulation and control is determined in a user-defined mode, various constraint conditions of the power system are considered, an optimal regulation and control strategy is solved, real-time regulation and control of the power system are achieved, and model parameters are adjusted according to real-time feedback. According to the intelligent dynamic power optimization regulation and control method, high-precision load prediction is realized, and new energy grid connection can be dynamically and stably controlled; meanwhile, full-link delay is accurately compensated to + / -5 milliseconds, intelligent regulation and control can be achieved to adapt to system changes, the multi-main-body cooperation efficiency is improved, the anti-delay robustness is high, and source network load storage cooperation optimization operation can be supported.

Description

Technical Field

[0001] This invention relates to the field of power regulation technology, and in particular to an intelligent dynamic power optimization regulation method and system. Background Technology

[0002] As a key area for carbon emissions, the power industry is actively promoting the transformation to a new power system based on wind, solar, and energy storage, with photovoltaic and wind power demonstrating strong momentum in renewable energy development. However, the traditional grid's "source-following-load" control model is proving inadequate in dealing with the intermittent nature of distributed energy resources. With the exponential growth of data from renewable energy power plants, traditional communication networks are revealing problems such as insufficient bandwidth and path congestion, resulting in dispatch response delays of up to seconds, far from meeting the grid's stringent real-time requirements. At the same time, the power industry also faces increasingly severe pressure to reduce emissions.

[0003] On the one hand, the volatility problem of renewable energy integration is difficult to solve.

[0004] When the actual output of wind and solar power fluctuates by more than 20% of the predicted value, AI scheduling models trained on historical data struggle to issue timely warnings of voltage anomalies due to data gaps and insufficient robustness under extreme conditions, which can easily trigger cascading failures. Although the application of generative AI has improved load forecasting accuracy to ±2.3%, its complex calculation process still results in a 15-minute prediction lag when encountering "non-stable load" situations caused by extreme weather, making it difficult to meet the needs of rapid response in the power system.

[0005] Secondly, there is a real-time bottleneck in communication and control.

[0006] In traditional power communication networks, there is a significant "last mile" blind spot in distribution networks below 10kV. Currently widely used 4G networks have an uplink bandwidth of only 1-5Mbps, which cannot meet the high bandwidth requirements of services such as high-definition video surveillance and real-time control. Even with advanced 5G RedCap technology, distribution network differential protection still faces latency challenges in the 10ms range. In virtual power plant scenarios, the coordinated control of distributed resources places even higher demands on communication response, requiring millisecond-level latency for efficient operation.

[0007] Third, the market and user needs are facing upgrades and changes.

[0008] A virtual power plant successfully improved its load response speed by 60% by introducing the DeepSeek AI system. However, in the process of aggregating dispersed resources such as residential V2G charging stations and building air conditioners, the problem of asynchronous dispatch commands and equipment responses still exists. For example, although a campus air conditioner group control system can complete load adjustment within 50 seconds, communication delays cause 20% of the devices to have actual response errors. Furthermore, a case study of real-time green certificate trading for residential photovoltaic systems using blockchain technology demonstrates that the power system's regulation model is gradually evolving from a simple "physical energy flow" to a collaborative optimization of "energy-value flow." In this new market environment, traditional static regulation methods are no longer adequate to meet the demands of dynamic pricing in market-based transactions.

[0009] Fourth, there are still shortcomings in the integrated application of mathematical optimization and intelligent algorithms.

[0010] The combination of Model Predictive Control (MPC) and reinforcement learning: Existing technologies propose a multi-energy collaborative control method that utilizes a physical simulation model to construct a safety constraint framework and combines it with a reinforcement learning agent to achieve dynamic scheduling of the power system. In practical operation, this technology can effectively coordinate and manage energy sources such as wind power, energy storage, and loads under normal operating conditions. However, when renewable energy output experiences drastic random fluctuations, such as significant changes in wind power output within a short period or a sudden drop in photovoltaic power generation due to weather changes, the limitations of this technology become apparent. Due to the high uncertainty of renewable energy output, in scenarios with multiple conflicting objectives, even with the introduction of a reinforcement learning agent, prior constraints are still required to avoid policy oscillations, making it impossible to achieve fully adaptive control of random fluctuations in renewable energy output. For example, in a practical application case, when wind power fluctuated by more than 30% of the predicted value within 10 minutes, the system was unable to quickly adjust its scheduling strategy, leading to temporary voltage instability in the regional power grid and affecting the reliability and stability of power supply.

[0011] The Challenges of Deep Learning in Load Forecasting: LSTM-based deep learning load forecasting models perform well under normal operating conditions, keeping prediction errors below 3%, providing a relatively reliable basis for power dispatch decisions. However, in special circumstances such as a surge in air conditioning load caused by extreme temperatures, the prediction error can increase dramatically, potentially exceeding 15%, due to the model's lack of deep generalization ability to the complex nonlinear relationship between weather and load. For example, in a city experiencing high summer temperatures, a sudden extreme heat wave caused a rapid increase in air conditioning load. The LSTM-based forecasting model failed to accurately predict the load changes, leading to delays in power dispatch and ultimately resulting in short-term power supply shortages in some areas.

[0012] Fifth, problems still exist in the engineering practice of communication and network technologies.

[0013] The Application Results and Challenges of SDN and 5G: After introducing SDN technology into a local power grid, significant results were achieved, with communication network bandwidth utilization increasing by 40% and core service response speed improving by 90%. However, this technology encountered a "traffic labeling conflict" problem at the edge nodes of the distribution network, causing 10% of data packet transmission delays to exceed 50ms. Meanwhile, although 5G virtual private networks achieved "zero security incident" operation for distribution network automation services, they have certain limitations in coverage. Fiber optic cables can only extend to the main network above 35kV, while terminals below 10kV still need to rely on wireless communication, posing a risk of signal blind spots.

[0014] Sixth, the bottleneck problem of virtual power plants and distributed control.

[0015] Most existing virtual power plants adopt a "centralized optimization + distributed execution" architecture. For example, in a test at a university, the coordinated control of 988 air conditioning units required real-time command issuance from a cloud server. When the number of devices being controlled simultaneously exceeded 1000, the cloud computing pressure increased, and the scheduling cycle extended to more than 10 seconds, failing to meet the requirement of second-level response. Furthermore, in multi-energy systems, the coordinated control of reactive power support from photovoltaic inverters and the charging and discharging of energy storage has not yet formed a unified safety constraint standard, which increases the risk of voltage exceeding limits by 20% in some scenarios.

[0016] Based on the above problems, this invention proposes an intelligent dynamic power optimization and control method and system. Summary of the Invention

[0017] To overcome the shortcomings of existing technologies, this invention provides a simple and efficient intelligent dynamic power optimization and control method.

[0018] This invention is achieved through the following technical solution:

[0019] A smart dynamic power optimization and control method and system, characterized by the following steps:

[0020] Step S1: Data Acquisition and Processing

[0021] In the power system, key nodes are selected by customization, and smart sensors, including voltage sensors, current sensors, power sensors and frequency sensors, are deployed at the key nodes to collect the operating parameters of the power system in real time, including voltage value, current value, active power, reactive power and frequency.

[0022] After preprocessing the collected data, it is stored in a data warehouse for later use.

[0023] In step S1, the collected data is cleaned, filtered, and interpolated to remove noise and fill in missing values ​​in order to improve data quality.

[0024] Step S2, Power System Modeling

[0025] Component modeling: Mathematical models are established for various components in the power system, including generators, transformers, transmission lines and loads, to describe their electrical characteristics and operating behavior;

[0026] System overall modeling: combining the models of the various components to form a model of the entire power system;

[0027] In step S2, the power system model considers the topology, power flow distribution, voltage stability and frequency stability of the power system, and can accurately reflect the behavior of the power system under different operating conditions.

[0028] Step S3, Power Forecasting

[0029] Load forecasting: Based on historical load data and relevant meteorological and economic data, time series analysis and machine learning algorithms are used to forecast the power load for a customized future period (such as short-term at the hour level, medium-term at the day level, and long-term at the month level or year level).

[0030] Renewable energy output forecasting: For wind and solar power generation, based on their geographical location, meteorological conditions, and equipment operating status information, the trained forecasting model is used to predict the output for a custom future period.

[0031] In step S3, a load forecasting system based on the TCN-Transformer-KAN hybrid model is constructed. The system captures short-term load fluctuation characteristics through dilated convolution, analyzes long-term time series dependencies using a self-attention mechanism, and approximates complex nonlinear mappings with the help of a Kolmogorov-Arnold network.

[0032] When forecasting renewable energy output, forecasting models include wind power forecasting models based on physical processes and photovoltaic power forecasting models based on machine learning.

[0033] In step S3, the power forecast includes short-term forecast, mid-term forecast and long-term forecast, where the short-term forecast is at the hour level, the mid-term forecast is at the day level, and the long-term forecast is at the month level or year level.

[0034] Step S4: Optimize the formulation of control strategies

[0035] Target setting: Based on the operating requirements and actual conditions of the power system, define and determine the optimization and control targets, including minimizing power generation costs, maximizing renewable energy consumption, maintaining system voltage and frequency stability, and improving power supply reliability;

[0036] Constraint determination: Consider various constraints of the power system, including generator output limits, transmission line capacity limits, transformer capacity limits, allowable ranges of system voltage and frequency, and equipment operation safety constraints, to ensure that the control strategy will not lead to equipment overload and system instability in the actual power system;

[0037] Algorithm solution: Using intelligent optimization algorithms or model predictive control methods, the optimal control strategy is solved under the premise of satisfying the constraints;

[0038] In step S4, the optimal control strategy is solved using a genetic algorithm, particle swarm optimization algorithm, or simulated annealing algorithm.

[0039] In step S4, a genetic algorithm is used to search for the optimal solution in the solution space by simulating the selection, crossover and mutation operations in the biological evolution process, so as to determine the control measures, including the output adjustment of each generator, the adjustment of the tap position of the transformer, and the switching of reactive power compensation equipment.

[0040] Step S5: Control Execution and Feedback

[0041] Control command issuance: The optimized control strategy is transformed into specific control commands, which are then issued to the corresponding execution agencies through the power system's automated control equipment to achieve real-time control of the power system;

[0042] Real-time feedback and adjustment: After the control is implemented, the power system's operating status changes are monitored in real time through the data acquisition system. The actual operating data is compared and analyzed with the predicted data and optimization targets. If the actual situation is found to be inconsistent with the expectations, or if new abnormal operating conditions occur, including sudden faults and load changes, the prediction model and control strategy are adjusted in a timely manner to form a closed-loop feedback control to ensure that the power system is always in the optimal operating state.

[0043] In step S5, when issuing control commands, the synchronous phasor measurement and Pade approximation algorithm are integrated to construct an adaptive delay compensator based on GPS timestamps; through time difference analysis of the measurement signal (first timestamp) and the control signal (second timestamp), a frequency domain compensation function is generated to reduce the phase lag error of the secondary frequency control to within ±5ms.

[0044] After the control is implemented, the Kalman filter algorithm is used to correct the parameters of the power grid component model online based on real-time measurement data, supporting the dynamic updating of key parameters;

[0045] Key parameters include generator rotational inertia and transformer turns ratio, breaking through the limitations of traditional fixed parameter modeling.

[0046] An intelligent dynamic power optimization and control system is provided to implement the above method, including:

[0047] The data acquisition and processing module is responsible for customizing and selecting key nodes in the power system, and deploying smart sensors, including voltage sensors, current sensors, power sensors, and frequency sensors, at these key nodes. It collects the operating parameters of the power system in real time, including voltage values, current values, active power, reactive power, and frequency. After preprocessing the collected data, it stores it in a data warehouse for later use.

[0048] The power system modeling module is responsible for establishing mathematical models of various components in the power system, including generators, transformers, transmission lines and loads, to describe their electrical characteristics and operating behavior; and combining the models of each component to form a model of the entire power system.

[0049] The power forecasting module is responsible for forecasting the power load for a defined future period (e.g., short-term at the hourly level, medium-term at the daily level, long-term at the monthly or yearly level) based on historical load data and relevant meteorological and economic data, using time series analysis and machine learning algorithms; and for forecasting the output of wind and solar power generation for a defined future period based on the geographical location, meteorological conditions, and equipment operating status information, using a trained forecasting model.

[0050] The optimization and control strategy formulation module is responsible for defining the optimization and control objectives based on the power system's operational requirements and actual conditions. These objectives include minimizing generation costs, maximizing renewable energy absorption, maintaining system voltage and frequency stability, and improving power supply reliability. Simultaneously, it considers various constraints of the power system, including generator output limits, transmission line capacity limits, transformer capacity limits, allowable system voltage and frequency ranges, and equipment operational safety constraints, to ensure that the control strategy does not lead to equipment overload or system instability in the actual power system. Using intelligent optimization algorithms or model predictive control methods, it solves for the optimal control strategy while satisfying the constraints.

[0051] The control execution and feedback module is responsible for translating the optimized control strategy into specific control instructions, which are then issued to the corresponding actuators through the power system's automated control equipment to achieve real-time control of the power system. After the control is executed, the module monitors the changes in the power system's operating status in real time through the data acquisition system, comparing and analyzing the actual operating data with the predicted data and optimization targets. If the actual situation does not match the expectations, or if new abnormal operating conditions occur, including sudden faults and load changes, the module promptly adjusts the prediction model and optimizes the control strategy to form a closed-loop feedback control, ensuring that the power system is always in the optimal operating state.

[0052] A smart dynamic power optimization and control device, characterized in that it includes a memory and a processor; the memory is used to store a computer program, and the processor is used to execute the computer program to implement the above-mentioned method steps.

[0053] A readable storage medium, characterized in that: a computer program is stored on the readable storage medium, and the computer program, when executed by a processor, implements the above-described method steps.

[0054] The beneficial effects of this invention are: the intelligent dynamic power optimization and control method achieves high-precision load forecasting and can dynamically and stably control the grid connection of new energy sources; at the same time, it accurately compensates the end-to-end delay to ±5 milliseconds, can intelligently adjust and adapt to system changes, improves the collaborative efficiency of multiple entities, and has strong anti-delay robustness, which can support the coordinated and optimized operation of source, grid, load and storage. Attached Figure Description

[0055] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0056] Appendix Figure 1 This is a schematic diagram of the intelligent dynamic power optimization and control method of the present invention. Detailed Implementation

[0057] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions in the embodiments of this invention will be clearly and completely described below in conjunction with the embodiments of this invention. Obviously, the described embodiments are merely some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this invention.

[0058] Traditional power dispatching relies heavily on experience and lacks accurate power data support. It does not fully consider the actual operating parameters of the power grid, making it difficult to accurately judge the safe and economical operating status of the power grid. This leads to a decrease in the effectiveness and safety of power dispatching, and an increase in the workload of subsequent dispatching with poor results.

[0059] Traditional power dispatching lacks flexibility and real-time capability, making it difficult to quickly adapt to dynamic changes in the power system, such as the intermittent integration of renewable energy sources and sudden load fluctuations. For example, when the power generation capacity of new energy sources changes suddenly, traditional dispatching methods may not be able to adjust in time, leading to grid instability.

[0060] Moreover, traditional power dispatching is mostly based on a single technology-based control method:

[0061] Adaptive control methods can automatically adjust control strategies according to changes in the internal and external environment of the power grid, but they require monitoring and adjusting the control parameters of the power system, consuming a large amount of data and computing resources.

[0062] Predictive control methods: These require the establishment of accurate predictive models, which depends on the collection and processing of large amounts of data. If the data is inaccurate or the model is imperfect, it will affect the control effect.

[0063] Early applications of intelligent agent technology:

[0064] As the number of agents increases, the computational complexity and coordination difficulty of the system rise significantly. Balancing the autonomy of agents with the integrity of the system becomes a key challenge. For example, in large-scale power systems, the collaborative work among numerous agents may become chaotic, affecting the system optimization effect.

[0065] Meanwhile, intelligent agents face challenges in communication and data interaction, including bandwidth limitations, data privacy, and security. For example, during data transmission, insufficient bandwidth may cause information delays, affecting real-time control, or data security vulnerabilities may lead to system attacks.

[0066] This intelligent dynamic power optimization and control method includes the following steps:

[0067] Step S1: Data Acquisition and Processing

[0068] In the power system, key nodes such as power plants, substations, transmission lines and important load terminals are selected by customization, and smart sensors, including voltage sensors, current sensors, power sensors and frequency sensors, are deployed at the key nodes to collect the operating parameters of the power system in real time, including voltage value, current value, active power, reactive power and frequency.

[0069] After preprocessing the collected data, it is stored in a data warehouse for later use.

[0070] In step S1, the collected data is cleaned, filtered, and interpolated to remove noise and fill in missing values ​​in order to improve data quality.

[0071] Step S2, Power System Modeling

[0072] Component modeling: Establishing accurate mathematical models for various components in the power system, including generators, transformers, transmission lines and loads, to describe their electrical characteristics and operating behavior; for example, generator models can consider their active power-frequency characteristics, reactive power-voltage characteristics, etc.; load models can be modeled according to the characteristics of different types of loads (such as industrial loads, residential loads, commercial loads, etc.).

[0073] System overall modeling: combining the models of the various components to form a model of the entire power system;

[0074] In step S2, the power system model considers the topology, power flow distribution, voltage stability and frequency stability of the power system, and can accurately reflect the behavior of the power system under different operating conditions.

[0075] Step S3, Power Forecasting

[0076] Load forecasting: Based on historical load data and relevant meteorological and economic data, time series analysis and machine learning algorithms (such as neural networks, support vector machines, etc.) are used to forecast the power load for a defined future period (such as short-term hourly, medium-term daily, long-term monthly or yearly). For example, by analyzing the load variation patterns during historical summer high-temperature weather and combining it with current weather forecasts, the peak load periods and load amounts for the next few days can be predicted.

[0077] Renewable energy output forecasting: For wind and solar power generation, based on their geographical location, meteorological conditions (such as wind speed and solar irradiance), and equipment operating status information, a trained forecasting model is used to predict the output for a defined future period. Due to the intermittency and uncertainty of renewable energy, accurate output forecasting is crucial for the optimized control of the power system.

[0078] In step S3, a load forecasting system based on the TCN-Transformer-KAN hybrid model is constructed. The system captures short-term load fluctuation characteristics through dilated convolution, analyzes long-term time series dependencies using a self-attention mechanism, and approximates complex nonlinear mappings with the help of a Kolmogorov-Arnold network.

[0079] When forecasting renewable energy output, forecasting models include wind power forecasting models based on physical processes and photovoltaic power forecasting models based on machine learning.

[0080] In step S3, the power forecast includes short-term forecast, mid-term forecast and long-term forecast, where the short-term forecast is at the hour level, the mid-term forecast is at the day level, and the long-term forecast is at the month level or year level.

[0081] Step S4: Optimize the formulation of control strategies

[0082] Target setting: Based on the operating requirements and actual conditions of the power system, define and determine the optimization and control targets, including minimizing power generation costs, maximizing renewable energy consumption, maintaining system voltage and frequency stability, and improving power supply reliability; for example, during off-peak hours, prioritize the use of renewable energy to reduce fossil fuel consumption and power generation costs; during peak hours, focus on ensuring system voltage and frequency stability to ensure power supply quality.

[0083] Constraint determination: Consider various constraints of the power system, including generator output limits, transmission line capacity limits, transformer capacity limits, allowable ranges of system voltage and frequency, and equipment operation safety constraints, to ensure that the control strategy is feasible in the actual power system and will not lead to equipment overload and system instability;

[0084] Algorithm solution: Using intelligent optimization algorithms or model predictive control methods, the optimal control strategy is solved under the premise of satisfying the constraints;

[0085] In step S4, the optimal control strategy is solved using a genetic algorithm, particle swarm optimization algorithm, or simulated annealing algorithm.

[0086] In step S4, a genetic algorithm is used to search for the optimal solution in the solution space by simulating the selection, crossover and mutation operations in the biological evolution process, so as to determine the control measures, including the output adjustment of each generator, the adjustment of the tap position of the transformer, and the switching of reactive power compensation equipment.

[0087] Step S5: Control Execution and Feedback

[0088] Issuance of control commands: The optimized control strategy is transformed into specific control commands, which are then issued to the corresponding actuators through the power system's automated control equipment (such as automatic generation control devices in power plants and remote control terminals in substations) to achieve real-time control of the power system; for example, issuing commands to generators to increase or decrease their output, controlling their active and reactive power output; issuing commands to substation switching equipment to realize the switching of transmission lines or the adjustment of transformer taps.

[0089] Real-time feedback and adjustment: After the control is implemented, the power system's operating status changes are monitored in real time through the data acquisition system. The actual operating data is compared and analyzed with the predicted data and optimization targets. If the actual situation is found to be inconsistent with the expectations, or if new abnormal operating conditions occur, including sudden faults and load changes, the prediction model and optimization control strategy are adjusted in a timely manner to form a closed-loop feedback control to ensure that the power system is always in the optimal or near-optimal operating state.

[0090] In step S5, GPS timestamps are embedded in the measurement end and the control end respectively to establish a time difference analysis system based on the first timestamp (measurement signal) and the second timestamp (control signal) to obtain communication link delay data in real time.

[0091] The Pade approximation algorithm is used to model the transmission delay in the frequency domain and construct an adaptive delay compensator based on GPS timestamps. By analyzing the time difference between the measurement signal (first timestamp) and the control signal (second timestamp), a frequency domain compensation function is generated, which reduces the phase lag error of the secondary frequency control to within ±5ms.

[0092] After the control is implemented, the Kalman filter algorithm is used to correct the parameters of the power grid component model online based on real-time measurement data, supporting the dynamic updating of key parameters;

[0093] Key parameters include generator rotational inertia and transformer turns ratio, breaking through the limitations of traditional fixed parameter modeling.

[0094] This intelligent dynamic power optimization and control system is used to implement the above methods, including

[0095] The data acquisition and processing module is responsible for customizing and selecting key nodes in the power system, and deploying smart sensors, including voltage sensors, current sensors, power sensors, and frequency sensors, at these key nodes. It collects the operating parameters of the power system in real time, including voltage values, current values, active power, reactive power, and frequency. After preprocessing the collected data, it stores it in a data warehouse for later use.

[0096] The power system modeling module is responsible for establishing accurate mathematical models of various components in the power system, including generators, transformers, transmission lines and loads, to describe their electrical characteristics and operating behavior; and combining the models of each component to form a model of the entire power system.

[0097] The power forecasting module is responsible for forecasting the power load for a defined future period (e.g., short-term at the hourly level, medium-term at the daily level, long-term at the monthly or yearly level) based on historical load data and relevant meteorological and economic data, using time series analysis and machine learning algorithms; and for forecasting the output of wind and solar power generation for a defined future period based on the geographical location, meteorological conditions, and equipment operating status information, using a trained forecasting model.

[0098] The optimization and control strategy formulation module is responsible for defining the optimization and control objectives based on the power system's operational requirements and actual conditions. These objectives include minimizing generation costs, maximizing renewable energy absorption, maintaining system voltage and frequency stability, and improving power supply reliability. Simultaneously, it considers various constraints of the power system, including generator output limits, transmission line capacity limits, transformer capacity limits, allowable system voltage and frequency ranges, and equipment operational safety constraints, to ensure that the control strategy does not lead to equipment overload or system instability in the actual power system. Using intelligent optimization algorithms or model predictive control methods, it solves for the optimal control strategy while satisfying the constraints.

[0099] The control execution and feedback module is responsible for translating the optimized control strategy into specific control instructions, which are then issued to the corresponding actuators through the power system's automated control equipment to achieve real-time control of the power system. After the control is executed, the module monitors the changes in the power system's operating status in real time through the data acquisition system, comparing and analyzing the actual operating data with the predicted data and optimization targets. If the actual situation is found to be inconsistent with expectations, or if new abnormal operating conditions occur, including sudden faults and load changes, the module promptly adjusts the prediction model and optimizes the control strategy to form a closed-loop feedback control, ensuring that the power system is always in an optimal or near-optimal operating state.

[0100] The intelligent dynamic power optimization and control device includes a memory and a processor; the memory is used to store a computer program, and the processor is used to execute the computer program to implement the above-described method steps.

[0101] The readable storage medium stores a computer program that, when executed by a processor, implements the above-described method steps.

[0102] Compared with existing technologies, this intelligent dynamic power optimization and control method has the following characteristics:

[0103] First, high-precision load prediction and stability control

[0104] The improved algorithm improves the accuracy of short-term load forecasting by 40% compared to the traditional LSTM, with the mean absolute percentage error (MAPE) as low as 1.25%. By dynamically adjusting the parameters of the virtual synchronous generator through adaptive inertia control technology, the frequency fluctuation amplitude of the two generators in parallel is reduced by more than 40%, effectively suppressing the active power oscillation caused by the grid connection of new energy sources.

[0105] Second, precise end-to-end latency compensation

[0106] Breaking through the limitations of traditional PID fixed parameter compensation, it supports a dynamic delay range of 0-100ms, compressing the phase lag error of secondary frequency control to within ±5ms, ensuring real-time control response in 5G communication and industrial IoT scenarios.

[0107] Third, intelligent dynamic optimization and control capabilities

[0108] It adapts in real time to intermittent renewable energy generation, random load fluctuations, and changes in grid structure, ensuring system stability and power quality under complex operating conditions; and achieves optimal allocation of resources on the generation side, grid side, and user side based on a multi-entity collaborative algorithm, thereby improving overall operating efficiency.

[0109] Fourth, efficient and meticulous management

[0110] It replaces the manual experience-based scheduling mode, relies on big data analysis and advanced algorithms to achieve comprehensive decision-making across all dimensions of variables, significantly improves the accuracy of regulation and computational efficiency, and supports real-time optimization processing of large-scale data.

[0111] Fifth, multi-entity collaborative interaction

[0112] Establish a linkage mechanism among aggregators, operators, and charging piles to achieve precise control of large-scale electric vehicle charging load; and be compatible with multiple entities such as distributed power sources, energy storage devices, and controllable loads to promote the coordinated operation of power generation, grid, load, and storage.

[0113] Sixth, robustness against delays and uncertainties

[0114] Built-in transmission delay compensation mechanism and uncertainty response strategy ensure that control commands can be executed in a timely and accurate manner under complex communication conditions, enhancing the system's adaptability to the time-varying characteristics of information.

[0115] The embodiments described above are merely one specific implementation of the present invention. Ordinary changes and substitutions made by those skilled in the art within the scope of the technical solution of the present invention should be included within the protection scope of the present invention.

Claims

1. A method and system for intelligent dynamic power optimization and control, characterized in that: Includes the following steps: Step S1: Data Acquisition and Processing In the power system, key nodes are selected by customization, and smart sensors, including voltage sensors, current sensors, power sensors and frequency sensors, are deployed at the key nodes to collect the operating parameters of the power system in real time, including voltage value, current value, active power, reactive power and frequency. The collected data is cleaned, filtered, and interpolated to remove noise, fill in missing values, and stored in a data warehouse for later use. Step S2, Power System Modeling Component modeling: Mathematical models are established for various components in the power system, including generators, transformers, transmission lines and loads, to describe their electrical characteristics and operating behavior; System overall modeling: combining the models of the various components to form a model of the entire power system; Step S3, Power Forecasting Load forecasting: Based on historical load data and relevant meteorological and economic data, time series analysis and machine learning algorithms are used to forecast the power load for a customized future period. Renewable energy output forecasting: For wind and solar power generation, based on their geographical location, meteorological conditions, and equipment operating status information, the trained forecasting model is used to predict the output for a custom future period. Step S4: Optimize the formulation of control strategies Target setting: Based on the operating requirements and actual conditions of the power system, define and determine the optimization and control targets, including minimizing power generation costs, maximizing renewable energy consumption, maintaining system voltage and frequency stability, and improving power supply reliability; Constraint determination: Consider various constraints of the power system, including generator output limits, transmission line capacity limits, transformer capacity limits, allowable ranges of system voltage and frequency, and equipment operation safety constraints, to ensure that the control strategy will not lead to equipment overload and system instability in the actual power system; Algorithm solution: Using intelligent optimization algorithms or model predictive control methods, the optimal control strategy is solved under the premise of satisfying the constraints; Step S5: Control Execution and Feedback Control command issuance: The optimized control strategy is transformed into specific control commands, which are then issued to the corresponding execution agencies through the power system's automated control equipment to achieve real-time control of the power system; Real-time feedback and adjustment: After the control is implemented, the power system's operating status changes are monitored in real time through the data acquisition system. The actual operating data is compared and analyzed with the predicted data and optimization targets. If the actual situation is found to be inconsistent with the expectations, or if new abnormal operating conditions occur, including sudden faults and load changes, the prediction model and control strategy are adjusted in a timely manner to form a closed-loop feedback control to ensure that the power system is always in the optimal operating state.

2. The intelligent dynamic power optimization and control method according to claim 1, characterized in that: In step S2, the power system model considers the topology, power flow distribution, voltage stability, and frequency stability of the power system to reflect the behavior of the power system under different operating conditions.

3. The intelligent dynamic power optimization and control method according to claim 1, characterized in that: In step S3, a load forecasting system based on the TCN-Transformer-KAN hybrid model is constructed. The system captures short-term load fluctuation characteristics through dilated convolution, analyzes long-term time series dependencies using a self-attention mechanism, and approximates complex nonlinear mappings with the help of a Kolmogorov-Arnold network. The prediction models include a wind power prediction model based on physical processes and a photovoltaic power prediction model based on machine learning.

4. The intelligent dynamic power optimization and control method according to claim 3, characterized in that: In step S3, the power forecast includes short-term forecast, mid-term forecast and long-term forecast, where the short-term forecast is at the hour level, the mid-term forecast is at the day level, and the long-term forecast is at the month level or year level.

5. The intelligent dynamic power optimization and control method according to claim 1, characterized in that: In step S4, the optimal control strategy is solved using a genetic algorithm, particle swarm optimization algorithm, or simulated annealing algorithm.

6. The intelligent dynamic power optimization and control method according to claim 5, characterized in that: In step S4, a genetic algorithm is used to search for the optimal solution in the solution space by simulating the selection, crossover and mutation operations in the biological evolution process, so as to determine the control measures, including the output adjustment of each generator, the adjustment of the tap position of the transformer, and the switching of reactive power compensation equipment.

7. The intelligent dynamic power optimization and control method according to claim 1, characterized in that: In step S5, when issuing control commands, the synchronous phasor measurement and Pade approximation algorithm are integrated to construct an adaptive delay compensator based on GPS timestamps; through time difference analysis of the measurement signal and the control signal, a frequency domain compensation function is generated to reduce the phase lag error of the secondary frequency control to within ±5 milliseconds. After the control is implemented, the Kalman filter algorithm is used to correct the parameters of the power grid component model online based on real-time measurement data, supporting the dynamic updating of key parameters; Key parameters include generator moment of inertia and transformer turns ratio.

8. An intelligent dynamic power optimization and control system, characterized in that: To implement the method according to any one of claims 1 to 7, comprising: The data acquisition and processing module is responsible for customizing and selecting key nodes in the power system, and deploying smart sensors, including voltage sensors, current sensors, power sensors, and frequency sensors, at these key nodes. It collects the operating parameters of the power system in real time, including voltage values, current values, active power, reactive power, and frequency. After preprocessing the collected data, it stores it in a data warehouse for later use. The power system modeling module is responsible for establishing mathematical models of various components in the power system, including generators, transformers, transmission lines and loads, to describe their electrical characteristics and operating behavior; and combining the models of each component to form a model of the entire power system. The power forecasting module is responsible for forecasting the power load for a defined future period based on historical load data and relevant meteorological and economic data, using time series analysis and machine learning algorithms; and for forecasting the power output for a defined future period based on the geographical location, meteorological conditions, and equipment operating status information of wind and solar power generation using trained forecasting models. The optimization and control strategy formulation module is responsible for defining the optimization and control objectives based on the power system's operational requirements and actual conditions. These objectives include minimizing generation costs, maximizing renewable energy absorption, maintaining system voltage and frequency stability, and improving power supply reliability. Simultaneously, it considers various constraints of the power system, including generator output limits, transmission line capacity limits, transformer capacity limits, allowable system voltage and frequency ranges, and equipment operational safety constraints, to ensure that the control strategy does not lead to equipment overload or system instability in the actual power system. Using intelligent optimization algorithms or model predictive control methods, it solves for the optimal control strategy while satisfying the constraints. The control execution and feedback module is responsible for translating the optimized control strategy into specific control instructions, which are then issued to the corresponding actuators through the power system's automated control equipment to achieve real-time control of the power system. After the control is executed, the module monitors the changes in the power system's operating status in real time through the data acquisition system, comparing and analyzing the actual operating data with the predicted data and optimization targets. If the actual situation does not match the expectations, or if new abnormal operating conditions occur, including sudden faults and load changes, the module promptly adjusts the prediction model and optimizes the control strategy to form a closed-loop feedback control, ensuring that the power system is always in the optimal operating state.

9. An intelligent dynamic power optimization and control device, characterized in that: It includes a memory and a processor; the memory is used to store a computer program, and the processor is used to execute the computer program to implement the steps of the method as described in any one of claims 1 to 7.

10. A readable storage medium, characterized in that: The readable storage medium stores a computer program that, when executed by a processor, implements the steps of the method as described in any one of claims 1 to 7.

Images