Method and device for quickly optimizing control parameters of one-pipe double-machine pumped storage system

Abstract

The invention belongs to the field of pumped storage unit operation and regulation and control, and particularly discloses a one-pipe double-machine pumped storage system control parameter rapid optimization method which comprises the steps that on the basis of a target pumped storage system research area, unit operation data and dynamic adjustment characteristic parameters are collected, and a frequency adjustment aggregation model of a one-pipe double-machine pumped storage system is established; a multi-dimensional parameter sample matrix is generated by constructing a tensor product; determining a stratified sampling strategy oriented to a multi-dimensional parameter space according to a control parameter-frequency modulation performance mapping relation of the adjusting system, sampling by using the stratified sampling strategy to obtain a multi-layer sample to form a sample set, and performing distributed traversal on unit control parameters by combining the sample set with a multi-dimensional parameter sample matrix and adopting a parallel computing strategy; and performing performance index calculation by using the frequency response data to obtain a target index minimum value, and realizing rapid optimization of the control parameter combination. According to the invention, rapid optimization of the control parameters of the double-unit pumped storage system can be realized.

Description

Technical Field

[0001] The present application relates to the field of operation and regulation of pumped storage units, and more specifically, to a method and device for rapid optimization of control parameters of a one-pipe, two-machine pumped storage system. Background Art

[0002] In recent years, new power systems have placed higher demands on the frequency regulation performance of hydropower. As a key component of system frequency support, pumped storage's rapid response and dynamic performance are crucial to the safe and stable operation of the power grid. Due to differences in the design and operation of pumped storage units, as well as equipment aging and dynamic changes in actual operating conditions, the dynamic characteristics of units operating in parallel within the same plant may deviate to a certain extent. Traditional parameter adjustment analysis and tuning methods often suffer from low computational efficiency and long optimization cycles when faced with multi-dimensional control parameter optimization, making it difficult to meet the real-time requirements of online optimization. Therefore, there is an urgent need for a rapid optimization method for pumped storage system control parameters that can adapt to the frequency regulation requirements of new power systems.

[0003] Currently, most pumped-storage unit control systems rely primarily on adjustments based on a single or a small number of key parameters, employing empirical rules or traditional numerical optimization methods for parameter tuning. However, pumped-storage units typically employ a dual-unit configuration. Under the complex operating conditions and multi-dimensional parameter coupling of two units operating in parallel, traditional control variable methods and parameter scanning methods struggle to effectively characterize the hydraulic coupling characteristics between the units. Furthermore, they struggle to fully capture the parameter space during optimization, limiting optimization results. Furthermore, these methods suffer from low computational resource utilization and excessive reliance on manual experience, making efficient and accurate parameter optimization difficult.

[0004] In addition, although the swarm optimization algorithm has certain advantages in global search effect and accuracy, its optimization time is generally long when dealing with the parameter-performance mapping relationship of complex systems, and it is difficult to achieve fast and real-time parameter optimization, which to a certain extent limits the rapid frequency support capability of the pumped storage system. Summary of the Invention

[0005] In view of the defects of the prior art, the purpose of this application is to provide a method and device for quickly optimizing the control parameters of a one-pipe, two-machine pumped storage system, aiming to solve the problems of long time and low efficiency in optimizing the control parameters of the prior art pumped storage system.

[0006] To achieve the above objectives, in a first aspect, the present application provides a method for rapidly optimizing control parameters of a one-pipe, two-machine pumped storage system, comprising: Based on the target pumped storage system research area, the unit operation data and dynamic regulation characteristic parameters are collected to establish a frequency regulation aggregation model for the one-pipe, two-unit pumped storage system. Selecting unit control parameters to be optimized from the dynamic adjustment characteristic parameters, defining feasible region boundaries and engineering constraints of the multidimensional parameters, and generating a multidimensional parameter sample matrix by constructing a tensor product based on the unit control parameter group, feasible region boundaries and engineering constraints; Based on the mapping relationship between the control parameters and frequency regulation performance of the regulation system, a stratified sampling strategy for the multidimensional parameter space is determined. The stratified sampling strategy is used to sample multiple layers of samples to form a sample set. The sample set is combined with the multidimensional parameter sample matrix using a parallel computing strategy to perform a distributed traversal of the unit control parameters to obtain the frequency response data of all samples. A comprehensive evaluation system for frequency modulation performance indicators is established and a multi-objective evaluation function is defined. The frequency response data is used to calculate the performance indicators, the minimum value of the target indicators is obtained, and the rapid optimization of the control parameter combination is achieved.

[0007] Optionally, the method for constructing the frequency regulation aggregation model specifically includes: Construct a hydraulic-electromechanical coupling mathematical model for a single unit; The hydraulic coupling relationship between the two units sharing the same water diversion pipeline is assumed, and an aggregate modeling method based on the equivalent pipeline assumption is adopted to normalize the actual pipeline network structure into a simplified model. Define the equivalent water head state at the bifurcated pipe inlet as the common water head of the unit, and construct an equivalent single-pipe model by modifying the hydraulic parameters; Establish the relationship between total flow and equivalent head, add up the flow of each unit to get the total flow of the pipeline, and perform equivalent treatment on the length and cross-sectional area of the system segmented pipelines; The total flow equation or transfer function of the pipeline is coupled with the output power equation and speed governor equation of each unit to obtain the frequency regulation aggregation model of the one-pipe two-unit pumped storage system. Among them, the total inertia and damping of the system are the weighted sum of the inertia and damping of each unit respectively.

[0008] Optionally, the hydraulic-electromechanical coupling mathematical model includes a water diversion system model, a turbine model and a generator motor model; the water diversion system model is used to reflect the elastic deformation of the pipeline and the water hammer wave propagation effect, the turbine model includes the dynamic equation of the guide vane opening adjustment, and the generator motor model is used to reflect the rotor inertia and damping characteristics.

[0009] Optionally, generating the multidimensional parameter sample matrix includes: According to the physical characteristics of the system and the control stability requirements, the value range of each control parameter is determined to constitute the parameter feasible region; Perform equally spaced node sampling on each parameter dimension, dividing the value range of each parameter into a preset number of intervals, with each interval corresponding to a discrete value; Perform Cartesian product or tensor product operations on the discrete values of each parameter dimension to generate a multidimensional grid point set covering the entire parameter feasible domain; A four-dimensional control parameter vector is selected as the spatial sample matrix according to the multi-dimensional grid point set.

[0010] Optionally, it also includes: According to the multivariable coupling constraints in actual engineering, the spatial sample matrix is screened to eliminate parameter combinations that do not meet the requirements; The frequency stability of the selected parameter combinations is checked to ensure that all parameter combinations meet the stable operation requirements of the system and obtain the checked multi-dimensional parameter sample matrix.

[0011] Optionally, the stratified sampling strategy includes equal probability stratification, equal interval stratification or sensitivity weighted stratification to achieve stratified hypercube sampling of the multidimensional parameter space.

[0012] Optionally, the sampling space and distributed traversal of the stratified sampling strategy specifically include: Select a stratified sampling strategy based on parameter characteristics; The stratified sampling strategy is determined to be equal interval stratification, dividing each parameter dimension into several non-overlapping sub-interval layers; Latin hypercube sampling or low-discrepancy sequence is used to generate samples in each subinterval layer; Combine the samples generated by each layer to obtain the final sample set; The sample set is divided into several subtask modules and assigned to multiple computing nodes for parallel simulation calculations.

[0013] Optionally, the process of selecting a control parameter combination according to the frequency modulation performance index includes: Selecting system frequency response performance indicators, the performance indicators including: absolute error of integral time, overshoot, and adjustment time; The performance indicators are weighted and combined into comprehensive indicators to comprehensively evaluate the performance of control parameters; Combined with the multi-dimensional parameter sample matrix, each parameter sample is simulated independently to obtain the corresponding performance index value; The parameter combination with the minimum comprehensive performance index and the corresponding multi-dimensional parameter sample matrix are selected to determine the optimal control parameter combination.

[0014] In a second aspect, the present application further provides a device for rapidly optimizing control parameters of a one-pipe, two-machine pumped storage system, comprising: A model building module is used to collect unit operating data and dynamic regulation characteristic parameters based on the target pumped storage system research area, and establish a frequency regulation aggregation model for the one-pipe, two-unit pumped storage system; a matrix generation module for selecting unit control parameters to be optimized from the dynamic adjustment characteristic parameters, defining the feasible region boundary and engineering constraints of the multidimensional parameters, and generating a multidimensional parameter sample matrix by constructing a tensor product based on the unit control parameter group, feasible region boundary and engineering constraints; A parallel computing module is used to determine a stratified sampling strategy for a multidimensional parameter space based on a mapping relationship between control parameters and frequency regulation performance of the regulation system, and to perform a distributed traversal of the unit control parameters using a parallel computing strategy on the multidimensional parameter sample matrix to obtain frequency response data of all samples; The parameter combination optimization module is used to establish a comprehensive evaluation system for frequency modulation performance indicators and define a multi-objective evaluation function, use the frequency response data to calculate the performance indicators, obtain the minimum value of the target indicator, and realize the rapid optimization of the control parameter combination.

[0015] In a third aspect, the present application provides an electronic device comprising: at least one memory for storing programs; and at least one processor for executing the programs stored in the memory. When the program stored in the memory is executed, the processor is used to execute the method described in the first aspect or any possible implementation of the first aspect.

[0016] In a fourth aspect, the present application provides a computer-readable storage medium, which stores a computer program. When the computer program runs on a processor, the processor executes the method described in the first aspect or any possible implementation of the first aspect.

[0017] In a fifth aspect, the present application provides a computer program product, which, when executed on a processor, enables the processor to execute the method described in the first aspect or any possible implementation of the first aspect.

[0018] It can be understood that the beneficial effects of the second to fifth aspects mentioned above can be found in the relevant description of the first aspect mentioned above, and will not be repeated here.

[0019] In general, the above technical solutions conceived by this application have the following beneficial effects compared with the existing technologies: (1) The stratified sampling strategy of this application divides the multidimensional parameter space into multiple subspaces. The parameter combinations in each subspace have similar characteristics. The multidimensional parameter sample matrix constructed by tensor product ensures the uniform distribution of samples in the parameter space, thereby effectively narrowing the range of parameter space that needs to be traversed, reducing the number of simulation calculations, and avoiding unnecessary repeated calculations. Secondly, a parallel computing strategy is adopted to distribute the sample set to multiple computing nodes for distributed traversal and simulation, which greatly shortens the simulation time, fully utilizes the computing resources of multi-core processors or computing clusters, and significantly improves the computing speed, so that more samples can be simulated within a limited time. By establishing a comprehensive evaluation system for frequency modulation performance indicators and defining a multi-objective evaluation function, the performance of each sample can be quickly evaluated and the optimal control parameter combination can be screened out. The solution process of the minimum value of the target indicator directly points to the parameter combination with the best performance, avoiding the tedious comparison and screening process, and realizing the rapid optimization of the control parameter combination, thereby reducing the parameter optimization time and improving the parameter optimization efficiency.

[0020] (2) The stratified sampling and tensor product method of this application ensures the diversity and representativeness of the samples, avoiding the deviation of the optimization results caused by uneven sample distribution. Parallel computing ensures that each sample can be fully simulated and calculated, ensuring the accuracy and reliability of the optimization results. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a flow chart of a method for quickly optimizing control parameters of a one-pipe dual-machine pumped storage system provided in an embodiment of the present application; Figure 2 This is a schematic diagram of the frequency regulation aggregation model structure of a one-pipe dual-machine pumped storage system provided in an embodiment of the present application; Figure 3 This is a schematic diagram of the control parameter sampling and traversal screening steps provided in an embodiment of the present application; Figure 4 Schematic diagram of four-dimensional parameter samples and distribution of pumped storage system provided by the embodiment of the present application; Figure 5 is a schematic diagram of a system frequency-time domain response curve provided in an embodiment of the present application; Figure 6 This is a schematic diagram comparing the computational efficiency of the proposed method provided in the embodiment of this application. Figure 7 This is a schematic diagram of the structure of a device for quickly optimizing control parameters of a one-pipe dual-machine pumped storage system provided in an embodiment of the present application; Figure 8 It is a structural diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0023] The term "and / or" as used herein describes an association between related objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A exists alone, A and B exist simultaneously, or B exists alone. The symbol " / " as used herein indicates that the related objects are in an "or" relationship, for example, A / B means either A or B.

[0024] The terms "first" and "second" in this specification and claims are used to distinguish different objects rather than to describe a specific order of objects. For example, "first response message" and "second response message" are used to distinguish different response messages rather than to describe a specific order of response messages.

[0025] In the embodiments of this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of this application should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0026] In the description of the embodiments of the present application, unless otherwise specified, "multiple" means two or more, for example, multiple processing units means two or more processing units, etc.; multiple elements means two or more elements, etc.

[0027] The embodiments of the present application are described below in conjunction with the drawings in the embodiments of the present application.

[0028] Reference Figure 1 The present application provides a method for rapidly optimizing control parameters of a one-pipe, two-machine pumped storage system, comprising: S101. Based on the target pumped storage system study area, collect unit operating data and dynamic regulation characteristic parameters, and establish a frequency regulation aggregation model for the one-pipe, two-unit pumped storage system; S102. Selecting the unit control parameters to be optimized from the dynamic adjustment characteristic parameters, defining the feasible region boundary and engineering constraints of the multidimensional parameters, and generating a multidimensional parameter sample matrix by constructing a tensor product based on the unit control parameter group, feasible region boundary and engineering constraints; S103. Determine a stratified sampling strategy for the multidimensional parameter space based on the mapping relationship between the control parameters of the regulation system and the frequency regulation performance. Use the stratified sampling strategy to obtain multi-layer samples to form a sample set. Combine the sample set with the multidimensional parameter sample matrix using a parallel computing strategy to perform a distributed traversal of the unit control parameters to obtain frequency response data for all samples. S104. Establish a comprehensive evaluation system for frequency modulation performance indicators and define a multi-objective evaluation function. Use the frequency response data to calculate performance indicators, obtain the minimum value of the target indicator, and achieve rapid optimization of control parameter combinations.

[0029] Specifically, this application first requires comprehensive collection of turbine operating data for a specific target research area. This data typically includes information such as turbine start and stop records, power output, head pressure, and guide vane opening. It details the actual operating status of the turbine over a period of time and serves as an important foundation for analyzing turbine performance and building models.

[0030] In addition to operational data, the unit's dynamic regulation characteristic parameters also need to be collected and analyzed. These parameters may include the unit's response time, regulation speed, stability index, etc. These parameters describe the unit's regulation capability and dynamic characteristics when facing frequency fluctuations.

[0031] Based on this collected operational data and dynamic regulation characteristic parameters, a frequency regulation aggregation model for a dual-unit pumped storage system can be established. This model simulates the frequency response behavior of the units under specific control parameters, providing a foundation for subsequent parameter optimization.

[0032] Secondly, after establishing the frequency regulation aggregation model, it is necessary to select the control parameters that significantly impact the unit's frequency regulation performance from the numerous dynamic regulation characteristic parameters as optimization targets. These control parameters may be variables directly related to the unit's regulation process. Then, based on the actual project and the physical limitations of the unit's operation, it is necessary to define the feasible region boundaries and engineering constraints for these control parameters to ensure that the optimization process is carried out within a safe and feasible range.

[0033] Based on the selected set of unit control parameters, the defined feasible region boundaries, and engineering constraints, a multidimensional parameter sample matrix can be constructed using the mathematical tensor product method. Each element in this matrix represents a specific set of control parameter combinations, covering the entire feasible region space, providing a comprehensive and evenly distributed sample space for subsequent parameter optimization. In this way, the impact of different control parameter combinations on the unit's frequency regulation performance can be systematically explored.

[0034] Furthermore, to more efficiently explore the multidimensional parameter space, it is necessary to determine an appropriate stratified sampling strategy based on the mapping between the control parameters of the regulation system and the frequency modulation performance. This strategy divides the multidimensional parameter space into several subspaces, or layers, where the parameter combinations within each subspace share certain similarities or characteristics. By sampling each subspace, we ensure sample diversity and representativeness while reducing the number of samples required to be traversed, improving computational efficiency.

[0035] Using a defined stratified sampling strategy, a certain number of samples are drawn from each subspace to form a multi-layered sample set. This sample set is then combined with the previously generated multidimensional parameter sample matrix, and a parallel computing strategy is employed to perform a distributed traversal of the unit control parameters. This involves distributing the samples across multiple computing nodes or processors for parallel simulation, significantly shortening simulation time and obtaining frequency response data for all samples. This parallel computing strategy fully utilizes modern computing resources and significantly accelerates parameter optimization.

[0036] Finally, after obtaining the frequency response data for all samples, a comprehensive frequency regulation performance evaluation system needs to be established, and a multi-objective evaluation function needs to be defined. This evaluation system may include multiple performance indicators, such as integrated time absolute error (ITAE), overshoot, and regulation time, which reflect the frequency regulation performance of the unit from different perspectives. The multi-objective evaluation function combines these performance indicators into a single evaluation standard to measure the performance of different control parameter combinations.

[0037] Using the acquired frequency response data, we calculate the target indicator value for each sample based on the defined performance indicators and evaluation functions. By comparing these indicator values, we find the control parameter combination that minimizes the target indicator. This combination is the optimal control parameter solution within the current sample space. This method enables rapid optimization of control parameter combinations, providing an optimal control strategy for frequency regulation in pumped storage systems.

[0038] Optionally, the method for constructing the frequency regulation aggregation model specifically includes: Construct a hydraulic-electromechanical coupling mathematical model for a single unit; The hydraulic coupling relationship between the two units sharing the same water diversion pipeline is assumed, and an aggregate modeling method based on the equivalent pipeline assumption is adopted to normalize the actual pipeline network structure into a simplified model. Define the equivalent water head state at the bifurcated pipe inlet as the common water head of the unit, and construct an equivalent single-pipe model by modifying the hydraulic parameters; Establish the relationship between total flow and equivalent head, add up the flow of each unit to get the total flow of the pipeline, and perform equivalent treatment on the length and cross-sectional area of the system segmented pipelines; The total flow equation or transfer function of the pipeline is coupled with the output power equation and speed governor equation of each unit to obtain the frequency regulation aggregation model of the one-pipe two-unit pumped storage system. Among them, the total inertia and damping of the system are the weighted sum of the inertia and damping of each unit respectively.

[0039] Specifically, a mathematical model of the regulation system of a single unit is first established, taking into account the hydraulic-electromechanical coupling. The hydraulic model of the water diversion system has a significant impact on the operation and regulation of the pumped storage unit. To this end, a mathematical model of the water diversion system that includes the elasticity of the pressure pipe and the inertia of the water flow is established as follows:

[0040] In the formula are the system head and flow changes respectively; is the hydraulic impedance of the pipeline, which reflects the hindering effect of pipeline characteristics on the propagation of pressure waves; a 1 is the pipeline correlation coefficient, A is the cross-sectional area of the pipe, g is the acceleration due to gravity; is the hydraulic inertia time constant; Because water hits each other, L is the pipeline distance, c is the water hammer wave velocity.

[0041] Expand the above into Taylor series and simplify the transfer function of the water diversion system to obtain:

[0042] In the formula A represents the cross-sectional area of the pipe, h w Indicates pipeline characteristic parameters, T r Indicates that water hammer is long. s represents the complex variable in the Laplace transform.

[0043] The dynamic characteristics of the pump-turbine-speed governor are the key link in the frequency control of the pumped storage system. Its mathematical description can be jointly characterized by the controller law, the relay action logic and the turbine power dynamic equation. Among them, the speed control system is used to adjust the pump-turbine guide vane opening to control the unit output. Its core controller generally adopts PI control, and the controller output then drives the hydraulic servo system through the relay. Assume that the unit frequency deviation As the input signal, the controller outputs the guide vane opening instruction, and the guide vane control equation satisfies:

[0044] Where, G 1( s) represents the transfer function between the guide vane opening and the frequency deviation, The Laplace transform of the guide vane opening variation is represented by: represents the Laplace transform of the frequency deviation, is the proportional gain, K i is the integral gain, T y is the response time constant of the unit servomotor.

[0045] The dynamic equations for the mechanical torque and flow of the pump-turbine as a function of the speed and opening of the group are as follows:

[0046] in, is the relative value of the pump-turbine torque deviation; is the relative value of pump-turbine flow deviation; is the relative value of the speed deviation, is the relative value of the pump-turbine head deviation; is the relative value of the guide vane opening deviation; is the transfer coefficient of pump-turbine torque to guide vane opening; is the pump-turbine torque-to-speed transfer coefficient; is the transfer coefficient of pump-turbine torque to working head; is the transfer coefficient of pump-turbine flow to guide vane opening; is the pump-turbine flow-to-speed transfer coefficient; is the transfer coefficient of pump-turbine flow to working head.

[0047] The dynamic response of the generator motor and the load system can be described by equivalent inertia and damping. Assume that the generator motor speed is n , the frequency response dynamic equation of the unit is:

[0048] in, J is the equivalent moment of inertia of the unit; P e is the electromagnetic power of the generator motor, which is related to the load demand; D is the system damping coefficient; n s is the synchronous angular velocity of the generator motor, is the unit angular velocity, P m is the mechanical power.

[0049] Next, a frequency regulation aggregation model of a one-pipe dual-machine pumped storage system is established.

[0050] In a dual-unit pumped-storage system, two units (units 1 and 2) share a common water supply pipeline. Due to the coupled hydraulic characteristics within the pipeline, flow changes in each unit will dynamically adjust the pressure distribution within the pipeline, thus influencing the flow and hydraulic head of the other unit. To accurately describe the hydraulic coupling between these units, a corresponding aggregate modeling method is required to characterize the impact of pipeline hydraulic characteristics on the unit's operating characteristics.

[0051] For pumped storage systems with complex pipe connection structures and different characteristic parameters, an aggregate modeling method based on the equivalent pipe assumption can be used. This method normalizes the actual pipe network structure composed of multiple pipe sections with different characteristic parameters (including length, diameter, friction coefficient, etc.) into a simplified model through equivalent processing, thereby characterizing the overall hydraulic characteristics of the system and optimizing the dynamic response analysis of the coupled units. An equivalent head state is defined at the inlet of the bifurcated pipe. H eq , regarded as the common water head of the unit, and by correcting the hydraulic parameters such as equivalent inertia time constant, equivalent wave velocity, etc., an equivalent single-pipe model is constructed, and then the mechanical power or flow part is split on the unit side. Specifically, the flow of each unit Q i The total flow rate of the pipeline is obtained by adding up:

[0052] in, Q total is the total flow rate in the pipeline, Q i For the i The flow rate of the unit, n is the number of units.

[0053] Then, by performing equivalent processing on the length and cross-sectional area of the system's segmented pipes, the relationship between the total flow and the equivalent head is obtained as follows:

[0054] Where, It is the equivalent hydraulic parameter, which can be obtained by converting the original pipeline into multiple branches or by fitting the test data; G v,i and are the opening change and weight coefficient of each unit respectively. G total is the transfer function between total flow and equivalent head, H eq It is the equivalent head state. By coupling the above pipeline total flow equation or transfer function with the output power equation and speed governor equation of each unit, we can obtain the frequency regulation aggregation model of the one-pipe, two-unit pumped storage system. The total inertia and damping of the system satisfy:

[0055] in, J total is the total inertia of the system, D total is the total damping of the system, is the inertia weight or damping weight of each unit, J i is the unit inertia, D i For the unit damping.

[0056] Furthermore, the hydraulic-electromechanical coupling mathematical model includes a water diversion system model, a turbine model and a generator motor model; the water diversion system model is used to reflect the elastic deformation of the pipeline and the water hammer wave propagation effect, the turbine model includes the dynamic equation for guide vane opening adjustment, and the generator motor model is used to reflect the rotor inertia and damping characteristics.

[0057] Optionally, generating the multidimensional parameter sample matrix includes: According to the physical characteristics of the system and the control stability requirements, the value range of each control parameter is determined to constitute the parameter feasible region; Perform equally spaced node sampling on each parameter dimension, dividing the value range of each parameter into a preset number of intervals, with each interval corresponding to a discrete value; Perform Cartesian product or tensor product operations on the discrete values of each parameter dimension to generate a multidimensional grid point set covering the entire parameter feasible domain; A four-dimensional control parameter vector is selected as the spatial sample matrix according to the multi-dimensional grid point set.

[0058] Optionally, it also includes: According to the multivariable coupling constraints in actual engineering, the spatial sample matrix is screened to eliminate parameter combinations that do not meet the requirements; The frequency stability of the selected parameter combinations is checked to ensure that all parameter combinations meet the stable operation requirements of the system and obtain the checked multi-dimensional parameter sample matrix.

[0059] Specifically, for the multi-dimensional control parameter optimization problem in the frequency regulation process of the pumped storage system, the parameter group of the dual PI controller of the speed regulation system is selected as the optimization object: Based on the physical characteristics of the system and the control stability requirements, the parameter feasible region is defined as follows:

[0060] Where, To control the parameter amplitude limit, is the time constant constraint, which guarantees the lower limit of the PI controller time constant.

[0061] In multidimensional parameter optimization problems, constructing a sample set covering the entire feasible domain is crucial to fully explore the impact of each control parameter on system performance. To this end, the tensor product method is used to generate a discretized sample matrix to ensure uniformity and representativeness of the sample.

[0062] The discretization strategies for each parameter dimension are as follows: 1) Single parameter discretization: First, perform equal-spaced node sampling on each parameter dimension and set the number of discrete points of the control parameter to n Kp , then the discretized set of system control parameters is:

[0063] 2) Multidimensional parameter tensor product construction: In order to efficiently traverse the multidimensional parameter space in numerical simulation or optimization, it is necessary to select several discrete sampling points in each dimension and perform a Cartesian product or tensor product of these discrete values in each dimension to form a multidimensional grid point set covering the parameter feasible domain. For the control parameter optimization problem of a single-pipe, dual-machine pumped storage system, a four-dimensional control parameter vector is selected as the spatial sample matrix:

[0064] The total number of sample points for obtaining control parameters , each sample point corresponds to a unique parameter combination that satisfies:

[0065] 3) Constraint embedding: In order to eliminate infeasible samples, a constraint operator is introduced ,in: In addition to the individual boundary constraints of each dimension, practical engineering often also involves coupling constraints between multiple variables. For example, to prevent the gain difference between two units from being too large, which would lead to uncoordinated regulation, the following constraints can be introduced:

[0066] In the formula It is the maximum difference allowed in the adjustment process of the two units.

[0067] After the tensor product constructs the complete sample matrix, all parameter sample amplitudes, time constants, and coupling constraints can be screened and verified to ensure the frequency stability of the control parameters:

[0068] Where, Kp1 、 K p2 are the proportional gains of the first unit and the second unit respectively, K i1 、 K i2 are the integration time of the first unit and the second unit respectively; is the time constant constraint, which ensures the lower limit of the PI controller time constant Optionally, the stratified sampling strategy includes equal probability stratification, equal interval stratification or sensitivity weighted stratification to achieve stratified hypercube sampling of the multidimensional parameter space.

[0069] The sampling space and distributed traversal of the stratified sampling strategy specifically include: Select a stratified sampling strategy based on parameter characteristics; The stratified sampling strategy is determined to be equal interval stratification, dividing each parameter dimension into several non-overlapping sub-interval layers; Latin hypercube sampling or low-discrepancy sequence is used to generate samples in each subinterval layer; Combine the samples generated by each layer to obtain the final sample set; The sample set is divided into several subtask modules and assigned to multiple computing nodes for parallel simulation calculations.

[0070] Specifically, the following are some common stratification criteria in the present application: 1) Equal probability stratification: each stratum contains parameter distributions of the same probability mass, which is applicable to normal distribution parameters; 2) Equal interval stratification: the parameter range is evenly divided into m i subintervals; 3) Sensitivity weighted stratification: high-sensitivity parameter dimensions are divided more densely.

[0071] Taking equal-interval stratification as an example, for the four-dimensional control parameter space , divide each parameter dimension into m i layers of non-overlapping subintervals to implement stratified hypercube sampling:

[0072] in, Indicates the k The first dimension parameter j subintervals.

[0073] On each layer In the code, Latin hypercube sampling (LHS) or Sobol sequence is used to generate n per_stratum samples to ensure uniform coverage of samples within the layer. The mathematical properties of LHS can guarantee:

[0074] Where, N ( B ) is the falling area The number of samples, n s represents the total number of samples in each layer, represents the measure of region B, C is a constant related to the sampling properties.

[0075] The total number of samples is:

[0076] Using the parallel simulation computing framework, the parameter sample matrix Θ is divided into P subtask modules, assigned to p Compute nodes:

[0077] Θ represents the parameter sample matrix, Θ p Indicates the assignment to p The subtask module of computing nodes, Θ q Indicates the assignment to q A subtask module for each computing node.

[0078] Assume that the single simulation time is t f , the communication overhead is t c , then the total parallel computing time is:

[0079] in, T parallel To calculate the total time, Indicates a round-up operation.

[0080] The speedup ratio and computational efficiency are:

[0081] in, S p Indicates the degree of acceleration of parallel computing relative to serial computing, T serial Indicates the total serial calculation time, Indicates computational efficiency.

[0082] when and When, there is approximation of , i.e. linear acceleration.

[0083] Optionally, the process of selecting a control parameter combination according to the frequency modulation performance index includes: Selecting system frequency response performance indicators, the performance indicators including: absolute error of integral time, overshoot, and adjustment time; The performance indicators are weighted and combined into comprehensive indicators to comprehensively evaluate the performance of control parameters; Combined with the multi-dimensional parameter sample matrix, each parameter sample is simulated independently to obtain the corresponding performance index value; The parameter combination with the minimum comprehensive performance index and the corresponding multi-dimensional parameter sample matrix are selected to determine the optimal control parameter combination.

[0084] Specifically, the system frequency response performance indicators are selected as follows: 1) Integral Time Absolute Error (ITAE): , reflects the controller's ability to suppress long-term frequency deviation, and is related to the integral gain K i Strong correlation.

[0085] in, t is the time variable, Indicates time t When the control parameter θ The frequency deviation caused by T is the total simulation time; 2) Overshoot:

[0086] in, is the steady-state frequency deviation.

[0087] 3) Adjustment time: , and the system equivalent inertia constant H eq Related to controller bandwidth.

[0088] Define comprehensive frequency modulation performance indicators:

[0089] in, is the standard value under the benchmark parameters, and the weight of each indicator satisfies . Combined with the aforementioned multi-dimensional parameter sample matrix, for each sample Perform independent simulations to calculate the evaluation function groups for all parameter combinations .

[0090] Among many samples, The minimum value and the corresponding sample matrix can quickly screen out the optimal dual-machine control parameter solution with the best frequency modulation performance index: .in, It is a comprehensive performance indicator.

[0091] The following embodiment takes the basic operating data of a domestic dual-unit pumped storage system as an example (corresponding model parameters are shown in Table 1) to establish a frequency regulation aggregation model for a dual-unit pumped storage system. Figure 2 shown. Figure 2 The system consists of an upper reservoir, a lower reservoir, a shared water diversion system, two parallel turbines, and their own regulating controllers. This hydraulic structure utilizes a dual-turbine configuration, resulting in significant fluid coupling between the two turbines. Furthermore, each turbine in the system consists of a pump-turbine, a generator motor, and an independent speed control system. The controller adjusts the turbine's power output in real time to respond to grid frequency fluctuations and maintain system stability. However, since the turbines share a shared water diversion system, independent control of each unit can lead to uncoordinated overall regulation, impacting system frequency stability. Therefore, the model must ensure both system frequency response speed and the overall system dynamic performance.

[0092] Table 1. Model parameters of a dual-unit pumped storage system

[0093] Aiming at the rapid optimization problem of dual PI control parameters of a one-pipe dual-machine pumped storage system, an optimization process based on multi-dimensional parameter sampling and hierarchical traversal screening is adopted. Figure 3 First, based on the system physical characteristics and engineering constraints, the four-dimensional control parameters are defined. K p1 、 K i1 、 K p2 、 K i2 Feasible region: controller parameters K p1 、 K p2 Normal distribution N (3.2,0.52), N (3.0,0.52), K i1 、 K i2 obey N (0.8,0.12), N (1.0,0.12), its sample and distribution are as follows Figure 4As shown in the figure, the amplitude limit and the lower limit of the time constant are set, and the unit gain difference constraint is introduced to limit the maximum difference value of the regulation process of the two units. Secondly, the candidate nodes of each dimension are generated by equidistant discretization sampling of a single parameter, and the full coverage sample matrix of the four-dimensional parameters is constructed using the tensor product. The constraint operator is embedded to eliminate invalid samples that do not meet the amplitude, time constant and coupling relationships. On this basis, the parameter space is divided into multiple layers of sub-intervals using an equidistant layering strategy, and the sample distribution density within the layer is optimized in combination with Latin hypercube sampling (LHS) to ensure fine coverage of high-sensitivity parameter dimensions. Finally, the screened parameter samples are simulated in blocks through a distributed parallel computing framework, and the frequency stability index is used as the evaluation criterion to quickly screen out the parameter combination with the best comprehensive regulation performance.

[0094] The optimization results of control parameters for the pumped storage system with two units per pipe are as follows: Figure 5 The system frequency-time response curve under the optimal parameter combination screened by the proposed method is shown. Compared with the initial parameter configuration, the optimized controller parameters K p1 、 K i1 、 K p2 、 K i2 The system's frequency stability under dynamic disturbances was significantly improved, manifested in reduced frequency overshoot, faster oscillation convergence, and shorter steady-state recovery time. This result demonstrates that multidimensional constraint screening and a stratified sampling strategy can effectively balance the dynamic characteristics and robustness requirements of dual-machine coordinated regulation.

[0095] Figure 6 The computational efficiency of the proposed stratified sampling-parallel computing framework was further compared with that of the traditional Monte Carlo method. Experimental results show that the proposed method generates a global sample matrix through tensor product discretization and optimizes the local sample distribution through stratified sampling, significantly reducing the redundant computation of invalid samples. At the same time, the distributed parallel architecture splits the computational task into multiple submodules, fully utilizing computing resources, thereby achieving faster optimization convergence speed at the same simulation scale. Compared with traditional methods, the proposed framework reduces the use of computing resources while ensuring the accuracy of parameter optimization, shortening the calculation speed to seconds, verifying its engineering applicability in complex multivariable control systems. Overall, this method provides an efficient and reliable solution for the optimization of control parameters of a single-pipe, dual-machine pumped storage system through the "global coverage-stratified refinement-parallel acceleration" approach.

[0096] Reference Figure 7 The present application also provides a device for rapidly optimizing control parameters of a one-pipe dual-machine pumped storage system, comprising: A model building module 710 is used to collect unit operating data and dynamic regulation characteristic parameters based on the target pumped storage system study area, and establish a frequency regulation aggregation model for the one-pipe two-unit pumped storage system; A matrix generation module 720 is configured to select unit control parameters to be optimized from the dynamic adjustment characteristic parameters, define the feasible region boundary and engineering constraints of the multi-dimensional parameters, and generate a multi-dimensional parameter sample matrix by constructing a tensor product based on the unit control parameter group, feasible region boundary and engineering constraints; A parallel computing module 730 is configured to determine a stratified sampling strategy for a multidimensional parameter space based on a mapping relationship between control parameters and frequency regulation performance of the regulation system, and to perform a distributed traversal of the unit control parameters using the parallel computing strategy on the multidimensional parameter sample matrix to obtain frequency response data for all samples; The parameter combination optimization module 740 is used to establish a comprehensive evaluation system for frequency modulation performance indicators and define a multi-objective evaluation function, use the frequency response data to calculate performance indicators, obtain the minimum value of the target indicator, and achieve rapid optimization of control parameter combinations.

[0097] Optionally, the method for constructing the frequency regulation aggregation model specifically includes: Construct a hydraulic-electromechanical coupling mathematical model for a single unit; The hydraulic coupling relationship between the two units sharing the same water diversion pipeline is assumed, and an aggregate modeling method based on the equivalent pipeline assumption is adopted to normalize the actual pipeline network structure into a simplified model. Define the equivalent water head state at the bifurcated pipe inlet as the common water head of the unit, and construct an equivalent single-pipe model by modifying the hydraulic parameters; Establish the relationship between total flow and equivalent head, add up the flow of each unit to get the total flow of the pipeline, and perform equivalent treatment on the length and cross-sectional area of the system segmented pipelines; The total flow equation or transfer function of the pipeline is coupled with the output power equation and speed governor equation of each unit to obtain the frequency regulation aggregation model of the one-pipe two-unit pumped storage system. Among them, the total inertia and damping of the system are the weighted sum of the inertia and damping of each unit respectively.

[0098] Optionally, the hydraulic-electromechanical coupling mathematical model includes a water diversion system model, a turbine model and a generator motor model; the water diversion system model is used to reflect the elastic deformation of the pipeline and the water hammer wave propagation effect, the turbine model includes the dynamic equation of the guide vane opening adjustment, and the generator motor model is used to reflect the rotor inertia and damping characteristics.

[0099] Optionally, generating the multidimensional parameter sample matrix includes: According to the physical characteristics of the system and the control stability requirements, the value range of each control parameter is determined to constitute the parameter feasible region; Perform equally spaced node sampling on each parameter dimension, dividing the value range of each parameter into a preset number of intervals, with each interval corresponding to a discrete value; Perform Cartesian product or tensor product operations on the discrete values of each parameter dimension to generate a multidimensional grid point set covering the entire parameter feasible domain; A four-dimensional control parameter vector is selected as the spatial sample matrix according to the multi-dimensional grid point set.

[0100] Optionally, it also includes: According to the multivariable coupling constraints in actual engineering, the spatial sample matrix is screened to eliminate parameter combinations that do not meet the requirements; The frequency stability of the selected parameter combinations is checked to ensure that all parameter combinations meet the stable operation requirements of the system and obtain the checked multi-dimensional parameter sample matrix.

[0101] Optionally, the stratified sampling strategy includes equal probability stratification, equal interval stratification or sensitivity weighted stratification to achieve stratified hypercube sampling of the multidimensional parameter space.

[0102] Optionally, the sampling space and distributed traversal of the stratified sampling strategy specifically include: Select a stratified sampling strategy based on parameter characteristics; The stratified sampling strategy is determined to be equal interval stratification, dividing each parameter dimension into several non-overlapping sub-interval layers; Latin hypercube sampling or low-discrepancy sequence is used to generate samples in each subinterval layer; Combine the samples generated by each layer to obtain the final sample set; The sample set is divided into several subtask modules and assigned to multiple computing nodes for parallel simulation calculations.

[0103] Optionally, the process of selecting a control parameter combination according to the frequency modulation performance index includes: Selecting system frequency response performance indicators, the performance indicators including: absolute error of integral time, overshoot, and adjustment time; The performance indicators are weighted and combined into comprehensive indicators to comprehensively evaluate the performance of control parameters; Combined with the multi-dimensional parameter sample matrix, each parameter sample is simulated independently to obtain the corresponding performance index value; The parameter combination with the minimum comprehensive performance index and the corresponding multi-dimensional parameter sample matrix are selected to determine the optimal control parameter combination.

[0104] It should be understood that the above-mentioned device is used to execute the method in the above-mentioned embodiment. The implementation principle and technical effect of the corresponding program module in the device are similar to those described in the above-mentioned method. The working process of the device can refer to the corresponding process in the above-mentioned method and will not be repeated here.

[0105] Based on the methods in the above embodiments, an embodiment of the present application provides an electronic device, which may include: a processor (Processor) 810, a communication interface (Communications Interface) 820, a memory (Memory) 830, and a communication bus 840, wherein the processor 810, the communication interface 820, and the memory 830 communicate with each other via the communication bus 840. The processor 810 can call the logic instructions in the memory 830 to execute the methods in the above embodiments.

[0106] In addition, the logic instructions in the aforementioned memory 830 can be implemented in the form of a software functional unit and, when sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the portion that contributes to the prior art, or the portion of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application.

[0107] Based on the method in the above embodiment, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When the computer program runs on a processor, the processor executes the method in the above embodiment.

[0108] Based on the method in the above embodiment, an embodiment of the present application provides a computer program product. When the computer program product runs on a processor, the processor executes the method in the above embodiment.

[0109] It is understood that the processor in the embodiments of the present application may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field programmable gate arrays (FPGA), other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. The general-purpose processor may be a microprocessor or any conventional processor.

[0110] The method steps in the embodiments of the present application can be implemented by hardware or by a processor executing software instructions. The software instructions can be composed of corresponding software modules, which can be stored in random access memory (RAM), flash memory, read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), registers, hard disks, mobile hard disks, CD-ROMs, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be an integral part of the processor. The processor and storage medium can be located in an ASIC.

[0111] The above embodiments can be implemented in whole or in part using software, hardware, firmware, or any combination thereof. When implemented using software, they can be implemented in whole or in part in the form of a computer program product. The computer program product comprises one or more computer instructions. When loaded and executed on a computer, the computer program instructions fully or partially produce the processes or functions described in the embodiments of this application. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted via the computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible by a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium can be magnetic media (e.g., floppy disk, hard disk, tape), optical media (e.g., DVD), or semiconductor media (e.g., solid-state drive (SSD)).

[0112] It will be understood that the various numerical numbers involved in the embodiments of the present application are merely distinctions for the convenience of description and are not intended to limit the scope of the embodiments of the present application.

[0113] It is easy for those skilled in the art to understand that the above is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. A method for rapid optimization of control parameters of a one-pipe dual-machine pumped storage system, characterized in that: include: Based on the target pumped storage system research area, the unit operation data and dynamic regulation characteristic parameters are collected to establish a frequency regulation aggregation model for the one-pipe, two-unit pumped storage system. Selecting unit control parameters to be optimized from the dynamic adjustment characteristic parameters, defining feasible region boundaries and engineering constraints of the multidimensional parameters, and generating a multidimensional parameter sample matrix by constructing a tensor product based on the unit control parameter group, feasible region boundaries and engineering constraints; Based on the mapping relationship between the control parameters and frequency regulation performance of the regulation system, a stratified sampling strategy for the multidimensional parameter space is determined. The stratified sampling strategy is used to sample multiple layers of samples to form a sample set. The sample set is combined with the multidimensional parameter sample matrix using a parallel computing strategy to perform a distributed traversal of the unit control parameters to obtain the frequency response data of all samples. A comprehensive evaluation system for frequency modulation performance indicators is established and a multi-objective evaluation function is defined. The frequency response data is used to calculate the performance indicators, the minimum value of the target indicators is obtained, and the rapid optimization of the control parameter combination is achieved.

2. The method for rapid optimization of control parameters of a one-pipe dual-machine pumped storage system according to claim 1 is characterized in that: The method for constructing the frequency regulation aggregation model specifically includes: Construct a hydraulic-electromechanical coupling mathematical model for a single unit; The hydraulic coupling relationship between the two units sharing the same water diversion pipeline is assumed, and an aggregate modeling method based on the equivalent pipeline assumption is adopted to normalize the actual pipeline network structure into a simplified model. Define the equivalent water head state at the bifurcated pipe inlet as the common water head of the unit, and construct an equivalent single-pipe model by modifying the hydraulic parameters; Establish the relationship between total flow and equivalent head, add up the flow of each unit to get the total flow of the pipeline, and perform equivalent treatment on the length and cross-sectional area of the system segmented pipelines; The total flow equation or transfer function of the pipeline is coupled with the output power equation and speed governor equation of each unit to obtain the frequency regulation aggregation model of the one-pipe two-unit pumped storage system. Among them, the total inertia and damping of the system are the weighted sum of the inertia and damping of each unit respectively.

3. The method for rapid optimization of control parameters of a one-pipe, two-machine pumped storage system according to claim 2 is characterized in that: The hydraulic-electromechanical coupling mathematical model includes a water diversion system model, a turbine model and a generator motor model; the water diversion system model is used to reflect the elastic deformation of the pipeline and the water hammer wave propagation effect, the turbine model includes the dynamic equation for guide vane opening adjustment, and the generator motor model is used to reflect the rotor inertia and damping characteristics.

4. The method for rapid optimization of control parameters of a one-pipe, two-machine pumped storage system according to claim 1 is characterized in that: The generation of the multidimensional parameter sample matrix includes: According to the physical characteristics of the system and the control stability requirements, the value range of each control parameter is determined to constitute the parameter feasible region; Perform equally spaced node sampling on each parameter dimension, dividing the value range of each parameter into a preset number of intervals, with each interval corresponding to a discrete value; Perform Cartesian product or tensor product operations on the discrete values of each parameter dimension to generate a multidimensional grid point set covering the entire parameter feasible domain; A four-dimensional control parameter vector is selected as the spatial sample matrix according to the multi-dimensional grid point set.

5. The method for rapid optimization of control parameters of a one-pipe, two-machine pumped storage system according to claim 4 is characterized in that: Also includes: According to the multivariable coupling constraints in actual engineering, the spatial sample matrix is screened to eliminate parameter combinations that do not meet the requirements; The frequency stability of the selected parameter combinations is checked to ensure that all parameter combinations meet the stable operation requirements of the system and obtain the checked multi-dimensional parameter sample matrix.

6. The method for rapid optimization of control parameters of a one-pipe, two-machine pumped storage system according to claim 1 is characterized in that: The stratified sampling strategy includes equal probability stratification, equal interval stratification or sensitivity weighted stratification to achieve stratified hypercube sampling in a multi-dimensional parameter space.

7. The method for rapid optimization of control parameters of a one-pipe, two-machine pumped storage system according to claim 6 is characterized in that: The sampling space and distributed traversal of the stratified sampling strategy specifically include: Select a stratified sampling strategy based on parameter characteristics; The stratified sampling strategy is determined to be equal interval stratification, dividing each parameter dimension into several non-overlapping sub-interval layers; Latin hypercube sampling or low-discrepancy sequence is used to generate samples in each subinterval layer; Combine the samples generated by each layer to obtain the final sample set; The sample set is divided into several subtask modules and assigned to multiple computing nodes for parallel simulation calculations.

8. The method for rapid optimization of control parameters of a one-pipe, two-machine pumped storage system according to claim 1 is characterized in that: The process of selecting a control parameter combination based on frequency modulation performance indicators includes: Selecting system frequency response performance indicators, the performance indicators including: absolute error of integral time, overshoot, and adjustment time; The performance indicators are weighted and combined into comprehensive indicators to comprehensively evaluate the performance of control parameters; Combined with the multi-dimensional parameter sample matrix, each parameter sample is simulated independently to obtain the corresponding performance index value; The parameter combination with the minimum comprehensive performance index and the corresponding multi-dimensional parameter sample matrix are selected to determine the optimal control parameter combination.

9. A device for rapid optimization of control parameters of a one-pipe dual-machine pumped storage system, characterized in that: include: A model building module is used to collect unit operating data and dynamic regulation characteristic parameters based on the target pumped storage system research area, and establish a frequency regulation aggregation model for the one-pipe, two-unit pumped storage system; a matrix generation module for selecting unit control parameters to be optimized from the dynamic adjustment characteristic parameters, defining the feasible region boundary and engineering constraints of the multidimensional parameters, and generating a multidimensional parameter sample matrix by constructing a tensor product based on the unit control parameter group, feasible region boundary and engineering constraints; A parallel computing module is used to determine a stratified sampling strategy for a multidimensional parameter space based on a mapping relationship between control parameters and frequency regulation performance of the regulation system, and to perform a distributed traversal of the unit control parameters using a parallel computing strategy on the multidimensional parameter sample matrix to obtain frequency response data of all samples; The parameter combination optimization module is used to establish a comprehensive evaluation system for frequency modulation performance indicators and define a multi-objective evaluation function, use the frequency response data to calculate the performance indicators, obtain the minimum value of the target indicator, and realize the rapid optimization of the control parameter combination.

10. An electronic device, characterized in that: include: at least one memory for storing a computer program; At least one processor is used to execute the program stored in the memory. When the program stored in the memory is executed, the processor is used to execute the method according to any one of claims 1 to 8.

Images