Parameter optimization method, device, equipment and product of multi-machine power system stabilizer

Abstract

The invention belongs to the technical field of multi-machine power systems, and particularly discloses a parameter optimization method and device for a multi-machine power system stabilizer, equipment and a product. According to the method, the damping ratio corresponding to the key electromechanical oscillation mode is calculated; generating a multi-dimensional target optimization function according to the damping ratio, the target deviation parameter and the comprehensive target function; and on the basis of a target weighted average optimization algorithm, searching an optimal parameter combination meeting an optimization constraint condition function according to the multi-dimensional target optimization function. Through the mode, the key electromechanical oscillation mode is identified by using the current oscillation signal of the multi-machine power system, the comprehensive target function which takes the damping ratio into account and takes multiple indexes into account is used when the multi-dimensional target optimization function is generated, then iterative solution is performed based on the target weighted average optimization algorithm, and the optimal parameter combination is searched, so that the multi-dimensional electromechanical oscillation mode is obtained. Therefore, the efficiency and the quality of parameter optimization can be effectively improved, and the robustness of the power system stabilizer during low-frequency oscillation suppression is further improved.

Description

Technical Field

[0001] This application belongs to the field of multi-machine power system technology, and more specifically, relates to parameter optimization methods, devices, equipment and products for multi-machine power system stabilizers. Background Technology

[0002] With the expansion of power system scale and the increase in the proportion of renewable energy integration, the low-frequency oscillation mechanism is becoming increasingly complex, placing higher demands on the parameter tuning of power system stabilizers (PSS), especially multi-machine power system stabilizers. Traditional parameter tuning strategies for multi-machine power system stabilizers can be phase compensation methods and local optimization methods, which suffer from problems such as long tuning time, poor adaptability, and susceptibility to getting trapped in local optima, and cannot effectively cope with the multi-condition operation and nonlinear characteristics of large interconnected power grids. The entire parameter tuning process is inseparable from parameter optimization. Currently, common methods for parameter optimization rely on intelligent algorithms such as particle swarm optimization and genetic algorithms. However, these intelligent algorithms still face shortcomings such as slow convergence speed, weak global search capability, and high difficulty in multi-machine coordinated optimization, resulting in low efficiency and quality of parameter optimization. Summary of the Invention

[0003] In view of the shortcomings of the prior art, the purpose of this application is to provide a parameter optimization method, device, equipment and product for a multi-machine power system stabilizer, which aims to solve the problems of low efficiency and quality of parameter optimization caused by slow convergence speed, weak global search capability and high difficulty of multi-machine coordinated optimization in the prior art.

[0004] To achieve the above objectives, in a first aspect, this application provides a parameter optimization method for a multi-machine power system stabilizer, comprising: The key electromechanical oscillation mode is determined based on the current oscillation signal of the multi-machine power system, and the damping ratio corresponding to the key electromechanical oscillation mode is calculated. A multi-dimensional target optimization function is generated based on the damping ratio, target deviation parameters, and comprehensive objective function. Obtain the parameter optimization variables of the multi-machine power system stabilizer, and construct the optimization constraint function based on the stabilizer parameter optimization variables; Based on the objective weighted average optimization algorithm, the optimal parameter combination that satisfies the optimization constraint function is searched according to the multi-dimensional objective optimization function.

[0005] In one embodiment, the step of determining the key electromechanical oscillation mode based on the current oscillation signal of the multi-machine power system and calculating the damping ratio corresponding to the key electromechanical oscillation mode includes: The current oscillation signal of the multi-machine power system is acquired, and the current oscillation signal is sampled multiple times according to the sampling time interval to obtain oscillation signal samples; The oscillation signal samples are analyzed according to the target signal analysis algorithm, and the key electromechanical oscillation modes of the multi-machine power system are identified based on the signal analysis results. Construct a target signal fitting model and determine the current fitting order of the target signal fitting model according to the second-order derivative algorithm; The multi-dimensional damping calculation parameters are solved based on the current fitting order and the key electromechanical oscillation mode. The damping ratio corresponding to the key electromechanical oscillation mode is calculated based on the multi-dimensional damping calculation parameters.

[0006] In one embodiment, before the step of generating a multi-dimensional target optimization function based on the damping ratio, the target deviation parameter, and the integrated target function, the method further includes: The basic optimization objective is determined based on the damping ratio and the set of key electromechanical oscillation modes; The penalty constraint function is determined based on the damping ratio, the set of key electromechanical oscillation modes, and the penalty coefficient, and the quadratic performance objective function is determined based on the basic optimization objective and the penalty constraint function. The speed integral time absolute error index is determined based on the speed deviation and time variable, and the speed peak deviation penalty index is determined based on the speed deviation, the preset speed deviation peak value and the speed penalty weighting coefficient. The target function for rotational speed is determined based on the absolute error index of rotational speed integral time and the penalty index for peak rotational speed deviation. The power integration time absolute error index is determined based on the transmission power deviation and the time variable, and the power peak deviation penalty index is determined based on the transmission power deviation, the preset power deviation peak value, and the power penalty weighting coefficient. The power objective function is determined based on the power integral time absolute error index and the power peak deviation penalty index, and the comprehensive objective function is determined based on the quadratic performance objective function, the speed objective function, and the power objective function.

[0007] In one embodiment, the step of searching for the optimal parameter combination that satisfies the optimization constraint function based on the objective weighted average optimization algorithm includes: The current optimization search space is determined based on the multi-dimensional objective optimization function; An initial candidate solution matrix is ​​generated in the current optimization search space, and the fitness value of each candidate solution in the initial candidate solution matrix is ​​calculated by the target fitness function. Obtain the population size of the initial candidate solution matrix, and determine the selection number based on the population size, the current iteration number, and the preset iteration number; The initial candidate solution matrix is ​​selected based on the fitness value and the number of selections to obtain the target candidate solution set; When the current iteration number is less than or equal to the preset iteration number, the current weighted average position is calculated based on the target candidate solution set using the target weighted average optimization algorithm. The algorithm stage balance parameters are calculated based on the current iteration number, the preset iteration number, and the multi-objective parameters. Based on the target weighted average optimization algorithm, the optimal parameter combination that satisfies the optimization constraint function is determined by the algorithm stage balance parameters and the current weighted average position.

[0008] In one embodiment, the step of optimizing the target weighted average algorithm by determining the optimal parameter combination that satisfies the optimization constraint function based on the algorithm stage balance parameters and the current weighted average position includes: When the balance parameter in the algorithm stage is greater than or equal to a preset threshold, the target weighted average optimization algorithm is determined to be in a local development stage. Based on the actual application scenario, at least one optimal position is selected from the individual optimal position and the global optimal position, and the position is updated based on the current weighted average position and the at least one optimal position; The target candidate solution is determined based on the first position update result, and the optimal parameter combination is determined when the target candidate solution satisfies the boundary conditions of the optimization constraint function.

[0009] In one embodiment, the step of optimizing the target weighted average algorithm by determining the optimal parameter combination that satisfies the optimization constraint function based on the algorithm stage balance parameters and the current weighted average position includes: When the balance parameter is less than a preset threshold in the algorithm stage, the target weighted average optimization algorithm is determined to be in the global search stage. When the random number is greater than the preset threshold, the target flight step length is calculated based on the power law exponent, the control random variable, and the adjustment random variable. The position is updated based on the target flight step size and the target's global optimal position at the current iteration number; When the random number is less than or equal to the preset threshold, the position is updated according to the minimum value of the upper and lower boundaries of the optimization constraint function, and the optimal parameter combination is determined according to the second position update result.

[0010] Secondly, this application provides a parameter optimization device for a multi-machine power system stabilizer, comprising: The determination module is used to determine the key electromechanical oscillation mode based on the current oscillation signal of the multi-machine power system, and to calculate the damping ratio corresponding to the key electromechanical oscillation mode; The generation module is used to generate a multi-dimensional target optimization function based on the damping ratio, target deviation parameters, and comprehensive objective function; A construction module is used to obtain the parameter optimization variables of the multi-machine power system stabilizer and construct the optimization constraint function based on the stabilizer parameter optimization variables; The search module is used to search for the optimal parameter combination that satisfies the optimization constraint function based on the objective weighted average optimization algorithm and the multi-dimensional objective optimization function.

[0011] Thirdly, this application provides an electronic device, comprising: at least one memory for storing a program; and at least one processor for executing the program stored in the memory, wherein when the program stored in the memory is executed, the processor is configured to execute the method described in the first aspect or any possible implementation thereof.

[0012] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when run on a processor, causes the processor to perform the method described in the first aspect or any possible implementation thereof.

[0013] Fifthly, this application provides a computer program product that, when run on a processor, causes the processor to perform the method described in the first aspect or any possible implementation thereof.

[0014] It is understood that the beneficial effects of the second to fifth aspects mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here.

[0015] Overall, the technical solutions conceived in this application have the following beneficial effects compared with the prior art: (1) After obtaining the current oscillation signal of the multi-machine power system, this application can directly identify the key electromechanical oscillation modes using the SDM-Prony algorithm, avoiding complex modeling and effectively improving the efficiency of determining the key electromechanical oscillation modes. In addition, when generating the multi-dimensional objective optimization function, this application not only considers the damping ratio, but also considers a comprehensive objective function that takes into account multiple indicators, so that the optimization results can simultaneously meet multiple performance requirements.

[0016] (2) This application uses the objective weighted average optimization algorithm when searching for the optimal parameter combination that satisfies the optimization constraint function. While maintaining the core advantages of the WAA algorithm, this objective weighted average optimization algorithm is adapted to address the performance of the WAA algorithm in multi-aircraft PSS parameter optimization problems. For example, it introduces an adaptive "exploration-development" balance strategy, an exploration mechanism based on Weibull flight, and a population optimization strategy based on quasi-oppositional learning to cope with the complex oscillation modes of multi-aircraft power systems, provide more stable and efficient global exploration, and improve the quality of the initial candidate solution matrix, making the algorithm's exploration process more efficient. Compared with traditional intelligent algorithms, it performs better in terms of convergence speed, solution accuracy, and avoiding local optima, making it extremely suitable for handling complex high-dimensional and nonlinear problems such as PSS parameter optimization. Through optimization and verification under various typical fault conditions, it is shown that the technical solution of this application can exhibit good oscillation suppression effects under different operating conditions, demonstrating excellent robustness.

[0017] In summary, this application determines key electromechanical oscillation modes based on the current oscillation signal of a multi-machine power system and calculates the damping ratio corresponding to the key electromechanical oscillation mode; it generates a multi-dimensional objective optimization function based on the damping ratio, target deviation parameters, and a comprehensive objective function; it obtains the parameter optimization variables of the multi-machine power system stabilizer and constructs an optimization constraint function based on the stabilizer parameter optimization variables; and it searches for the optimal parameter combination that satisfies the optimization constraint function based on the objective weighted average optimization algorithm. Through this method, key electromechanical oscillation modes are identified using the current oscillation signal of the multi-machine power system, and a comprehensive objective function considering the damping ratio and multiple indicators is used when generating the multi-dimensional objective optimization function. Then, iterative solutions are performed based on the objective weighted average optimization algorithm to search for the optimal parameter combination, thereby effectively improving the efficiency and quality of parameter optimization, and thus enhancing the robustness of the power system stabilizer in suppressing low-frequency oscillations. Attached Figure Description

[0018] Figure 1 This is one of the flowcharts illustrating the parameter optimization method for a multi-machine power system stabilizer provided in this application embodiment; Figure 2 This is a structural block diagram of the multi-machine power system stabilizer provided in the embodiments of this application; Figure 3 This is a comparison chart of convergence performance curves provided in the embodiments of this application; Figure 4 This is a second schematic flowchart of the parameter optimization method for a multi-machine power system stabilizer provided in the embodiments of this application; Figure 5 This is a schematic diagram of the module structure of the parameter optimization device for a multi-machine power system stabilizer provided in the embodiments of this application; Figure 6 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0020] In this article, the term "and / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. The symbol " / " in this article indicates that the related objects are in an "or" relationship; for example, A / B means A or B.

[0021] The terms "first" and "second," etc., used in the specification and claims herein are used to distinguish different objects, not to describe a specific order of objects. For example, "first response message" and "second response message," etc., are used to distinguish different response messages, not to describe a specific order of response messages.

[0022] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0023] Based on this, embodiments of this application provide a parameter optimization method for a multi-machine power system stabilizer, referring to... Figure 1 , Figure 1 This is one of the flowcharts illustrating the parameter optimization method for a multi-machine power system stabilizer provided in this application embodiment. In this embodiment, the parameter optimization method for the multi-machine power system stabilizer includes steps S10 to S40: Step S10: Determine the key electromechanical oscillation mode based on the current oscillation signal of the multi-machine power system, and calculate the damping ratio corresponding to the key electromechanical oscillation mode.

[0024] It should be noted that a multi-generator power system refers to a power system in which multiple generators are interconnected. This multi-generator power system can be an IEEE four-generator, two-area power system. Compared with conventional power systems, its structure is more complex, and the low-frequency oscillation mechanism is also more complex. Key electromechanical oscillation modes include, but are not limited to, local oscillation modes and inter-area oscillation modes. Before calculating the damping ratio, it is necessary to determine the key electromechanical oscillation modes based on the current oscillation signal of the multi-generator power system.

[0025] It should be understood that the damping ratio is a dimensionless parameter used to measure the rate of energy dissipation in key electromechanical oscillation modes, reflecting the ability of a multi-machine power system to suppress oscillations. It is typically represented by the symbol ζ. For example, ζ=0 indicates undamped oscillation with a constant amplitude; 0<ζ<1 indicates underdamped oscillation with a gradually decreasing amplitude; ζ=1 indicates critical damping, with no oscillation and the fastest return to steady state; and ζ>1 indicates overdamped oscillation with a slow, oscillating return to steady state. For example, Mode 1 (interval oscillation, =0.641 Hz, ξ=-0.0262), Mode 2 ( =1.122Hz, ξ=0.081), Mode 3 ( =1.158Hz, ξ=0.079), at this time, ξmin<0.05, indicating that parameter optimization is needed.

[0026] Further, step S10 includes: acquiring the current oscillation signal of the multi-machine power system, and sampling the current oscillation signal multiple times according to the sampling time interval to obtain oscillation signal samples; analyzing the oscillation signal samples according to the target signal analysis algorithm, and identifying the key electromechanical oscillation modes of the multi-machine power system according to the signal analysis results; constructing a target signal fitting model, and determining the current fitting order of the target signal fitting model according to the second-order derivative algorithm; solving for multi-dimensional damping calculation parameters according to the current fitting order and the key electromechanical oscillation modes; and calculating the damping ratio corresponding to the key electromechanical oscillation modes according to the multi-dimensional damping calculation parameters.

[0027] Understandably, after acquiring the current oscillation signal of the multi-machine power system, the current oscillation signal can be sampled multiple times according to the sampling time interval, wherein the number of multiple samples can be... The target signal analysis algorithm used to analyze oscillating signal samples can be the SDM-Prony algorithm, which can be described in principle as follows:

[0028] in, This represents the fitted data for the oscillating signal samples. Represents the complex amplitude coefficient. Represents the complex pole. Indicates the sampling sequence number.

[0029] It should be noted that this can be achieved by minimizing Actual sampling data The square of the error between the two can be used to approximate their substitution. Furthermore, this embodiment will construct a target signal fitting model and determine the current fitting order of the target signal fitting model using a second-order derivative algorithm. This current fitting order can be considered the optimal fitting order, and can be expressed as... And needs to meet ≥ The target signal fitting model can be the SDM-Prony model. Here, the complex amplitude coefficients... and complex poles It can be represented as:

[0030] in, Indicates amplitude. Indicates the initial phase of the signal. Indicates the attenuation factor. Indicates the signal oscillation frequency. Indicates the sampling time interval.

[0031] It should be understood that in determining the complex amplitude coefficient and complex poles After obtaining the expression, the complex magnitude coefficients can be adjusted by constructing an extended matrix. and complex poles The solution is performed to obtain multi-dimensional damping calculation parameters, which include, but are not limited to, amplitude. Initial phase of the signal Attenuation factor and signal oscillation frequency Specifically, it can be expressed as:

[0032] It should be noted that the effective rank of the extended matrix can be determined by the second-order derivative algorithm, that is, by performing singular value decomposition on the extended matrix and taking the second derivative of the obtained singular values ​​from largest to smallest. The position corresponding to the point where the derivative value is zero is the rank value.

[0033] Step S20: Generate a multi-dimensional target optimization function based on the damping ratio, target deviation parameters, and comprehensive objective function.

[0034] Understandably, the multi-dimensional objective optimization function refers to a function used to evaluate the speed and smoothness of the power recovery process. This multi-dimensional objective optimization function comprehensively considers factors such as damping ratio, speed deviation, and line transmission power deviation.

[0035] Further, before step S20, the method includes: determining a basic optimization objective based on the damping ratio and the set of key electromechanical oscillation modes; determining a penalty constraint function based on the damping ratio, the set of key electromechanical oscillation modes, and penalty coefficients, and determining a quadratic performance objective function based on the basic optimization objective and the penalty constraint function; determining a speed integral time absolute error index based on speed deviation and time variables, and determining a speed peak deviation penalty index based on the speed deviation, a preset speed deviation peak value, and a speed penalty weight coefficient; determining a speed objective function based on the speed integral time absolute error index and the speed peak deviation penalty index; determining a power integral time absolute error index based on the transmission power deviation and the time variables, and determining a power peak deviation penalty index based on the transmission power deviation, a preset power deviation peak value, and a power penalty weight coefficient; determining a power objective function based on the power integral time absolute error index and the power peak deviation penalty index, and determining a comprehensive objective function based on the quadratic performance objective function, the speed objective function, and the power objective function.

[0036] It should be understood that the basic optimization objective refers to the optimization objective that maximizes the minimum damping ratio of the multi-machine power system, and the penalty constraint function refers to the constraint function that penalizes cases where the damping ratio is lower than the target threshold. At this point, the quadratic performance objective function can be determined based on the basic optimization objective and the penalty constraint function, specifically:

[0037] in, Basic optimization goals Indicates the damping ratio. This represents the set of key electromechanical oscillation modes. This represents the penalty constraint function. Indicates the penalty coefficient. This represents the target threshold.

[0038] It is understandable that the speed objective function refers to the objective function designed based on the speed deviation, the speed integral time absolute error index refers to the index determined taking into account the speed deviation, and the speed peak deviation penalty index refers to the index used to penalize objects that exceed the preset speed deviation peak. At this point, the speed objective function can be determined based on the speed integral time absolute error index and the speed peak deviation penalty index, specifically:

[0039] in, The objective function for rotational speed is... This indicates the absolute error of the integral time of rotational speed. This indicates the penalty index for peak speed deviation. Indicates the speed deviation. Represents a time variable. This indicates the preset peak speed deviation. This represents the speed penalty weighting coefficient.

[0040] It should be noted that the power objective function can be determined in the same way as described above. The power objective function refers to the objective function designed based on the transmission power deviation. The speed integral time absolute error index refers to the index determined taking into account the transmission power deviation. The speed peak deviation penalty index refers to the index used to penalize objects that exceed the preset power deviation peak. In this case, the power objective function can be determined based on the power integral time absolute error index and the power peak deviation penalty index, specifically as follows:

[0041] in, Represents the power objective function, This indicates the absolute error of the power integration time. This indicates the penalty index for peak power deviation. Indicates power deviation. Represents a time variable. Indicates the preset power deviation peak value. This represents the power penalty weighting coefficient.

[0042] Step S30: Obtain the parameter optimization variables of the multi-machine power system stabilizer, and construct the optimization constraint function based on the stabilizer parameter optimization variables.

[0043] It should be understood that the parameter optimization variables for a multi-machine power system stabilizer include the gain coefficient and the time constant of the lead-lag element. The optimization constraint function constructed based on these stabilizer parameter optimization variables ensures parameter feasibility and can be specifically expressed as follows:

[0044] in, This represents the minimum value of the coefficient. Indicates the gain coefficient. This represents the maximum value of the coefficient. Represents the minimum value of a constant. This represents the time constant of the lead-lag process. This represents the maximum value of the constant.

[0045] It should be noted that the reference Figure 2 , Figure 2This is a block diagram of a multi-machine power system stabilizer, using the PSS1A model as an example. Specifically, it consists of six modules: a gain module, an inertial module, a torsional vibration filter module, an isolation module, and a lead-lag module. The gain module adjusts the signal amplitude, the inertial module filters high-frequency noise, the torsional vibration filter suppresses shaft torsional vibration, the isolation module removes the DC component, and the lead-lag module provides phase compensation. Through the processing and adjustment of these modules, the output signal is phase-matched to the input signal.

[0046] Step S40: Based on the objective weighted average optimization algorithm, search for the optimal parameter combination that satisfies the optimization constraint function according to the multi-dimensional objective optimization function.

[0047] It is understandable that the Weighted Average Algorithm (WAA) is a novel metaheuristic algorithm. Its core idea is to calculate the "weighted average position" of the population in each iteration and dynamically balance global search and local exploitation based on this, thereby efficiently guiding the search process to approach the global optimum. Since the WAA algorithm has shown excellent performance in handling multidimensional constrained optimization problems, it can be introduced into the parameter optimization task of multi-machine power system stabilizers to search for the optimal parameter combination. The target weighted average optimization algorithm in this embodiment can be the Improved Weighted Average Algorithm (IWAA). While maintaining the core advantages of the WAA algorithm, it is adapted to the performance of the WAA algorithm in multi-machine PSS parameter optimization problems. For example, an adaptive "exploration-exploitation" balancing strategy, an exploration mechanism based on Weibull flight, and a population optimization strategy based on quasi-adversarial learning are introduced.

[0048] It should be understood that once the optimal parameter combination satisfying the optimization constraint function is obtained, this optimal parameter combination can be loaded into the PSS1A model defined by the IEEE standard and applied to the generator excitation system to achieve low-frequency oscillation suppression. Under small disturbances, in the set of tie-line power transmission curves, the power attenuation amplitude of this embodiment is the smallest after the disturbance ends, only -0.052 pu. At the same time, under the premise of ensuring that the power recovery time meets the requirements, its overshoot is almost 0, which is much lower than the overshoot of other methods. In the speed deviation curve, the technical solution of this embodiment also shows obvious advantages. Not only are the initial impact amplitude and overshoot the smallest, but the oscillation can also be quickly smoothed out in 2 seconds. Compared with other technical solutions, the adjustment time is reduced by more than 50%. Under large disturbances, when a multi-machine power system is subjected to permanent faults and degraded structure operation, under the control of the technical solution in this embodiment, the multi-machine power system can reach near steady state the fastest. This not only achieves stable power transmission to the lower-level areas, but also ensures that the overshoot of the tie line power is almost zero. On the other hand, the technical solution in this embodiment can limit the speed deviation to below 1.8×10-3 pu. When facing transient faults, the power oscillation curve converges the fastest, which is significantly better than other methods.

[0049] It should also be noted that this embodiment will record the dominant damping ratio of the multi-machine power system under various operating conditions before and after optimization, as detailed in Table 1: Table 1:

[0050] It is important to understand that, as shown in the table above, without PSS configuration, the damping ratio of the multi-machine power system is negative in all scenarios, indicating instability. After configuring PSS, the damping of the multi-machine power system is improved and becomes positive. Due to the weak inter-system connection, the damping ratio is lower in single-line operation scenarios than in dual-line operation scenarios. Furthermore, the parameter optimization results of the technical solution (IWAA-PSS) in this embodiment are optimal in the above three scenarios. Taking the large disturbance (dual-line) scenario as an example, IWAA-PSS increases the damping ratio to 0.536, which is 37.8%, 25.8%, and 11.4% higher than the unoptimized IEEE-PSS (IEEE standard defined multi-machine power system stabilizer), PSO-PSS (multi-machine power system stabilizer with parameter optimization based on particle swarm optimization algorithm), and MFO-PSS (multi-machine power system stabilizer with parameter optimization based on moth-to-a-flame optimization algorithm), respectively. This further proves that the optimal parameter combination searched by the IWAA algorithm can most effectively enhance the stability of the system under various operating modes.

[0051] It should also be noted that the convergence performance of the algorithm is an important criterion for evaluating its optimization efficiency and global search capability. Taking the parameter optimization process under a small disturbance scenario as an example, with the population size set to 30 and the maximum number of iterations to 100, the reference... Figure 3 , Figure 3 The following is a comparison of convergence performance curves: purple represents the convergence performance curve of the objective weighted average optimization algorithm, yellow represents the convergence performance curve of the particle swarm optimization algorithm, and blue represents the convergence performance curve of the moth-to-a-flame optimization algorithm. Through the above comparison, it can be seen that the objective weighted average optimization algorithm in this embodiment has the best convergence performance, indicating that its optimization efficiency and global search capability are the strongest.

[0052] This embodiment determines the key electromechanical oscillation mode based on the current oscillation signal of the multi-machine power system and calculates the damping ratio corresponding to the key electromechanical oscillation mode. A multi-dimensional objective optimization function is generated based on the damping ratio, target deviation parameters, and a comprehensive objective function. Parameter optimization variables of the multi-machine power system stabilizer are obtained, and an optimization constraint function is constructed based on these variables. An optimal parameter combination satisfying the optimization constraint function is searched using a target weighted average optimization algorithm. By using the current oscillation signal of the multi-machine power system to identify the key electromechanical oscillation mode, and by using a comprehensive objective function that considers the damping ratio and multiple indicators when generating the multi-dimensional objective optimization function, and then iteratively solving the problem using a target weighted average optimization algorithm to search for the optimal parameter combination, the efficiency and quality of parameter optimization can be effectively improved, thereby enhancing the robustness of the power system stabilizer in suppressing low-frequency oscillations.

[0053] In one specific implementation, this application provides steps for searching the optimal combination of parameters that satisfies the optimization constraint function. Please refer to... Figure 4 , Figure 4 This is the second flowchart illustrating the parameter optimization method for a multi-machine power system stabilizer provided in this application embodiment. Step S40 includes steps S401 to S407: Step S401: Determine the current optimization search space based on the multi-dimensional objective optimization function.

[0054] It should be noted that the current optimization search space refers to the optimization search space determined based on the multi-dimensional objective optimization function, which is used to search for the optimal parameter combination that satisfies the optimization constraint function.

[0055] Step S402: Generate an initial candidate solution matrix in the current optimization search space, and calculate the fitness value of each candidate solution in the initial candidate solution matrix using the target fitness function.

[0056] Understandably, after determining the current optimization search space, random candidate solutions are randomly generated within the current optimization search space according to a preset function, which can be expressed as:

[0057] in, This represents a random candidate solution. This represents the current random number between 0 and 1. Given the first problem A lower bound value, For the given problem, the first There are upper bound values.

[0058] It should be noted that, in order to effectively improve the quality of the initial candidate solution matrix and make the algorithm's exploration process more efficient, a quasi-opposition based learning (QOBL) strategy can be introduced into the weighted average optimization algorithm. This QOBL strategy refers to a strategy that considers both "opposite solutions" and "quasi-opposite" solutions simultaneously during population initialization or iteration. In this case, the opposition vector and quasi-opposite vector are calculated separately, specifically as follows:

[0059] in, Represents the opposite vectors. Given the first problem A lower bound value, For the given problem, the first An upper bound value, This represents a quasi-opposite vector.

[0060] It should be understood that complementary vectors and quasi-complementary vectors can also serve as candidate solutions. After generating random candidate solutions, the complementary vectors and quasi-complementary vectors can be combined to form an initial candidate solution matrix, which can be specifically represented as:

[0061] in, This represents the initial candidate solution matrix. No. The candidate solution at the th... The position in the middle, This represents the total number of candidate solutions. The dimension of the problem.

[0062] It should be noted that, for each candidate solution in the initial candidate solution matrix, in order to evaluate the quality of the candidate solutions, it is also necessary to calculate the fitness value of each candidate solution using the objective fitness function. This objective fitness function can be: The function then sorts the entire candidate solution population based on the fitness value: if it is a large-to-small-best (LTB) optimization problem, the candidate solution with the larger fitness value is selected; if it is a small-to-large-best (STB) optimization problem, the candidate solution with the smaller fitness value is selected.

[0063] Step S403: Obtain the population size of the initial candidate solution matrix, and determine the selection number based on the population size, the current iteration number, and the preset iteration number.

[0064] It should be understood that before selecting candidate solutions from the initial candidate solution matrix, the number of selections needs to be determined based on the population size, the current iteration number, and the preset iteration number, specifically:

[0065] in, Indicates the number of selections. Indicates population size. Indicates the current iteration number. Indicates the preset number of iterations.

[0066] Step S404: Select the initial candidate solution matrix according to the fitness value and the selection quantity to obtain the target candidate solution set.

[0067] It should be noted that after determining the number of selections, the top candidates are selected from the initial candidate solution matrix based on their fitness values. There are candidate solutions, at which point the previous... The candidate solutions constitute the target candidate solution set.

[0068] Step S405: When the current iteration number is less than or equal to the preset iteration number, calculate the current weighted average position based on the target candidate solution set using the target weighted average optimization algorithm.

[0069] Understandably, before calculating the current weighted average position, it's necessary to determine if the current iteration number is less than or equal to the preset iteration number. If so, the current weighted average position needs to be calculated based on the target candidate solution set; otherwise, if the current iteration number is greater than the preset iteration number, it indicates that no calculation is needed. In this case, the current weighted average position can be calculated based on the target weighted average optimization algorithm and the target candidate solution set, specifically as follows:

[0070] in, This indicates the current position of the weighted average, determined from largest to smallest. This indicates the current position of the weighted average, determined from smallest to largest. This represents the sum of fitness values. Indicates the first There are 10 candidate solutions. This represents a function for calculating fitness values.

[0071] Step S406: Calculate the algorithm stage balance parameters based on the current iteration number, the preset iteration number, and the multi-objective parameters.

[0072] It should be understood that the algorithm stage balance parameters refer to the parameters used to determine the stage of the objective weighted average optimization algorithm. These multi-objective parameters include, but are not limited to, balance control constants, current random numbers, adaptive coefficients, modal diversity indices, and decay exponents. After calculating the current weighted average position, the algorithm stage balance parameters can be calculated based on the current iteration number, the preset iteration number, and the multi-objective parameters, specifically:

[0073] in, Represents the balance parameters in the algorithm phase. This represents the balance control constant, used to control the balance between the development and exploration phases. This represents the current random number between 0 and 1. Indicates the current iteration number. Indicates the preset number of iterations. Represents the adaptive coefficient. Indicates the decay index, The modal diversity index is used to quantify the damping ratio variance. The formula for determining the modal diversity index can be expressed as:

[0074] in, Indicators representing modal diversity Indicates the damping ratio. This represents the average value of the identification damping ratio; when When it is large, It will be easier to switch to a global exploration phase to break out of the current area.

[0075] Step S407: Based on the target weighted average optimization algorithm, determine the optimal parameter combination that satisfies the optimization constraint function according to the algorithm stage balance parameters and the current weighted average position.

[0076] Further, step S407 includes: when the balance parameter in the algorithm stage is greater than or equal to a preset threshold, determining that the target weighted average optimization algorithm is in a local development stage; selecting at least one optimal position from the individual optimal position and the global optimal position according to the actual application scenario, and updating the position according to the current weighted average position and the at least one optimal position; determining the target candidate solution according to the first position update result, and determining the optimal parameter combination when the target candidate solution satisfies the boundary conditions of the optimization constraint function.

[0077] It should be noted that after calculating the algorithm stage balance parameters, the stage of the algorithm stage balance parameters needs to be determined based on the comparison results between the algorithm stage balance parameters and the preset threshold. The strategy for updating the position is different for different stages. For example, when the objective weighted average optimization algorithm is in the local development stage, the search direction will be adjusted through three movement strategies to move towards the solution space with a higher probability of obtaining a new global optimum. When the objective weighted average optimization algorithm is in the global search stage, two strategies will be used to explore the far regions of the solution space to avoid getting trapped in local optima.

[0078] It should be understood that when the target weighted average optimization algorithm is in the local development stage, at this time, at least one optimal position can be selected from the individual optimal position and the global optimal position according to the actual application scenario, and the position can be updated according to the current weighted average position and at least one optimal position. The position combination for updating the position at this time can be: (1) the current weighted average position, the individual optimal position and the global optimal position; (2) the current weighted average position and the individual optimal position; (3) the current weighted average position and the global optimal position.

[0079] On the one hand, the method of updating the position based on the position combination (1) can be expressed as:

[0080] in, This indicates that the result is updated at the first position. These represent weight constants, used to adjust the optimal position of an individual. Global optimal position Compared with the current weighted average position Balancing the search process.

[0081] On the other hand, the method of updating the position based on the position combination (2) can be expressed as:

[0082] in, This indicates that the result is updated at the first position. denoted as weight constants respectively. Compared with the above position combination (1), position combination (2) tends to improve convergence speed and accuracy by guiding the individual's optimal position to move towards the current weighted average position.

[0083] On the other hand, the method of updating the position based on the position combination (3) can be expressed as:

[0084] in, This indicates that the result is updated at the first position. denoted as weight constants respectively. Since the solution space of position combination (3) is narrower than that of position combination (1) and (2), the convergence speed and accuracy of position combination (3) are extremely high.

[0085] Further, step S407 includes: when the balance parameter in the algorithm stage is less than a preset threshold, determining that the target weighted average optimization algorithm is in the global search stage; when the random number is greater than the preset threshold, calculating the target flight step size based on the power law exponent, control random variables, and adjustment random variables; updating the position based on the target flight step size and the target global optimal position at the current iteration number; when the random number is less than or equal to the preset threshold, updating the position based on the minimum value of the upper and lower boundaries of the optimization constraint function, and determining the optimal parameter combination based on the second position update result.

[0086] It should be noted that the preset threshold can be set to 0.5. When the target weighted average optimization algorithm is determined to be in the global search stage based on the algorithm stage balance parameters, it is necessary to further compare the random number with the preset threshold and update the position according to the comparison result. Specifically: On the one hand, when the random number exceeds a preset threshold, it indicates that the search range needs to be narrowed. At this time, the Weibull flight strategy can be used. Compared with the Lévy flight strategy, which has uneven exploration, the Weibull flight strategy has both long-tail characteristics and adjustability in the step size distribution of random walks, and can provide more stable and efficient global exploration. The calculation method of its target flight step size can be expressed as:

[0087] in, Indicates the target flight step length. Represents a random value generated by the Weibull distribution. This represents a random vector within the range [0,1]. Indicates the dimensions of the problem. This represents the direction control function. Let represent the distance norm. Based on the target flight step length, we have:

[0088] in, This represents the target global optimal position at the current iteration number, which can be the [number]th iteration. The location of the global optimal solution. This represents the position update result after narrowing the search range, and can be used for iteration. Next time The solution of the first... One location, Indicates the target flight step length.

[0089] On the other hand, when the random number is less than or equal to a preset threshold, it is determined that the solution is trapped in a local optimum. In this case, it is necessary to expand the search range, specifically:

[0090] in, This indicates the location update results when the search scope is expanded. This represents the minimum value of the upper bound of each dimension. This represents the minimum value of the lower bound of each dimension.

[0091] It should be noted that, in this embodiment, after updating the location, the system will continue to determine the next step. candidate solutions If the boundary conditions of the optimization constraint function are met, then the fitness value needs to be calculated and the individual's optimal position updated. Global optimal position The current iteration number at this point = +1, and return to the execution judgment. Is it less than or equal to the preset number of iterations? The steps.

[0092] This embodiment determines the current optimization search space based on the multi-dimensional objective optimization function; generates an initial candidate solution matrix in the current optimization search space, and calculates the fitness value of each candidate solution in the initial candidate solution matrix using the objective fitness function; obtains the population size of the initial candidate solution matrix, and determines the selection quantity based on the population size, the current iteration number, and the preset iteration number; selects from the initial candidate solution matrix based on the fitness value and the selection quantity to obtain the objective candidate solution set; when the current iteration number is less than or equal to the preset iteration number, calculates the current weighted average position based on the objective weighted average optimization algorithm and the objective candidate solution set; calculates the algorithm stage balance parameters based on the current iteration number, the preset iteration number, and the multi-objective parameters; and, based on the objective weighted average optimization algorithm, determines the optimal parameter combination that satisfies the optimization constraint function based on the algorithm stage balance parameters and the current weighted average position. By using the above method, after calculating the number of selected solutions, candidate solutions are selected from the initial candidate solution matrix in combination with the fitness value. Then, it is determined whether the current iteration number is less than or equal to the preset iteration number. If so, the optimal parameter combination is determined based on the target weighted average optimization algorithm, according to the algorithm stage balance parameter and the current weighted average position, thereby effectively improving the accuracy of obtaining the optimal parameter combination.

[0093] The parameter optimization device for a multi-machine power system stabilizer provided in this application is described below. The parameter optimization device described below corresponds to the parameter optimization method described above. Please refer to... Figure 5 , Figure 5 This is a schematic diagram of the module structure of the parameter optimization device for a multi-machine power system stabilizer provided in this application embodiment, including: The determination module T10 is used to determine the key electromechanical oscillation mode based on the current oscillation signal of the multi-machine power system, and to calculate the damping ratio corresponding to the key electromechanical oscillation mode.

[0094] The generation module T20 is used to generate a multi-dimensional target optimization function based on the damping ratio, target deviation parameters, and comprehensive objective function.

[0095] The construction module T30 is used to obtain the parameter optimization variables of the multi-machine power system stabilizer and construct the optimization constraint function based on the stabilizer parameter optimization variables.

[0096] The search module T40 is used to search for the optimal parameter combination that satisfies the optimization constraint function based on the objective weighted average optimization algorithm and the multi-dimensional objective optimization function.

[0097] This embodiment determines the key electromechanical oscillation mode based on the current oscillation signal of the multi-machine power system and calculates the damping ratio corresponding to the key electromechanical oscillation mode. A multi-dimensional objective optimization function is generated based on the damping ratio, target deviation parameters, and a comprehensive objective function. Parameter optimization variables of the multi-machine power system stabilizer are obtained, and an optimization constraint function is constructed based on these variables. An optimal parameter combination satisfying the optimization constraint function is searched using a target weighted average optimization algorithm. By using the current oscillation signal of the multi-machine power system to identify the key electromechanical oscillation mode, and by using a comprehensive objective function that considers the damping ratio and multiple indicators when generating the multi-dimensional objective optimization function, and then iteratively solving the problem using a target weighted average optimization algorithm to search for the optimal parameter combination, the efficiency and quality of parameter optimization can be effectively improved, thereby enhancing the robustness of the power system stabilizer in suppressing low-frequency oscillations.

[0098] It is understood that the detailed functional implementation of each of the above modules can be found in the description of the aforementioned method embodiments, and will not be repeated here.

[0099] It should be understood that the above-described device is used to execute the methods in the above embodiments. The implementation principle and technical effect of the corresponding program modules in the device are similar to those described in the above methods. The working process of the device can be referred to the corresponding process in the above methods, and will not be repeated here.

[0100] Based on the methods in the above embodiments, this application provides an electronic device, please refer to... Figure 6 , Figure 6 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application.

[0101] It should be noted that the system may include: a processor 10, a communications interface 20, a memory 30, and a communication bus 40. The processor 10, communications interface 20, and memory 30 communicate with each other via the communication bus 40. The processor 10 can invoke logical instructions stored in the memory 30 to execute the methods described in the above embodiments.

[0102] Furthermore, the logical instructions in the aforementioned memory 30 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application.

[0103] Based on the methods in the above embodiments, this application provides a computer-readable storage medium storing a computer program that, when run on a processor, causes the processor to execute the methods in the above embodiments.

[0104] Based on the methods in the above embodiments, this application provides a computer program product that, when run on a processor, causes the processor to execute the methods in the above embodiments.

[0105] It is understood that the processor in the embodiments of this application can be a central processing unit, or other general-purpose processors, digital signal processors, application-specific integrated circuits, field-programmable gate arrays, or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. A general-purpose processor can be a microprocessor or any conventional processor.

[0106] The method steps in this application embodiment can be implemented in hardware or by a processor executing software instructions. The software instructions can consist of corresponding software modules, which can be stored in random access memory, flash memory, read-only memory, programmable read-only memory, erasable programmable read-only memory, electrically erasable programmable read-only memory, registers, hard disks, portable hard disks, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor.

[0107] It is understood that the various numerical designations used in the embodiments of this application are merely for descriptive convenience and are not intended to limit the scope of the embodiments of this application. Those skilled in the art will readily understand that the above descriptions are merely preferred embodiments of this application and are not intended to limit this application. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A parameter optimization method for a multi-machine power system stabilizer, characterized in that, include: The key electromechanical oscillation mode is determined based on the current oscillation signal of the multi-machine power system, and the damping ratio corresponding to the key electromechanical oscillation mode is calculated. A multi-dimensional target optimization function is generated based on the damping ratio, target deviation parameters, and comprehensive objective function. Obtain the parameter optimization variables of the multi-machine power system stabilizer, and construct the optimization constraint function based on the stabilizer parameter optimization variables; Based on the objective weighted average optimization algorithm, the optimal parameter combination that satisfies the optimization constraint function is searched according to the multi-dimensional objective optimization function.

2. The method as described in claim 1, characterized in that, The step of determining the key electromechanical oscillation mode based on the current oscillation signal of the multi-machine power system and calculating the damping ratio corresponding to the key electromechanical oscillation mode includes: The current oscillation signal of the multi-machine power system is acquired, and the current oscillation signal is sampled multiple times according to the sampling time interval to obtain oscillation signal samples; The oscillation signal samples are analyzed according to the target signal analysis algorithm, and the key electromechanical oscillation modes of the multi-machine power system are identified based on the signal analysis results. Construct a target signal fitting model and determine the current fitting order of the target signal fitting model according to the second-order derivative algorithm; The multi-dimensional damping calculation parameters are solved based on the current fitting order and the key electromechanical oscillation mode. The damping ratio corresponding to the key electromechanical oscillation mode is calculated based on the multi-dimensional damping calculation parameters.

3. The method as described in claim 1, characterized in that, Before the step of generating a multi-dimensional target optimization function based on the damping ratio, target deviation parameters, and integrated target function, the method further includes: The basic optimization objective is determined based on the damping ratio and the set of key electromechanical oscillation modes; The penalty constraint function is determined based on the damping ratio, the set of key electromechanical oscillation modes, and the penalty coefficient, and the quadratic performance objective function is determined based on the basic optimization objective and the penalty constraint function. The speed integral time absolute error index is determined based on the speed deviation and time variable, and the speed peak deviation penalty index is determined based on the speed deviation, the preset speed deviation peak value and the speed penalty weighting coefficient. The target function for rotational speed is determined based on the absolute error index of rotational speed integral time and the penalty index for peak rotational speed deviation. The power integration time absolute error index is determined based on the transmission power deviation and the time variable, and the power peak deviation penalty index is determined based on the transmission power deviation, the preset power deviation peak value, and the power penalty weighting coefficient. The power objective function is determined based on the power integral time absolute error index and the power peak deviation penalty index, and the comprehensive objective function is determined based on the quadratic performance objective function, the speed objective function, and the power objective function.

4. The method as described in claim 1, characterized in that, The step of searching for the optimal parameter combination that satisfies the optimization constraint function based on the objective weighted average optimization algorithm includes: The current optimization search space is determined based on the multi-dimensional objective optimization function; An initial candidate solution matrix is ​​generated in the current optimization search space, and the fitness value of each candidate solution in the initial candidate solution matrix is ​​calculated by the target fitness function. Obtain the population size of the initial candidate solution matrix, and determine the selection number based on the population size, the current iteration number, and the preset iteration number; The initial candidate solution matrix is ​​selected based on the fitness value and the number of selections to obtain the target candidate solution set; When the current iteration number is less than or equal to the preset iteration number, the current weighted average position is calculated based on the target candidate solution set using the target weighted average optimization algorithm. The algorithm stage balance parameters are calculated based on the current iteration number, the preset iteration number, and the multi-objective parameters. Based on the target weighted average optimization algorithm, the optimal parameter combination that satisfies the optimization constraint function is determined by the algorithm stage balance parameters and the current weighted average position.

5. The method as described in claim 4, characterized in that, The step of the target-based weighted average optimization algorithm, which involves the optimal combination of parameters satisfying the optimization constraint function based on the algorithm's stage balance parameters and the current weighted average position, includes: When the balance parameter in the algorithm stage is greater than or equal to a preset threshold, the target weighted average optimization algorithm is determined to be in a local development stage. Based on the actual application scenario, at least one optimal position is selected from the individual optimal position and the global optimal position, and the position is updated based on the current weighted average position and the at least one optimal position; The target candidate solution is determined based on the first position update result, and the optimal parameter combination is determined when the target candidate solution satisfies the boundary conditions of the optimization constraint function.

6. The method as described in claim 4, characterized in that, The step of the target-based weighted average optimization algorithm, which involves the optimal combination of parameters satisfying the optimization constraint function based on the algorithm's stage balance parameters and the current weighted average position, includes: When the balance parameter is less than a preset threshold in the algorithm stage, the target weighted average optimization algorithm is determined to be in the global search stage. When the random number is greater than the preset threshold, the target flight step length is calculated based on the power law exponent, the control random variable, and the adjustment random variable. The position is updated based on the target flight step size and the target's global optimal position at the current iteration number; When the random number is less than or equal to the preset threshold, the position is updated according to the minimum value of the upper and lower boundaries of the optimization constraint function, and the optimal parameter combination is determined according to the second position update result.

7. A parameter optimization device for a multi-machine power system stabilizer, characterized in that, include: The determination module is used to determine the key electromechanical oscillation mode based on the current oscillation signal of the multi-machine power system, and to calculate the damping ratio corresponding to the key electromechanical oscillation mode; The generation module is used to generate a multi-dimensional target optimization function based on the damping ratio, target deviation parameters, and comprehensive objective function; A construction module is used to obtain the parameter optimization variables of the multi-machine power system stabilizer and construct the optimization constraint function based on the stabilizer parameter optimization variables; The search module is used to search for the optimal parameter combination that satisfies the optimization constraint function based on the objective weighted average optimization algorithm and the multi-dimensional objective optimization function.

8. An electronic device, characterized in that, include: At least one memory for storing computer programs; At least one processor is configured to execute a program stored in the memory, wherein when the program stored in the memory is executed, the processor is configured to perform the method as described in any one of claims 1-6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is run on the processor, it causes the processor to perform the method as described in any one of claims 1-6.

10. A computer program product, characterized in that, When the computer program product is run on a processor, the processor causes the processor to perform the method as described in any one of claims 1-6.

Images