AGC regulation method and system based on multi-objective optimization

Abstract

The invention discloses an AGC adjustment method and system based on multi-objective optimization, and the method comprises the steps: obtaining the output, load demands and new energy output information of each generator set in a real-time operation database of a power system, and obtaining the response capability and cost data of an adjustable load through a demand side management system, the input is used for constructing a multi-objective optimization model; aiming at the constructed high-dimensional nonlinear non-convex multi-objective optimization model, adopting an improved particle swarm optimization algorithm for solving, and improving the convergence speed and the global search capability of the algorithm by introducing an adaptive inertia weight and a diversity retention mechanism to obtain a Pareto optimal solution set; and selecting an optimal frequency modulation scheme from the obtained Pareto optimal solution set according to safety and stability constraints and frequency modulation performance requirements of the power system, determining output adjustment amount and adjustment time of each generator set, and generating a corresponding automatic power generation control instruction.

Description

Technical Field

[0001] The present invention belongs to the technical field of AGC control, and in particular relates to an AGC control method and system based on multi-objective optimization. Background Art

[0002] In the automatic generation control (AGC) of the power system, multiple objectives need to be considered comprehensively, such as frequency regulation cost, frequency regulation speed, frequency regulation accuracy, and environmental protection indicators. These objectives are often contradictory and conflicting, and it is difficult to meet them all at the same time. For example, the pursuit of higher frequency regulation accuracy may increase the frequency regulation cost; the pursuit of faster frequency regulation speed may affect the stability of the system; the pursuit of lower environmental protection indicators may reduce the output of the unit. How to find a balance and compromise between these conflicting objectives is a challenging technical problem.

[0003] In addition, the operating conditions of the power system are complex and changeable, and the load and renewable energy output fluctuate over time. AGC needs to respond to system changes in real time and dynamically adjust the output of each unit. This requires the AGC regulation method to be able to quickly solve the multi-objective optimization model and generate the best regulation strategy online. However, multi-objective optimization models often have the characteristics of high dimension, nonlinearity, and non-convexity, and traditional optimization algorithms are difficult to solve effectively. Therefore, how to design an efficient optimization algorithm is another key technical issue in realizing multi-objective AGC regulation. Summary of the invention

[0004] The present invention proposes an AGC adjustment method and system based on multi-objective optimization to solve the problems existing in the above-mentioned prior art.

[0005] To achieve the above object, the present invention provides an AGC adjustment method based on multi-objective optimization, comprising the following steps:

[0006] Obtain the responsiveness and cost data of adjustable loads from the power system real-time operation database and demand-side management system;

[0007] Based on the real-time operation database of the power system and the responsiveness and cost data of the adjustable load, a multi-objective optimization model is constructed in combination with environmental protection index data;

[0008] The multi-objective optimization model is solved by a particle swarm optimization algorithm to obtain a Pareto optimal solution set;

[0009] According to the safety and stability constraints and frequency regulation performance requirements of the power system, an optimal frequency regulation scheme is selected from the Pareto optimal solution set;

[0010] While running the optimal frequency modulation scheme, the operating parameters of the motor group are collected in real time, and adaptive optimization control is performed according to the operating parameters of the motor group.

[0011] Preferably, the power system real-time operation database includes the output, load demand and new energy output information of each generator set.

[0012] Preferably, the multi-objective optimization model is a network model based on a non-dominated sorting genetic algorithm.

[0013] Preferably, the constraints of the optimization model include the output of the generator set, the ramp rate, the network transmission capacity and the system spare capacity.

[0014] Preferably, an adaptive inertia weight and diversity retention mechanism are introduced into the solution through the particle swarm optimization algorithm to improve the convergence speed and global search capability of the algorithm;

[0015] According to the adaptive inertia weight strategy, the inertia weight is dynamically adjusted; when the algorithm falls into the local optimum, the inertia weight is increased to expand the search range; when the algorithm is close to the global optimum, the inertia weight is reduced to speed up the convergence speed; a diversity maintenance mechanism is introduced. When the diversity of the particle swarm drops below the preset threshold, the positions of some particles are randomly perturbed to maintain the population diversity.

[0016] Preferably, selecting the optimal frequency modulation scheme includes:

[0017] According to the selected optimal frequency regulation scheme, the output adjustment amount and adjustment time of each generator set are determined; the output adjustment amount and adjustment time of the generator set are optimized by using a genetic algorithm to minimize the frequency regulation cost and adjustment time of the system; the optimized output adjustment amount and adjustment time are converted into automatic power generation control instructions.

[0018] Preferably, the method also includes: evaluating the automatic power generation control process, counting the frequency regulation contribution and economy of each generating unit, forming a frequency regulation performance report, using the frequency regulation performance report as input for the next round of optimization, updating the frequency regulation cost function of each unit, adjusting the unit's participation in frequency regulation weights and the constraint condition parameters in the optimization model.

[0019] The present invention also provides an AGC adjustment system based on multi-objective optimization, comprising:

[0020] Data acquisition module, used to obtain the responsiveness and cost data of adjustable loads of the power system real-time operation database and demand side management system;

[0021] The multi-objective optimization modeling module is used to build a multi-objective optimization model based on the real-time operation database of the power system and the responsiveness and cost data of the adjustable load, combined with the environmental protection index data;

[0022] An optimization model solving module, used for solving the multi-objective optimization model by using a particle swarm optimization algorithm to obtain a Pareto optimal solution set;

[0023] A frequency regulation scheme generation module is used to select an optimal frequency regulation scheme from the Pareto optimal solution set according to the safety and stability constraints of the power system and the frequency regulation performance requirements;

[0024] A frequency modulation control execution module, used for collecting the operating parameters of the motor group in real time while running the optimal frequency modulation scheme, and performing adaptive optimization control according to the operating parameters of the motor group;

[0025] The frequency regulation performance evaluation module is used to evaluate the automatic power generation control process, count the frequency regulation contribution and economy of each generator set, form a frequency regulation performance report, use the data of the frequency regulation performance report as the input for the next round of optimization, update the frequency regulation cost function of each unit, adjust the unit's participation in frequency regulation weights, and optimize the constraint condition parameters in the model.

[0026] The present invention also provides a computer device, comprising a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method.

[0027] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the method are implemented.

[0028] Compared with the prior art, the present invention has the following advantages and technical effects:

[0029] The present invention discloses an AGC regulation method and system based on multi-objective optimization. The method obtains data such as generator set output, load demand, new energy output, and adjustable load response capability from a real-time operation database and a demand-side management system, and constructs a multi-objective optimization model that considers frequency regulation cost, speed, accuracy, and environmental protection indicators. An improved particle swarm algorithm is used to solve the high-dimensional nonlinear non-convex model to obtain a Pareto optimal solution set. The optimal frequency regulation scheme is selected according to system constraints to generate automatic power generation control instructions. During the execution process, the unit frequency regulation response is monitored and evaluated in real time, and a new scheme is re-optimized and generated when necessary. The control process is evaluated to form a frequency regulation performance report for updating the input parameters of the next round of optimization. The present invention effectively balances the economy, stability, and environmental protection of frequency regulation through multi-objective optimization and adaptive control, and improves the overall performance of automatic power generation control of the power system. At the same time, the coordinated and optimized utilization of various power generation resources and demand-side resources is realized, providing strong support for the safe and economical operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0031] Figure 1 The figure is a flow chart of a method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0033] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0034] Embodiment 1

[0035] like Figure 1 As shown, this embodiment provides an AGC adjustment method based on multi-objective optimization, including the following steps:

[0036] Obtain the responsiveness and cost data of adjustable loads from the power system real-time operation database and demand-side management system;

[0037] Based on the real-time operation database of the power system and the responsiveness and cost data of the adjustable load, a multi-objective optimization model is constructed in combination with environmental protection index data;

[0038] For the multi-objective optimization model, the particle swarm optimization algorithm is used to solve it and obtain the Pareto optimal solution set;

[0039] According to the safety and stability constraints of the power system and the frequency regulation performance requirements, the optimal frequency regulation scheme is selected from the Pareto optimal solution set;

[0040] While running the optimal frequency modulation solution, the operating parameters of the motor group are collected in real time, and adaptive optimization control is performed based on the operating parameters of the motor group.

[0041] The specific steps include:

[0042] Step S101, obtain the output, load demand and new energy output information of each generator set in the real-time operation database of the power system, and at the same time obtain the response capability and cost data of the adjustable load through the demand side management system as input for building a multi-objective optimization model.

[0043] Specifically, the real-time output data, load demand data and new energy output data of each generator set are obtained through the real-time operation database of the power system. The obtained data are preprocessed, abnormal data are eliminated, and missing data are interpolated to obtain standardized input data. According to the adjustable load response capacity and cost data obtained by the demand-side management system, a response model of demand-side resources is established, and the regulation potential and cost function of different loads are determined as one of the constraints of the optimization model.

[0044] In this embodiment, the optimal dispatching of the power system is a complex multi-objective optimization problem. First, it is necessary to obtain data such as the output of the generator set, load demand, and new energy output from the real-time operation database. For example, the database of a power grid company updates the real-time output data of each power plant every 5 minutes, including thermal power, hydropower, nuclear power, etc. For abnormal data, such as the sudden zero output value of a thermal power unit, it is necessary to remove it and replace it with the average value of the previous and next moments. For the missing data of the output of new energy, the interpolation method can be used to complete it. For example, the output of a wind farm at 15:00 and 16:00 is 100MW and 140MW respectively, then the output at 15:30 can be estimated as 120MW. The demand side management system provides the response capacity and cost data of adjustable loads. For example, the air conditioning load of an industrial park can be reduced by 20MW during the peak period of electricity consumption in summer, and the response cost is 500 yuan / MWh. When establishing a demand-side resource response model, the characteristics of different types of loads need to be considered. For example, the regulation potential of residential electricity consumption is small but the response speed is fast, and the regulation potential of industrial electricity consumption is large but may affect production. These characteristics will serve as constraints for the optimization model.

[0045] Step S102, constructing a multi-objective optimization model based on the acquired frequency regulation cost, frequency regulation speed, frequency regulation accuracy and environmental protection index data, wherein the model comprehensively considers the actual constraints and operation restrictions of the power system, including the upper and lower limits of the output of the generator set, the ramp rate limit, the network transmission capacity limit and the system backup capacity requirements.

[0046] In this embodiment, the actual operation data of the power system is the basis for optimizing the frequency regulation strategy. Taking a certain power grid as an example, the output range of the generator set is 100MW to 500MW, the ramp rate is 2% of the rated capacity per minute, the network transmission capacity is 1000MW, and the system spare capacity is 10% of the total load. These parameters constitute the constraints of the optimization model to ensure the feasibility of the frequency regulation strategy. Frequency regulation cost, speed, accuracy and environmental protection indicators are key factors in multi-objective optimization. Frequency regulation cost includes fuel consumption and equipment loss, which can be expressed as cost per megawatt-hour; frequency regulation speed reflects the system's ability to respond to frequency fluctuations, usually in seconds; frequency regulation accuracy reflects the system's ability to maintain frequency stability, which can be measured by frequency deviation; environmental protection indicators take into account pollutant emissions during power generation, such as carbon dioxide emissions. Intelligent optimization algorithms play an important role in solving complex multi-objective optimization problems. For example, the non-dominated sorting genetic algorithm (NSGA-II) continuously iterates to produce better solutions. For example, the frequency regulation strategy can be encoded as a chromosome, and new strategies can be generated through operations such as crossover and mutation, and their fitness can be evaluated according to the objective function. The particle swarm optimization algorithm simulates group intelligence. Each particle represents a possible frequency modulation strategy and searches for the optimal solution by updating the position and speed.

[0047] Step S103, for the constructed high-dimensional nonlinear non-convex multi-objective optimization model, an improved particle swarm optimization algorithm is used to solve it, and the convergence speed and global search ability of the algorithm are improved by introducing adaptive inertia weight and diversity preservation mechanism to obtain the Pareto optimal solution set.

[0048] Specifically, according to business needs, the optimization objective function and constraints are determined, and a high-dimensional nonlinear non-convex multi-objective optimization mathematical model is constructed. The position and speed of the particle swarm are initialized, and the algorithm parameters are set, including the number of particles, the maximum number of iterations, the learning factor, etc. In each iteration, for each particle, its fitness value is evaluated. The fitness value is determined by the objective function value and the degree of constraint violation. The individual optimal position and the global optimal position of each particle are updated. The individual optimal position is the optimal position of the particle in history, and the global optimal position is the optimal position of the entire particle swarm. According to the adaptive inertia weight strategy, the inertia weight is dynamically adjusted. When the algorithm falls into a local optimum, the inertia weight is increased to expand the search range; when the algorithm is close to the global optimum, the inertia weight is reduced to speed up the convergence. A diversity maintenance mechanism is introduced to prevent the particle swarm from converging prematurely. When the diversity of the particle swarm drops below the preset threshold, the positions of some particles are randomly perturbed to maintain the diversity of the population. Repeat the iterative steps until the maximum number of iterations is reached or the termination condition is met. Output the non-dominated solution set as the Pareto optimal solution for decision makers to select the final solution according to their preferences.

[0049] In this embodiment, when constructing a high-dimensional nonlinear non-convex multi-objective optimization mathematical model, it is necessary to determine the optimization objectives and constraints according to specific business needs. Taking the frequency regulation of the power system as an example, the optimization objectives may include minimizing the frequency regulation cost, maximizing the frequency regulation accuracy, minimizing the environmental impact, etc. The constraints include the upper and lower limits of the output of the generator set, the climbing rate limit, the network transmission capacity limit, etc. This multi-objective optimization problem is usually difficult to solve using traditional mathematical programming methods, so it is more appropriate to use intelligent optimization algorithms such as particle swarm algorithms. The core idea of ​​the particle swarm algorithm is to simulate the foraging behavior of bird flocks. At the beginning of the algorithm, the particle swarm needs to be initialized, and each particle represents a feasible solution to the problem. Taking the frequency regulation strategy optimization as an example, a particle may represent the output adjustment plan of each generator set. The position and speed of the particle are randomly initialized, the position represents the current solution, and the speed represents the update direction and step size of the solution. The setting of the algorithm parameters has an important influence on the optimization effect. For example, too few particles may lead to local optimality, and too many particles will increase the amount of calculation. The maximum number of iterations determines the convergence time of the algorithm, and it is necessary to balance the solution accuracy and efficiency. During the iteration process, the fitness of each particle needs to be evaluated. For the frequency regulation strategy optimization problem, fitness can be composed of multiple objective function values ​​such as frequency regulation cost, accuracy, and environmental impact. At the same time, the degree of constraint violation should also be considered, such as the degree of exceeding the output limit of the generator set. Fitness evaluation provides a basis for subsequent particle updates. Particle update is the core step of the algorithm. The individual optimal position records the best performance of the particle in history, and the global optimal position represents the best performance of the entire group. These two pieces of information guide the particles to move to a better area. In the frequency regulation strategy optimization, this means constantly adjusting the output plan of the generator set and gradually approaching the optimal solution. The adaptive inertia weight strategy can improve the algorithm's search ability. When the algorithm falls into a local optimum, for example, all particles gather around a suboptimal frequency regulation solution, increasing the inertia weight can expand the search range and help jump out of the local optimum. On the contrary, when the algorithm is close to the global optimum, reducing the inertia weight can speed up the convergence and find the exact solution faster. The diversity maintenance mechanism is an important means to prevent the algorithm from converging prematurely. In the frequency regulation strategy optimization, if the particle group gathers around a solution too early, other potential excellent solutions may be missed. By randomly perturbing some particles, the diversity of the population can be maintained and the chance of discovering a better solution can be increased. For example, the output of some generators can be randomly adjusted to explore new frequency regulation strategies. The termination conditions of the algorithm usually include reaching the maximum number of iterations or meeting the preset accuracy requirements. For frequency regulation strategy optimization, a possible termination condition is that the improvement in frequency regulation cost and accuracy is less than a certain threshold. The algorithm outputs a set of non-dominated solutions, namely the Pareto optimal solution set. Each solution represents the optimal trade-off on certain objectives. For example, a solution may sacrifice some frequency regulation accuracy while having a lower frequency regulation cost. Decision makers can select the final solution from the Pareto optimal solution set based on actual needs.For example, when the power system load is light, a solution with high frequency regulation accuracy may be preferred; while during peak power consumption, frequency regulation cost may be more important. This flexibility enables the optimization results to better adapt to different operating scenarios. Through this multi-objective optimization method, a set of high-quality solutions can be found for the power system frequency regulation problem while considering multiple objectives and complex constraints.

[0050] Step S104, from the obtained Pareto optimal solution set, according to the safety and stability constraints and frequency regulation performance requirements of the power system, select the optimal frequency regulation scheme, determine the output adjustment amount and adjustment time of each generator set, and generate corresponding automatic power generation control instructions.

[0051] Specifically, according to the real-time operating status and load demand of the power system, a Pareto optimal solution set is obtained, which satisfies the safety and stability constraints and frequency regulation performance requirements of the system. An optimal frequency regulation scheme is selected from the Pareto optimal solution set, which can maximize the frequency regulation performance and economic benefits of the system while satisfying the constraints. According to the selected optimal frequency regulation scheme, the output adjustment amount and adjustment time of each generator set are determined to ensure the power balance and frequency stability of the system. Intelligent optimization algorithms, such as particle swarm optimization algorithms or genetic algorithms, are used to optimize the output adjustment amount and adjustment time of the generator sets to minimize the frequency regulation cost and adjustment time of the system. The optimized output adjustment amount and adjustment time are converted into automatic power generation control instructions.

[0052] In this embodiment, each solution in the optimal solution set represents a frequency regulation scheme, including the output adjustment amount and adjustment time of each generator set. These schemes can meet the safety and stability constraints of the system, such as voltage not exceeding the limit, line not overloading, etc., and the frequency regulation performance indicators, such as frequency deviation and adjustment time, have reached a relatively good level. For example, scheme A: the output of unit 1 increases by 10 megawatts, and the adjustment time is 5 seconds; the output of unit 2 increases by 15 megawatts, and the adjustment time is 6 seconds. Scheme B: the output of unit 1 increases by 8 megawatts, and the adjustment time is 4 seconds; the output of unit 2 increases by 18 megawatts, and the adjustment time is 7 seconds. Scheme C: the output of unit 1 increases by 12 megawatts, and the adjustment time is 6 seconds; the output of unit 2 increases by 12 megawatts, and the adjustment time is 5 seconds. These schemes have their own advantages and disadvantages in terms of frequency regulation performance and economy, and need to be selected according to actual conditions. Selecting the optimal frequency regulation scheme requires comprehensive consideration of frequency regulation performance and economic benefits. Frequency regulation performance mainly refers to frequency deviation and adjustment time. It is hoped that the deviation is as small as possible and the adjustment time is as short as possible. Economic benefits mainly refer to frequency regulation costs, and it is hoped that the cost is as low as possible. Assuming that in the current situation, more attention is paid to the frequency regulation speed, then option B can be selected. Although its cost is slightly higher than option A, the adjustment time is shorter and the system frequency can be restored faster. After selecting option B, it is necessary to determine the specific output adjustment amount and adjustment time of each generator set. Option B gives a total adjustment amount and an approximate adjustment time range. In practical applications, it is also necessary to further refine the adjustment strategy according to constraints such as the ramp rate of the unit. For example, the output of unit 1 increases by 8 MW and the adjustment time is 4 seconds. If the maximum ramp rate of unit 1 is 2 MW / s, the output can be smoothly increased by 8 MW within 4 seconds. The goal of the intelligent optimization algorithm is to minimize the frequency regulation cost and adjustment time. The frequency regulation cost mainly includes fuel cost, start-stop cost, etc. The adjustment time directly affects the frequency regulation performance. For example, the objective function can be set as: total cost = fuel cost coefficient × output adjustment amount + start-stop cost coefficient × number of unit starts and stops + adjustment time coefficient × adjustment time. By adjusting the weights of each coefficient, the importance of different goals can be reflected. Then, according to the actual situation of the system, set constraints, such as upper and lower limits of unit output, ramp rate limit, line transmission capacity limit, etc. Use intelligent optimization algorithms such as particle swarm optimization algorithm or genetic algorithm to solve and obtain the optimal solution that meets the constraints, that is, the output adjustment amount and adjustment time of each unit. For example, the final optimization result is: the output of unit 1 increases by 8 megawatts, which is completed within 4 seconds; the output of unit 2 increases by 18 megawatts, which is completed within 7 seconds. After obtaining the optimized output adjustment amount and adjustment time, it is necessary to convert them into control instructions. For example, if the output of unit 1 increases by 8 megawatts, the control system can send instructions to the speed governor of unit 1 to increase the valve opening and increase the steam intake, thereby increasing the output. At the same time, send instructions to the excitation system to adjust the excitation current to maintain voltage stability. These instructions need to be customized according to the specific control system and unit model.

[0053] The particle swarm optimization algorithm is used to optimize the output adjustment amount and adjustment time of the generator set to minimize the frequency regulation cost and adjustment time of the system. Specifically, according to the historical operation data of the generator set, a mathematical model of the output adjustment amount and adjustment time of the generator set is established to obtain the optimization objective function. Using the particle swarm optimization algorithm, a group of particles are initialized, and each particle represents a set of solutions for the output adjustment amount and adjustment time. The fitness value of each particle is calculated, and the fitness value is the value of the optimization objective function, that is, the weighted sum of the system frequency regulation cost and adjustment time. The individual optimal solution and the global optimal solution of each particle are updated. The individual optimal solution is the optimal solution of the particle so far, and the global optimal solution is the optimal solution of all particles so far. According to the individual optimal solution and the global optimal solution of the particle, the position and speed of the particle are updated to make the particle move closer to a better solution. It is judged whether the termination condition is met. If it is met, the global optimal solution is output as the final output adjustment amount and adjustment time. Otherwise, it is returned and the iterative optimization is continued. The optimized output adjustment amount and adjustment time are applied to the actual frequency regulation control of the generator set to minimize the frequency regulation cost and adjustment time of the system and improve the economy and stability of the system.

[0054] Step S105, during the execution of the automatic power generation control instruction, the operating parameters of the generator set are collected in real time, and the frequency regulation response of the unit is monitored and evaluated. If it is determined that the frequency regulation performance of the unit does not meet the requirements, the optimization model is re-solved to generate a new frequency regulation scheme to achieve adaptive optimization control of automatic power generation control.

[0055] Specifically, the operating parameters of the generator set, including speed, power, voltage, etc., are collected in real time through sensors, and the collected data is transmitted to the control center. In the control center, the received operating parameter data is analyzed using a machine learning algorithm to evaluate the frequency regulation performance of the generator set and determine whether it meets the preset performance requirement threshold. If the evaluation results show that the frequency regulation performance of the generator set does not meet the requirements, the optimization model is re-solved to generate a new optimal frequency regulation scheme. The optimization model can use a heuristic algorithm, such as a genetic algorithm or a particle swarm optimization algorithm, with the operating parameters of the generator set and the frequency regulation performance evaluation results as input to solve the optimal frequency regulation strategy and control parameters. The new frequency regulation scheme obtained by solving the optimization model is sent to the control system of the generator set, and its frequency regulation control parameters are adjusted to achieve adaptive optimization control.

[0056] In this embodiment, various key parameters of the unit operation are collected in real time through sensors distributed in key parts of the generator set. For example, the speed sensor of a steam turbine generator set will monitor whether its speed is stable at 3000 revolutions per minute in real time, and the power sensor will continuously monitor its active power and reactive power output to ensure that it matches the grid dispatching instructions. The voltage sensor monitors its terminal voltage to ensure that the voltage is stable near the rated value, such as about 10.5 kilovolts. These data are like the "vital signs" information of the unit, and are transmitted to the control center in real time through the communication network. The control center is equipped with a powerful data analysis system, which integrates intelligent algorithms based on machine learning. Taking a coal-fired generator set as an example, the control center will receive historical data such as speed, power, valve opening, and grid frequency deviation data in the past month. Using machine learning algorithms such as support vector machines (SVM) or artificial neural networks (ANN), the system can learn the laws contained in these data and build a frequency regulation performance evaluation model for the unit. For example, by analyzing historical data, it is found that when the load rate of the unit is 80%, its frequency regulation response speed is slow, and the duration of the frequency deviation exceeding 0.2 Hz exceeds 5 seconds. When the load rate is lower than 60%, its frequency regulation response speed is significantly improved. Based on these analyses, the system can evaluate the current frequency regulation performance of the unit and predict its frequency regulation performance under different working conditions. If the evaluation results show that the frequency regulation performance of a generator set is lower than the set threshold, for example, the response speed of frequency regulation is lower than 0.05 Hz per second, or the steady-state error of frequency deviation exceeds 0.1 Hz, the system will determine that its frequency regulation performance does not meet the requirements. At this time, the system will automatically trigger the re-solution of the optimization model. The goal of the optimization model is to find a set of optimal frequency regulation control parameters under the constraints of meeting the safe and stable operation of the power grid, so that the frequency regulation performance of the generator set is improved. For example, for a hydropower generator set, its frequency regulation control parameters may include guide vane opening, blade angle, etc. The optimization model will comprehensively consider the current operating status of the unit, the load demand of the power grid, and the evaluation results of the frequency regulation performance, and use heuristic algorithms such as genetic algorithms or particle swarm optimization algorithms to solve. Genetic algorithms simulate the "survival of the fittest" mechanism in the biological evolution process, and find the optimal control parameter combination through continuous iteration and "mutation". For example, the initial population may contain 100 different control parameter groups, each of which represents a possible frequency regulation strategy. By calculating the frequency regulation performance indicators corresponding to each group of parameters, such as the integrated time absolute error (ITAE) of the frequency deviation, the parameter group with better performance is selected for "crossover" and "mutation" to generate a new parameter group. After multiple generations of iterations, a set of optimal control parameters can be found to achieve the best frequency regulation performance of the unit. The particle swarm optimization algorithm simulates the foraging behavior of a flock of birds. Each "particle" represents a possible control parameter combination. The particles learn from each other and collaborate with each other to continuously update their positions and eventually converge to the optimal solution.Once the optimization model finds a new optimal frequency regulation solution, the control center will send these new control parameters to the control system of the generator set. For example, for a gas turbine, the new control parameters may include the opening of the gas valve, the speed of the compressor, etc. The control system of the generator set will adjust its operating state according to these new parameters, thereby changing its frequency regulation characteristics.

[0057] Step S106, evaluate the automatic power generation control process, count the frequency regulation contribution and economy of each generator set, form a frequency regulation performance report, and use the report data as input for the next round of optimization. By updating the frequency regulation cost function of each unit, adjusting the unit's participation in frequency regulation weight and the constraint condition parameters in the optimization model, the multi-objective optimization level and actual operation effect of automatic power generation control are continuously improved.

[0058] Specifically, according to the preset frequency regulation contribution and economic evaluation model, the frequency regulation contribution and economic score of each unit are calculated to generate a frequency regulation performance report. The data in the frequency regulation performance report is used as the input for the next round of optimization, and the frequency regulation cost function of each unit is updated and fitted through the machine learning algorithm. According to the updated frequency regulation cost function, combined with the real-time operation status of the power grid and the load forecast data, the optimal frequency regulation weight of each unit is calculated through the optimization algorithm. The calculated unit frequency regulation weight is used as a constraint condition, and a multi-objective optimization model is constructed in combination with other operation constraints such as the upper and lower limits of the unit output. The heuristic optimization algorithm is used to solve the multi-objective optimization model, and the optimal output scheduling scheme of each unit in the automatic power generation control process is obtained to achieve a balance between frequency regulation performance and economy. The optimized unit output scheduling scheme is applied to the actual automatic power generation control system, the system operation effect is continuously monitored, and the monitoring data is fed back to the evaluation and optimization links to form a closed-loop optimization.

[0059] In this embodiment, the automatic power generation control system evaluates the frequency regulation performance by collecting the operating data of each generator set in real time. Taking a thermal power plant as an example, the system collects data such as the power generation power, frequency and coal consumption rate of the unit every 5 seconds. These data are input into the preset evaluation model, which calculates the frequency regulation contribution and economic score based on factors such as the frequency response speed, regulation accuracy and economy of the unit. For example, the average frequency regulation contribution of Unit 1 in the last hour is 0.85, the economic score is 0.92, and the comprehensive score is 0.88. The evaluation results are integrated into a frequency regulation performance report as an important basis for optimization. Machine learning algorithms such as support vector regression can use these historical data to continuously update and fit the frequency regulation cost function of each unit. For example, by analyzing the data of the past month, it is found that the frequency regulation cost function of Unit 2 has changed from y=0.05x^2+2x to y=0.04x^2+1.8x, indicating that its frequency regulation efficiency has been improved. The optimization algorithm will combine the updated cost function, the real-time load of the power grid, and the load forecast for the next 24 hours to calculate the optimal weight of each unit for participating in frequency regulation. For example, in a certain period of time, the weight of unit 1 is 0.3, unit 2 is 0.5, and unit 3 is 0.2. These weights are used as constraints to construct a multi-objective optimization model together with other constraints such as the upper and lower limits of the unit output. This model must ensure both the stability of the system frequency and the optimal economy. Heuristic methods such as particle swarm algorithm are used to solve the multi-objective optimization problem and obtain the optimal output scheduling plan for each unit. For example, in the next scheduling cycle, the target output of unit 1 is 280MW, unit 2 is 420MW, and unit 3 is 200MW. This plan can not only meet the frequency regulation needs of the system, but also achieve the lowest overall power generation cost. The optimized scheduling plan is applied to the actual automatic power generation control system. The system will continuously monitor the operating results, such as frequency deviation, AGC performance indicators, etc. These monitoring data will be fed back to the evaluation and optimization links to form a closed-loop optimization process. In this way, the system can continuously adapt to changes in the grid's operating status and always maintain optimal frequency regulation performance and economy.

[0060] This embodiment also provides an AGC adjustment system based on multi-objective optimization, including:

[0061] Data acquisition module, used to obtain the responsiveness and cost data of adjustable loads of the power system real-time operation database and demand side management system;

[0062] The multi-objective optimization modeling module is used to build a multi-objective optimization model based on the real-time operation database of the power system and the responsiveness and cost data of the adjustable load, combined with the environmental protection index data;

[0063] The optimization model solving module is used to solve the multi-objective optimization model through the particle swarm optimization algorithm to obtain the Pareto optimal solution set;

[0064] The frequency regulation scheme generation module is used to select the optimal frequency regulation scheme from the Pareto optimal solution set according to the safety and stability constraints of the power system and the frequency regulation performance requirements;

[0065] The frequency modulation control execution module is used to collect the operating parameters of the motor group in real time while running the optimal frequency modulation solution, and perform adaptive optimization control according to the operating parameters of the motor group;

[0066] The frequency regulation performance evaluation module is used to evaluate the automatic power generation control process, count the frequency regulation contribution and economy of each generating unit, form a frequency regulation performance report, use the data of the frequency regulation performance report as the input for the next round of optimization, update the frequency regulation cost function of each unit, adjust the unit's participation in frequency regulation weights, and optimize the constraint parameters in the model.

[0067] This embodiment further provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method.

[0068] This embodiment also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the method are implemented.

[0069] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. An AGC adjustment method based on multi-objective optimization, characterized in that: The following steps are involved: Obtain the responsiveness and cost data of adjustable loads in the real-time operation database of the power system and the demand side management system; Based on the real-time operation database of the power system and the responsiveness and cost data of the adjustable load, a multi-objective optimization model is constructed in combination with environmental protection index data; The multi-objective optimization model is solved by a particle swarm optimization algorithm to obtain a Pareto optimal solution set; According to the safety and stability constraints and frequency regulation performance requirements of the power system, an optimal frequency regulation scheme is selected from the Pareto optimal solution set; While running the optimal frequency modulation scheme, the operating parameters of the motor group are collected in real time, and adaptive optimization control is performed according to the operating parameters of the motor group.

2. The method according to claim 1, characterized in that The power system real-time operation database includes the output, load demand and new energy output information of each generator set.

3. The method according to claim 1, characterized in that The multi-objective optimization model is a network model based on a non-dominated sorting genetic algorithm.

4. The method according to claim 1, characterized in that: The constraints of the optimization model include the output of the generator set, the ramp rate, the network transmission capacity and the system reserve capacity.

5. The method according to claim 1, characterized in that By introducing adaptive inertia weight and diversity maintenance mechanism into the solution through particle swarm optimization algorithm, the convergence speed and global search ability of the algorithm are improved; According to the adaptive inertia weight strategy, the inertia weight is dynamically adjusted; when the algorithm falls into the local optimum, the inertia weight is increased to expand the search range; when the algorithm is close to the global optimum, the inertia weight is reduced to speed up the convergence speed; a diversity maintenance mechanism is introduced. When the diversity of the particle swarm drops below the preset threshold, the positions of some particles are randomly perturbed to maintain the population diversity.

6. The method according to claim 1, characterized in that After selecting the optimal frequency modulation solution, it includes: According to the selected optimal frequency regulation scheme, the output adjustment amount and adjustment time of each generator set are determined; the output adjustment amount and adjustment time of the generator set are optimized by using a genetic algorithm to minimize the frequency regulation cost and adjustment time of the system; the optimized output adjustment amount and adjustment time are converted into automatic power generation control instructions.

7. The method according to claim 1, characterized in that The method also includes: evaluating the automatic power generation control process, counting the frequency regulation contribution and economy of each generating unit, forming a frequency regulation performance report, using the frequency regulation performance report as input for the next round of optimization, updating the frequency regulation cost function of each unit, adjusting the unit's participation in frequency regulation weights and the constraint condition parameters in the optimization model.

8. An AGC regulation system based on multi-objective optimization, characterized in that: include: Data acquisition module, used to obtain the responsiveness and cost data of adjustable loads of the power system real-time operation database and demand side management system; The multi-objective optimization modeling module is used to build a multi-objective optimization model based on the real-time operation database of the power system and the responsiveness and cost data of the adjustable load, combined with the environmental protection index data; An optimization model solving module, used for solving the multi-objective optimization model by using a particle swarm optimization algorithm to obtain a Pareto optimal solution set; A frequency regulation scheme generation module is used to select an optimal frequency regulation scheme from the Pareto optimal solution set according to the safety and stability constraints of the power system and the frequency regulation performance requirements; A frequency modulation control execution module, used for collecting the operating parameters of the motor group in real time while running the optimal frequency modulation scheme, and performing adaptive optimization control according to the operating parameters of the motor group; The frequency regulation performance evaluation module is used to evaluate the automatic power generation control process, count the frequency regulation contribution and economy of each generator set, form a frequency regulation performance report, use the data of the frequency regulation performance report as the input for the next round of optimization, update the frequency regulation cost function of each unit, adjust the unit's participation in frequency regulation weights, and optimize the constraint condition parameters in the model.

9. A computer device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images