CN-121076908-B - New energy control parameter self-adaptive identification method based on double-layer particle swarm optimization

CN121076908BCN 121076908 BCN121076908 BCN 121076908BCN-121076908-B

Abstract

The invention provides a new energy control parameter self-adaptive identification method based on double-layer particle swarm optimization, and belongs to the technical field of simulation analysis of electric power systems. The method comprises the steps of integrating typical control strategies of heterogeneous new energy units, constructing a strategy database, performing verification, obtaining test data, processing and extracting features to obtain high-low pass threshold values, current initial states and test feature data, constructing an adaptability function based on the strategy database, screening optimal strategy combinations from the strategy database by using a discrete particle swarm algorithm in a global optimization layer, and continuously optimizing a parameter set of a selected strategy by combining momentum factors and learning rates in a local optimization layer. According to the invention, the characteristic value of the actual unit test data is extracted as the input of a double-layer particle swarm algorithm, the optimal strategy-parameter pair is judged and output through the weighting of the total error index, and the accuracy of the strategy-parameter pair is checked, so that the modeling efficiency and accuracy of heterogeneous new energy equipment are effectively improved.

Inventors

ZENG ZHIJIE
Lin Baogen
LIN XIAOQING
BAO GUOJUN
Wu Jialuo
ZHANG YAJIE
SHI JIYIN
LIN FANG
CHEN DAWEI
TIAN YE

Assignees

福建中试所电力调整试验有限责任公司
国网福建省电力有限公司电力科学研究院

Dates

Publication Date: 20260512
Application Date: 20251105

Claims (16)

1. The new energy control parameter self-adaptive identification method based on double-layer particle swarm optimization is characterized by comprising the following steps of: The method comprises the steps of integrating typical control strategies of four types of heterogeneous new energy units, namely wind power, photovoltaic, energy storage and SVG, and constructing a strategy database comprising voltage-current response characteristic mathematical models of various different typical control strategies, a value range of associated parameters of each control strategy and initial parameter values, wherein the voltage-current response characteristic mathematical models are in a piecewise function form; acquiring actual unit test data, and performing feature extraction on the test data to obtain test feature data; Constructing a double-layer particle swarm optimization framework comprising a global optimization layer and a local optimization layer based on a strategy database, and setting a layered fitness function; Inputting test characteristic data into a double-layer particle swarm optimization framework, and outputting an optimal strategy-parameter pair, wherein a global optimization layer adopts a discrete particle swarm algorithm, and selects an optimal strategy combination by taking a strategy error function as an fitness function, and the strategy error function is expressed as follows in a formula: ; In the formula, Is an optimal strategy combination; the method comprises the steps of primarily evaluating the number of data samples for evaluating the strategy quality; is the first Predicted current values corresponding to the individual samples; is the first The current effective values corresponding to the samples; expressed in policy database Finding a strategy that minimizes the strategy error function value among all the strategy combinations of (1) ; For policy combination The local optimization layer adopts a continuous particle swarm algorithm, optimizes the optimal strategy combination association parameters by taking the parameter error function as a fitness function, and is expressed as follows: ; In the formula, Associating optimal parameters for optimal strategy combination; Representing a unit control model, combining according to an optimal strategy Combining associated parameters for optimal policies And External input of time of day Calculating the unit Predicted current value at time; Is that A current effective value at a moment; For comprehensively evaluating the number of data samples; expressed in the associated parameter set Searching for associated parameters that minimize parameter error function values 。
2. The method for adaptively identifying new energy control parameters based on double-layer particle swarm optimization according to claim 1, wherein the strategy database comprises a strategy sub-library of SVG, wherein the strategy sub-library of SVG comprises SVG reactive power control strategies, specifically comprises voltage control current and superimposed initial reactive current components, specified reactive current control, specified reactive power control, SVG current remains unchanged, and SVG reactive current generates specified value changes and voltage control current on the basis of pre-fault and does not superimpose initial reactive current components.
3. The method for adaptively identifying the new energy control parameters based on double-layer particle swarm optimization according to claim 1, wherein the strategy database comprises a strategy sub-library oriented to wind power and photovoltaic, the strategy sub-library oriented to wind power and photovoltaic comprises a wind-light reactive power control strategy and a wind-light active power control strategy, wherein the wind-light reactive power control strategy comprises a designated reactive power control and a designated reactive power control, and the wind-light active power control strategy comprises a designated active power control, a designated active current control, a pre-fault active current control and a low-voltage limit active current control.
4. The method for adaptively identifying the new energy control parameters based on double-layer particle swarm optimization according to claim 1, wherein the strategy database comprises an energy storage strategy sub-library, the energy storage strategy sub-library comprises an energy storage reactive power control strategy, an energy storage active power control strategy in a charging and discharging mode, wherein the energy storage reactive power control strategy comprises a designated reactive power control and a designated reactive power control, and the energy storage active power control strategy in the charging and discharging mode comprises a designated active power control, a designated active current control, an active current control before failure and a low-voltage limit active current control.
5. The method for adaptively identifying new energy control parameters based on double-layer particle swarm optimization according to claim 1, further comprising: After the strategy database is constructed, the accuracy of the strategy database is checked through the simulation platform, so that the simulation data is ensured to be completely matched with the dynamic response characteristic curve corresponding to the voltage-current response characteristic mathematical model in the strategy database.
6. The method for adaptively identifying new energy control parameters based on double-layer particle swarm optimization according to claim 1, further comprising: And synchronously drawing a stepped voltage curve when the characteristic extraction is carried out on the actual unit test data, and when the current mutation is detected, taking the voltage critical value of the stepped voltage curve as a high-low penetration threshold value, and defining the piecewise function interval boundary of the voltage-current response characteristic mathematical model in the strategy database through the high-low penetration threshold value.
7. The method for adaptively identifying new energy control parameters based on double-layer particle swarm optimization according to claim 6, further comprising: When the feature extraction is carried out on the test data, through fault ride-through test data under different reactive initial values in the test data, whether the reactive current changes of the different reactive initial values in the fault period are consistent or not is analyzed, and a current superposition mode is judged according to the correlation, and the method is specific: judging whether the change of reactive current of different reactive initial values is consistent or not by comparing the change trend and the amplitude of the reactive current of different reactive initial values along with time, and calculating the correlation coefficient between the reactive current changes under different reactive initial values; When the correlation coefficient of the reactive current changes of different reactive initial values in the fault period is larger than a preset correlation coefficient threshold value, determining as an incremental mathematical model; when the correlation coefficient of the reactive current changes of different reactive initial values in the fault period is smaller than or equal to a preset correlation coefficient threshold value, determining an absolute value type mathematical model; the current superposition mode is used for controlling the optimal strategy combination selection process of the global optimization layer, and only the mathematical model of the corresponding type in the strategy database is activated.
8. The method for adaptively identifying the new energy control parameters based on double-layer particle swarm optimization according to claim 1, wherein the feature extraction of the test data adopts a fault partition method, the fault partition method is divided into a low-voltage passing region, a steady-state region and a high-voltage passing region, active current effective values, reactive current effective values and voltage effective values in each partition are extracted, and the extracted values are used as test feature data.
9. The method for adaptively identifying new energy control parameters based on double-layer particle swarm optimization according to claim 1, further comprising the steps of constructing a value range and an initial parameter value of each control strategy association parameter as an association parameter set 。
10. The method for adaptively identifying the new energy control parameters based on double-layer particle swarm optimization according to claim 1, wherein the particle update of the discrete particle swarm algorithm in the iterative process is expressed as: ; In the formula, Representation of Global speed of time; Representation of Global speed of time; Is an inertial weight; And Is a learning factor; And To at the same time A random number within; the optimal position is the particle history; is a global optimal position; is the current position of the particle; in the particle updating process, the discrete particle swarm algorithm is combined with strategy error functions to evaluate different strategy combinations in a strategy database, wherein the strategy error functions are used for measuring the quality of the current strategy combination, and the optimal strategy combination is obtained based on the strategy error functions.
11. The method for adaptively identifying the new energy control parameters based on double-layer particle swarm optimization according to claim 1, wherein the optimization process of the continuous particle swarm algorithm is specifically as follows: the local optimization layer corrects the particle speed in real time based on parameter error feedback, and the particle speed is expressed as follows: ; In the formula, Is that Local speed of time; Is that Local speed of time; is the learning rate; Is a momentum factor; combining associated parameters for optimal policies A corresponding parameter error function; combining associated parameters for the optimal strategy; The parameter error is expressed as: ; In the formula, Is that A voltage effective value at a moment; and optimizing the selected optimal strategy combination association parameters by combining the parameter errors to obtain the optimal strategy combination association optimal parameters.
12. The method for adaptively identifying the new energy control parameters based on double-layer particle swarm optimization according to claim 1, wherein the output optimal strategy-parameter pairs are specifically as follows: in the double-layer particle swarm optimization framework, the global optimization layer carries out circulation for preset times, and carries out parameter optimization circulation of the local optimization layer after carrying out strategy optimization circulation of the global optimization layer, and the total error index of each strategy-parameter pair is calculated through double-layer cooperative optimization circulation and expressed as the following formula: ; In the formula, Is the total error index; to be an optimal policy combination A corresponding policy error function; Associating optimal parameters for optimal policy combinations A corresponding parameter error function; When the total error index is smaller than the preset error threshold, the optimization is considered to be completed, the current strategy-parameter pair is output as the optimal strategy-parameter pair, and otherwise, the optimization is re-performed.
13. The method for adaptively identifying the new energy control parameters based on double-layer particle swarm optimization according to claim 1, further comprising the steps of constructing a simulation model based on an optimal strategy-parameter pair, performing feature extraction on simulation data generated by the simulation model to obtain simulation feature data, and performing error check on test feature data and simulation feature data, wherein the error check adopts a mean square error MSE, a mean absolute error MAE or a relative error.
14. The method for adaptively identifying new energy control parameters based on double-layer particle swarm optimization according to claim 1, further comprising the step of visually displaying error checking results.
15. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the new energy control parameter adaptive recognition method based on double-layer particle swarm optimization according to any of claims 1 to 14 when executing the program.
16. A computer-readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the new energy control parameter adaptive recognition method based on double-layer particle swarm optimization according to any of claims 1 to 14.

Description

New energy control parameter self-adaptive identification method based on double-layer particle swarm optimization Technical Field The invention relates to a new energy control parameter self-adaptive identification method based on double-layer particle swarm optimization, and belongs to the technical field of simulation analysis of electric power systems. Background Along with the rapid increase of the installed capacity of new energy sources such as wind power, photovoltaic and the like, the permeability of new energy source equipment in an electric power system continuously rises, and the large-scale access of the new energy source equipment to a power grid enables the accurate identification of control strategies and parameters to become the key for guaranteeing the stable operation of the system, so that the importance of new energy source simulation modeling work on the operation mode analysis of the electric power system is increasingly highlighted. In the new energy structured model building process, a normalized new energy simulation model is built, so that a convenient and quick verification method and simulation analysis tool can be provided for new energy grid-connected characteristic analysis, control strategy verification and optimization improvement, and theoretical and technical support can be provided for large-scale wind power grid-connected planning design, operation control, fault analysis and the like. The new energy structuring model modeling covers new energy unit modeling standardization technical process construction, unit key control parameter identification algorithm research and new energy unit model test verification. However, related work currently faces many challenges. The existing parameter identification method focuses on parameter optimization of a single equipment type, and the convergence efficiency is improved by improving an intelligent algorithm. For example, the Chinese patent application with publication number CN117335486A discloses a photovoltaic inverter parameter identification method and system based on a chaotic particle swarm algorithm, which are characterized in that a photovoltaic inverter double-loop control model is constructed, and a chaotic disturbance and fitness variance judgment mechanism is adopted to optimize parameter search, so that identification accuracy in a photovoltaic scene is remarkably improved. The patent model architecture is solidified in a control strategy of specific equipment (such as a photovoltaic inverter), a strategy library of diversified new energy equipment such as a wind turbine generator cannot be compatible, the optimization process only processes a mixed optimization problem through a single particle swarm frame, global exploration and local development are difficult to balance, in addition, a verification mechanism of the patent model architecture depends on static calculation of current errors, and multi-dimensional closed loop verification based on dynamic response characteristics is lacking, so that the model is insufficient in generalization in transient scenes such as actual power grid fault ride-through. In view of the foregoing, a new energy control parameter identification method supporting multi-strategy compatibility, layered collaborative optimization and dynamic feature verification is needed to break through key technical bottlenecks such as poor adaptability of heterogeneous equipment, low strategy-parameter coupling optimization efficiency, and disconnection of a model and an actual working condition, and provide theoretical support and engineering tools for stable control of a high-proportion new energy power grid. Disclosure of Invention In order to solve the problems in the prior art, the invention provides a new energy control parameter self-adaptive identification method based on double-layer particle swarm optimization. The technical scheme of the invention is as follows: In one aspect, the invention provides a new energy control parameter self-adaptive identification method based on double-layer particle swarm optimization, which comprises the following steps: The method comprises the steps of integrating typical control strategies of four types of heterogeneous new energy units, namely wind power, photovoltaic, energy storage and SVG, and constructing a strategy database comprising voltage-current response characteristic mathematical models of various different typical control strategies, a value range of associated parameters of each control strategy and initial parameter values, wherein the voltage-current response characteristic mathematical models are in a piecewise function form; acquiring actual unit test data, and performing feature extraction on the test data to obtain test feature data; Constructing a double-layer particle swarm optimization framework comprising a global optimization layer and a local optimization layer based on a strategy database, and setting a layered fitness function;