SVG parameter dynamic setting method and device, electronic equipment and storage medium

Abstract

The invention discloses an SVG parameter dynamic setting method and device, electronic equipment and a storage medium, which are used for solving the technical problems of incomplete consideration, low optimization efficiency, easiness in falling into a local optimal solution and poor expansibility when an objective function is constructed in the prior art. The method comprises the following steps: constructing a control model of the SVG, and determining a to-be-optimized parameter set of the control model; voltage response of the control model during transient simulation is considered, and a fitness function based on voltage recovery segmented penalty is constructed; based on the fitness function, performing parameter optimization on the to-be-optimized parameter set through particle swarm optimization to obtain an optimal parameter set; and taking the optimal parameter set as a control parameter set of the SVG in practical application.

Description

Technical Field

[0001] This invention relates to the field of power system automation and intelligent control technology, and in particular to a method, apparatus, electronic device and storage medium for dynamic tuning of SVG parameters. Background Technology

[0002] With the large-scale integration of new energy sources into the power system, the traditional "source-grid-load" structure is gradually evolving into a new system form with deep coupling of "source-grid-load-storage". Especially against the backdrop of the rapid increase in the proportion of renewable energy such as wind power and photovoltaics, increasingly fluctuating electricity load, and frequent changes in system topology, the voltage stability of the power system faces unprecedented challenges. When disturbances occur in the power grid (such as short circuits, sudden changes in power sources or loads), whether the voltage at key nodes of the system can recover quickly and meet grid specifications has become an important indicator for measuring the system's safety margin and dynamic stability capability.

[0003] Static Var Generators (SVG), as high-speed, continuously adjustable dynamic reactive power compensation devices, have been widely used in ultra-high voltage transmission systems, important load centers, and renewable energy access points. SVG mitigates voltage dips and enhances transient voltage support capabilities by rapidly adjusting the reactive current injection, making it a key control method in the voltage control system of new power systems. However, the actual adjustment effect of SVG is significantly affected by the internal parameter settings of its controller, which directly influence the dynamic response speed and amplitude of the SVG during low-voltage disturbances.

[0004] Currently, in engineering applications, these parameters mostly rely on manufacturer default settings or are set by commissioning personnel based on experience, making it difficult to meet the optimized response requirements under different disturbance scenarios and system structures. In practice, if the control parameters are set improperly, the SVG response process may exhibit delays, overshoot, undershoot, and other phenomena. In severe cases, it may even cause the voltage response to fail to meet the relevant requirements for Low Voltage Ride Through (LVRT), affecting the stable operation of the equipment and thus threatening the overall system safety.

[0005] To overcome the subjectivity and limitations of manual parameter tuning, some studies have introduced intelligent optimization algorithms to automatically adjust SVG control parameters. However, these methods still have the following drawbacks: the constructed objective function often focuses on the controller output or error integral, lacking direct constraints on actual grid compliance indicators; the optimization efficiency is low, and it is prone to getting trapped in local optima in the multidimensional nonlinear search space; the morphological characteristics and compliance checks of the simulated voltage curve during disturbance periods are ignored, failing to guarantee the physical feasibility and engineering applicability of the final output; and there is a lack of scalability to adapt to different SVG control structures and diverse system operating conditions.

[0006] Therefore, there is an urgent need for an intelligent optimization method that is oriented towards the disturbance response process, takes voltage compliance as the optimization goal, and has global search capabilities to dynamically tune and correct the SVG controller parameters, so as to improve its robustness and response accuracy to low voltage disturbances. Summary of the Invention

[0007] This invention provides a method, apparatus, electronic device, and storage medium for dynamic tuning of SVG parameters, which solves or partially solves the technical problems of current related technologies, such as insufficient consideration when constructing objective functions, low optimization efficiency, easy getting trapped in local optima, and poor scalability.

[0008] This invention provides a method for dynamically tuning SVG parameters, the method comprising:

[0009] Construct a control model for the SVG and determine the set of parameters to be optimized for the control model;

[0010] Considering the voltage response of the control model during transient simulation, a fitness function based on voltage recovery piecewise penalty is constructed.

[0011] Based on the fitness function, the optimal parameter set is obtained by optimizing the parameter set to be optimized through particle swarm optimization.

[0012] The optimal parameter set is used as the control parameter set for the SVG in practical applications.

[0013] Optionally, the step of considering the voltage response of the control model during transient simulation and constructing a fitness function based on voltage recovery piecewise penalty includes:

[0014] Considering the voltage response of the control model during transient simulation, the voltage response is divided into four key time periods;

[0015] A minimum voltage threshold is set for each of the key time periods, and a segmented penalty weight is set for each of the minimum voltage thresholds.

[0016] Based on the minimum voltage threshold and segmented penalty weights for each critical time period, and combined with the voltage response, a fitness function based on voltage recovery segmented penalty is constructed.

[0017] Optionally, the minimum voltage threshold and segmented penalty weight for each of the key time periods are set as follows:

[0018] First period Corresponding to the first minimum voltage threshold First segment penalty weight ;

[0019] Second period Corresponding to the second lowest voltage threshold Second segment penalty weight ;

[0020] Third period Corresponding to the third lowest voltage threshold The third segment penalty weight ;

[0021] Fourth period Corresponding to the fourth lowest voltage threshold The fourth segment penalty weight ;

[0022] The time unit for each period is a cycle (Zodiac); This refers to the time required to clear the fault. Indicates the time.

[0023] Optionally, the step of optimizing the parameter set to be optimized through particle swarm optimization based on the fitness function to obtain the optimal parameter set includes:

[0024] The set of parameters to be optimized is updated iteratively through particle swarm optimization. In each iteration:

[0025] A transient simulation was performed on the control model, and the voltage time series of each node was recorded.

[0026] Based on the voltage time series of each node, a voltage response curve is generated;

[0027] Based on the fitness function, the voltage response curve and the low-breakdown reference curve are time-aligned and error is calculated to complete the fitness evaluation and generate multiple fitness values ​​for the current iteration round.

[0028] After the current iteration ends, save the simulation results when the current fitness value is optimal;

[0029] Through continuous iterative optimization, when the maximum number of iterations is reached, or when the preset low-pass requirement is met, the final optimal parameter set is output.

[0030] Optionally, the simulation results include the voltage response curve, low-breakdown reference curve, and optimal fitness value generated when the control model performs transient simulation; the method further includes:

[0031] Output and save the voltage response curve, low-breakdown reference curve, and optimal fitness value after each iteration in a structured table format.

[0032] Optionally, constructing the control model of the SVG and determining the set of parameters to be optimized for the control model includes:

[0033] A control model for the SVG is constructed, and several key control parameters in the control model are selected as variables to be optimized.

[0034] Set upper and lower bounds for each of the variables to be optimized, and construct a parameter search space;

[0035] The various variables to be optimized located in the parameter search space are integrated as the set of parameters to be optimized for the control model.

[0036] Optionally, the variables to be optimized in the set of parameters to be optimized include the characteristic time constant of the filtering and measurement circuit, the lead time constant of the first feedforward control, the lag time constant of the first feedforward control, the lead time constant of the second feedforward control, the lag time constant of the second feedforward control, the time constant of the proportional element, the gain of the proportional element, the gain of the integral element, and the slope of the voltage-current characteristic curve.

[0037] The present invention also provides an SVG parameter dynamic tuning device, comprising:

[0038] The parameter set determination unit is used to construct the control model of the SVG and determine the parameter set to be optimized of the control model;

[0039] The fitness function construction unit is used to consider the voltage response of the control model during transient simulation and construct a fitness function based on voltage recovery piecewise penalty.

[0040] The parameter optimization unit is used to optimize the parameter set to be optimized based on the fitness function and through particle swarm optimization to obtain the optimal parameter set.

[0041] The parameter set application unit is used to use the optimal parameter set as the control parameter set of the SVG in actual application.

[0042] The present invention also provides an electronic device, the device comprising a processor and a memory:

[0043] The memory is used to store program code and transmit the program code to the processor;

[0044] The processor is configured to execute the SVG parameter dynamic tuning method as described above, according to the instructions in the program code.

[0045] The present invention also provides a computer-readable storage medium for storing program code for executing the SVG parameter dynamic tuning method as described in any of the preceding claims.

[0046] As can be seen from the above technical solutions, the present invention has the following advantages:

[0047] A method for dynamic parameter tuning of SVG based on particle swarm optimization is presented. First, a control model of the SVG is constructed, and the set of parameters to be optimized is determined. Then, considering the voltage response of the control model during transient simulation, a fitness function based on voltage recovery piecewise penalty is constructed. Next, based on the fitness function, particle swarm optimization is used to optimize the parameter set to obtain the optimal parameter set. Finally, the optimal parameter set is used as the control parameter set for the SVG in practical applications. Thus, by establishing a fitness function with voltage transient response performance as the objective and using the particle swarm optimization algorithm to optimize the key control parameters of the SVG, automatic parameter setting and dynamic adjustment are achieved. This method has advantages such as strong generalization ability, high adaptability, and high adjustment accuracy, effectively improving the dynamic adjustment capability of the SVG in complex power grids. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 A flowchart illustrating the steps of a method for dynamically tuning SVG parameters;

[0050] Figure 2 This is a schematic diagram of a process for dynamically tuning SVG parameters through particle swarm optimization.

[0051] Figure 3 This is a schematic diagram of SVG control parameter optimization;

[0052] Figure 4 This is a schematic diagram comparing a voltage response curve with a low-breakdown reference curve.

[0053] Figure 5This is a structural block diagram of an SVG parameter dynamic tuning device. Detailed Implementation

[0054] This invention provides a method, apparatus, electronic device, and storage medium for dynamically tuning SVG parameters, which solves or partially solves the technical problems of current related technologies, such as insufficient consideration when constructing objective functions, low optimization efficiency, easy getting trapped in local optima, and poor scalability.

[0055] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0056] As an example, in engineering applications, the internal control parameters of SVG (Static Var Generator) mostly rely on manufacturer default settings or are set by commissioning personnel based on experience, making it difficult to meet the optimized response requirements under different disturbance scenarios and system structures. In reality, if the control parameters are set improperly, the SVG response process may exhibit delays, overshoot, undershoot, or other phenomena. In severe cases, it may even cause the voltage response to fail to meet the low-voltage drive-through specifications, affecting the stable operation of the equipment and thus threatening the overall system safety.

[0057] To overcome the subjectivity and limitations of manual parameter tuning, some studies have introduced intelligent optimization algorithms to automatically adjust SVG control parameters. However, these methods still have the following drawbacks: the constructed objective function often focuses on the controller output or error integral, lacking direct constraints on actual grid compliance indicators; the optimization efficiency is low, and it is prone to getting trapped in local optima in the multidimensional nonlinear search space; the morphological characteristics and compliance checks of the simulated voltage curve during disturbance periods are ignored, failing to guarantee the physical feasibility and engineering applicability of the final output; and there is a lack of scalability to adapt to different SVG control structures and diverse system operating conditions.

[0058] Therefore, there is an urgent need for an intelligent optimization method that is oriented towards the disturbance response process, takes voltage compliance as the optimization goal, and has global search capabilities to dynamically tune and correct the SVG controller parameters, so as to improve its robustness and response accuracy to low voltage disturbances.

[0059] Based on this, one of the core inventive points of this invention is: to improve the voltage support capability and dynamic response performance of SVG (Static Var Generator) under transient conditions such as fault disturbances in power systems, a dynamic tuning method for SVG control parameters based on the Particle Swarm Optimization (PSO) algorithm is proposed. The method includes: constructing an SVG system model; setting an objective function and quantifying voltage response characteristics; using the PSO algorithm to iteratively optimize within the parameter space; and applying the optimization results to the SVG controller to achieve stable voltage control of the power grid under different operating conditions. The technical solution provided by this invention establishes a fitness function with voltage transient response performance as the objective, and uses the PSO algorithm to optimize and search for key control parameters of the SVG, achieving automatic parameter setting and dynamic adjustment. It has advantages such as strong generalization ability, high adaptability, and high adjustment accuracy, and can effectively improve the dynamic adjustment capability of SVG in complex power grids.

[0060] Reference Figure 1 The diagram illustrates a flowchart of a dynamic tuning method for SVG parameters provided by an embodiment of the present invention, which specifically includes the following steps:

[0061] Step 101: Construct the control model of the SVG and determine the set of parameters to be optimized for the control model;

[0062] In some embodiments, a control model of the SVG is constructed, and the set of parameters to be optimized in the control model is determined. Specifically, this can be done by: constructing a control model of the SVG and selecting multiple key control parameters in the control model as variables to be optimized; setting upper and lower bounds for each variable to be optimized to construct a parameter search space; and integrating the variables to be optimized located in the parameter search space as the set of parameters to be optimized in the control model.

[0063] Specifically, in this embodiment of the invention, nine key parameters in the SVG control model are selected as optimization variables to construct a set of parameters to be optimized. These parameters cover multiple control paths, including feedforward lag elements, response delays, and proportional-integral controllers.

[0064] Specifically, the variables to be optimized in the set of parameters to be optimized may include the characteristic time constants of the filtering and measurement circuits. The lead time constant of primary feedforward control Lag time constant of primary feedforward control The lead time constant of secondary feedforward control Lag time constant of secondary feedforward control Proportional link time constant Proportional gain Integral gain The slope of the voltage-current characteristic curve (differential control factor) .

[0065] The above key parameters are presented in vector form as follows:

[0066]

[0067] To improve the controllability and convergence of the optimization process, reasonable upper and lower bounds are set for each key parameter, forming a search space for the parameters. For example, it is possible to set... The value range is [0.01, 0.4]. The range is [0.01, 20]. The range is [0.001, 0.2]. The range is [0.001, 2], etc. By setting the search space, it is possible to fully cover all possible control and regulation solutions while ensuring the practicality of the project.

[0068] Step 102: Considering the voltage response of the control model during transient simulation, construct a fitness function based on voltage recovery piecewise penalty;

[0069] To accurately evaluate the effect of each set of control parameters on the SVG disturbance response, this invention constructs a penalized fitness function based on a voltage recovery segmentation standard to measure the dynamic recovery performance of the SVG node voltage after fault clearance. This function divides the voltage response into four critical time periods, sets a minimum voltage threshold for each critical time period, and applies segmented penalties based on the degree to which the response voltage falls below the voltage threshold.

[0070] In the specific implementation, considering the voltage response of the control model during transient simulation, the implementation process of the fitness function based on voltage recovery piecewise penalty is constructed, including the following sub-steps S1 to S3:

[0071] Step S1: Consider the voltage response of the control model during transient simulation, and divide the voltage response into four key time periods;

[0072] Step S2: Set minimum voltage thresholds for each key time period, and set segmented penalty weights for each minimum voltage threshold.

[0073] set up For the first Each node at time... The voltage response; This is the time required to clear the fault.

[0074] The minimum voltage threshold and segmented penalty weights for each critical time period are set as follows:

[0075] First period Corresponding to the first minimum voltage threshold First segment penalty weight ;

[0076] Second period Corresponding to the second lowest voltage threshold Second segment penalty weight ;

[0077] Third period Corresponding to the third lowest voltage threshold The third segment penalty weight ;

[0078] Fourth period Corresponding to the fourth lowest voltage threshold The fourth segment penalty weight ;

[0079] The time unit for each period is a cycle (Zodiac); This refers to the time required to clear the fault. Indicates the time.

[0080] In this invention, "cycle" refers to a time measurement method using the AC cycle of a power system as the unit. For example, in a 50Hz system, 1 cycle is equivalent to 20ms. That is, the first time period is: fault clearing time < ≤ Fault clearing time + 16.5 cycles. If the power grid frequency is 50Hz, then 16.5 cycles = 16.5 × 1 / 50 = 0.33 seconds. All time periods (16.5, 25, 75) are measured in cycles.

[0081] It should be noted that the time period values ​​for the key time periods mentioned above are based on typical time period divisions defined by the low-voltage characteristic curve in relevant industry standards. These figures are not the only selectable values ​​and can be appropriately adjusted within a reasonable range in practical applications. Those skilled in the art can set other durations according to the specific simulation platform or application scenario. However, the following setting prerequisites must be followed when setting: the segmentation settings should cover the entire voltage transient process (such as 0~1.5 seconds after fault clearance); the segment intervals should not overlap; the threshold and time period should be reasonably matched to reflect the gradual recovery trend.

[0082] The minimum voltage thresholds for each critical period are set based on the standard requirements for transient voltage recovery in the power grid. That is, on the one hand, the voltage cannot immediately recover to its rated value after the short circuit is cleared; it should exhibit a gradual increase. On the other hand, different minimum voltage requirements are set for different stages. For example, the minimum voltage requirement is only ≥0.7 pu immediately after the fault is cleared (the first period); while in the later recovery stages, the voltage is required to be ≥0.9 pu or 0.95 pu. This segmented voltage recovery design of the present invention reflects the principle of combining the grid's tolerance and constraints on the gradual voltage recovery of equipment after a fault.

[0083] Furthermore, the penalty weights corresponding to the four voltage recovery stages mentioned above can be manually set based on actual needs. By weighting the voltage response according to the system's level of attention at different stages, the control capability during critical time periods is strengthened. Specifically, those skilled in the art can directly set the weight coefficient for each critical time period in the system configuration (e.g., 1.0 for the first time period, 1.5 for the second, 2.0 for the third, and 2.5 for the fourth). The system weights and accumulates the voltage deviations for each stage in the fitness function according to the set weights, forming the final loss value. This allows for flexible adjustment based on system operating characteristics, control objectives, or actual scheduling requirements. For example, the weight of the initial response stage can be increased to emphasize the initial fault recovery speed, or the weight of the later stages can be increased to improve voltage stability. This setting does not affect the optimization algorithm structure and has good adjustability and adaptability.

[0084] Step S3: Based on the minimum voltage threshold and segmented penalty weights for each key time period, and combined with the voltage response, construct a fitness function based on voltage recovery segmented penalty.

[0085] The fitness function constructed in this step is shown below:

[0086]

[0087] in, For the first The minimum voltage threshold of the segment; This represents the time interval corresponding to this segment.

[0088] The physical meaning of the fitness function described above is to impose a quantitative penalty on all portions of the voltage below the undervoltage threshold. The larger the deviation, the longer the duration, and the more nodes involved, the higher the penalty. In this embodiment of the invention, the voltage thresholds for each time period show an increasing trend, which is set with reference to the undervoltage threshold. This setting reflects the reasonable requirement for the gradual voltage recovery process after fault clearance. This type of setting can both ensure that the equipment is not damaged by instantaneous low voltage and impose clear phased requirements on the system response performance.

[0089] To further enhance the adjustment flexibility at different stages of the voltage recovery process, this invention also sets segmented penalty weight parameters. Those skilled in the art can set penalty weight values ​​for different time periods (e.g., based on the system's focus) according to their specific needs. =1.0, =1.5, =2.0, =2.5). The weight values ​​participate as weighting factors in the fitness function to accumulate voltage deviations for each segment, forming the overall evaluation result. By adjusting the weights, the voltage control capability of the SVG in a specific time period can be enhanced, improving the diversity and practicality of the optimization objectives.

[0090] Step 103: Based on the fitness function, perform parameter optimization on the parameter set to be optimized through particle swarm optimization to obtain the optimal parameter set;

[0091] In some embodiments, based on the fitness function, particle swarm optimization is used to optimize the parameter set to obtain the optimal parameter set. Specifically, this can be as follows:

[0092] Particle swarm optimization is used to iteratively update the parameter set to be optimized. In each iteration, transient simulation is performed on the control model, and the voltage time series of each node is recorded. Based on the voltage time series of each node, a voltage response curve is generated. Based on the fitness function, the voltage response curve is time-aligned with the low-voltage reference curve and the error is calculated to complete the fitness evaluation and generate multiple fitness values ​​for the current iteration. After the current iteration ends, the simulation result when the current fitness value is optimal is saved. Through continuous iterative optimization, when the maximum number of iterations is reached, or when the preset low-voltage requirement is met, the final optimal parameter set is output.

[0093] In each iteration, each particle in the particle swarm represents a set of control parameters to be optimized. This parameter set is automatically written into the simulation platform's configuration file (e.g., a .SWI file). Power system transient simulation software is called to perform calculations, and the simulation output file (e.g., a .CHT file) records the voltage time series of each node. The system automatically extracts the voltage response curves of the SVG nodes, performs time alignment and error calculation with the standard low-voltage breakover curve, and completes the fitness assessment.

[0094] The aforementioned standard low-voltage curve can be viewed as a piecewise defined reference voltage lower limit curve. This curve is formed by splicing different time intervals and corresponding voltage thresholds. This curve serves as a benchmark for voltage response compliance; any region in the simulated output voltage response curve below this curve will trigger a piecewise penalty, used to calculate the fitness function. Essentially, the standard low-voltage curve divides the voltage transient recovery process into multiple key time intervals and sets a minimum allowable voltage value for each key time period, thus constructing a "stepped" standard voltage lower limit curve. This curve serves as a compliance reference for the voltage response in the fitness function; all simulated SVG voltage response curves must be compared with this standard curve time-by-time. After comparison, any portion below this curve will trigger a penalty mechanism to guide the optimization direction.

[0095] The optimization process employs the standard particle swarm optimization algorithm for iterative group updates. The velocity and position updates for each particle follow the following formula:

[0096]

[0097]

[0098] in, and The first The particle in the first The speed and position of the generation; This represents the optimal position in the individual particle's history. This is the globally optimal position for the entire population; Inertia factor; , As a learning factor, , It is a random number between [0,1].

[0099] In each iteration, the system automatically saves the simulation results corresponding to the current optimal control parameter set (corresponding to the optimal fitness value). These simulation results include the voltage response curve, low-breakdown reference curve, and optimal fitness value generated during transient simulation by the control model. After each iteration, the system outputs and saves the voltage response curve, low-breakdown reference curve, and optimal fitness value in structured table format for subsequent comparative analysis or archiving by the user.

[0100] Through continuous iterative optimization, the particle swarm optimization algorithm can effectively converge to the optimal SVG control parameter set that meets the low-penetration compliance requirements.

[0101] Step 104: Use the optimal parameter set as the control parameter set for the SVG in actual application.

[0102] The final output of the optimal parameter set can be used for actual deployment of SVG controllers or for policy migration testing in multi-condition simulation environments, demonstrating high engineering practicality and versatility.

[0103] It should be noted that the technical solution provided in this embodiment of the invention focuses on the dynamic tuning of SVG control parameters during power system disturbances (such as short-circuit faults). By using the deviation between the simulated voltage response curve and the low-voltage ride-through specification curve as the core of the fitness function, the compliance and dynamic support capabilities of the disturbance response are effectively improved. In practical applications, when it is necessary to extend the solution to everyday operating conditions, the method provided by this invention has good architectural portability. In real-world scenarios, only the definition of steady-state operating objectives (such as maintaining voltage stability, reducing power oscillations, etc.) needs to be added, and the fitness function and simulation task scheduling mechanism need to be appropriately adjusted to achieve joint tuning and optimization of SVG parameters under various operating conditions. Therefore, the technical solution provided by this invention has the feasibility and practical value of being extended to a wider range of real-world operating scenarios.

[0104] This invention provides a method for dynamic tuning of SVG control parameters based on particle swarm optimization. First, the deviation between the SVG node voltage response and the low-voltage ride-through characteristic curve of new energy grid connection is used as the core of fitness evaluation to quantify the compliance of the SVG during voltage transient processes after disturbances. Next, by analyzing the deviation between the simulated output voltage response curve and the low-voltage ride-through reference limit curve, an optimization objective function reflecting the SVG response compliance and regulation effect is constructed to guide the parameter search direction. Finally, combining the efficient convergence characteristics of the particle swarm optimization algorithm in continuous space, the optimal SVG controller parameter set is automatically found, achieving automatic optimization and dynamic tuning of the control parameters. By adopting the method provided in this invention, on the one hand, based on the matching relationship between the simulation output and the standard curve, it has stronger applicability, versatility, and ease of engineering deployment, and can be widely applied to grid operating environments with stringent voltage disturbance compliance requirements, promoting the intelligent upgrading of SVG parameter tuning and improving the voltage stability control capability of new power systems. On the other hand, it can achieve simulation-driven automatic optimization of SVG control parameters, significantly improving the SVG's control capability for voltage stability during disturbances. Meanwhile, the method has a clear structure, high computational efficiency, and can be adapted to various SVG control models and power grid operation scenarios. It has significant engineering value and promising prospects for promoting intelligent voltage control and improving the stability of power systems.

[0105] For better explanation, refer to Figure 2 This illustration shows a flowchart of dynamic tuning of SVG parameters through particle swarm optimization according to an embodiment of the present invention. It should be noted that this embodiment only provides a brief description of the general process of particle swarm optimization; the specific implementation process of each step can be understood by referring to the relevant content in the foregoing embodiments, and will not be elaborated here. It is understood that the present invention does not impose any limitations on this.

[0106] Step 201: Initialize particle swarm parameters, including the population size and number of iterations; initialize the position and velocity of each particle (i.e., the set of control parameters to be optimized); jump to step 202;

[0107] Step 202: Write the parameters to the configuration file; Proceed to step 203;

[0108] Step 203: Run SVG transient simulation, extract the voltage response curve, compare the voltage response curve with the low-through reference curve, and calculate the fitness value of each particle; jump to step 204;

[0109] Step 204: Select the particle with the best fitness value (smallest value) as the global best particle in the current iteration, and proceed to the next iteration, updating the particle's velocity and position; jump to step 205;

[0110] Step 205: Compare the voltage response curve after position update with the low-penetration reference curve, and recalculate the fitness value of each particle; jump to step 206;

[0111] Step 206: Determine if the convergence condition is met; if yes, proceed to step 207; if no, proceed to step 205.

[0112] Step 207: Select the particle with the best fitness value (smallest value) as the final optimal particle, use the optimal control parameter set corresponding to the final optimal particle as the SVG control parameter set, output the optimal parameters, plot the voltage response curve at the last iteration, and save the results; end the optimization process.

[0113] This invention proposes a dynamic tuning method for SVG parameters based on particle swarm optimization. The aim is to automatically adjust the SVG controller parameters through simulation-driven mechanisms, ensuring good voltage support performance under power system fault disturbances and meeting the technical requirements of low-voltage ride-through standards. This method can automatically complete parameter configuration, dynamic simulation, voltage assessment, and intelligent optimization without manual intervention, exhibiting strong engineering deployability and versatility.

[0114] To enable those skilled in the art to better understand the technical solutions of the present invention, the following specific example is used to illustrate the embodiments of the present invention.

[0115] In the specific implementation process, a power system transient simulation platform is used as the basic supporting environment. The selected optimized control parameters may include: the characteristic time constants of the filtering and measurement circuits. The lead time constant of primary feedforward control Lag time constant of primary feedforward control The lead time constant of secondary feedforward control Lag time constant of secondary feedforward control Proportional link time constant Proportional gain Integral gain The slope of the voltage-current characteristic curve (differential control factor) .

[0116] The control vector composed of the above parameters is:

[0117]

[0118] To ensure the parameter search process operates within a reasonable physical range, the system sets upper and lower boundaries for each parameter. In this example, the specific settings are as follows: ∈ [0.01, 0.4]; to ∈ [0.001, 0.2];

[0119] ∈ [0.001, 0.2]; ∈ [0.1, 100.0]; ∈ [0.01, 20.0]; ∈[0.001, 2].

[0120] Reference Figure 3 This diagram illustrates an optimization of SVG control parameters in this example.

[0121] Figure 3 The block diagram of the SVG controller shown illustrates the various control components involved in parameter optimization in this invention. These include a filter module, primary and secondary feedforward control channels, a PI regulator (proportional gain and integral), a response delay stage, and a derivative control module. Based on the SVG structure, this invention optimizes several key parameters, including the time constants of the filtering and measurement stages, the lead and lag time constants in the primary and secondary feedforward channels, the time constant of the proportional stage, the PI controller gain parameters, and the slope of the voltage-current characteristic curve. These parameters are distributed across the main control modules of the SVG and directly influence its disturbance response characteristics.

[0122] In the actual optimization process, the particle swarm optimization algorithm will randomly initialize the parameter vector and velocity vector of each particle within these boundaries to form a parameter population.

[0123] In each iteration, each particle in the particle swarm corresponds to a set of control parameters to be optimized. The system writes this set of control parameters into the simulation configuration file (e.g., a .SWI file), to the specific line containing the controller parameters (e.g., line 2751). The writing process must maintain consistent formatting precision to ensure the simulation platform correctly recognizes the parameters.

[0124] After the parameters are written, the simulation program is automatically invoked to run the specified transient condition. The simulation inputs include: fault type and location configuration; fault clearing time (…). ); the bus number where the SVG is located.

[0125] Simulation results are output in .CHT format. They include time-series data for voltage, current, active power, and reactive power at each node in the system.

[0126] After simulation, the system parses the .CHT binary file and automatically selects the voltage response curves corresponding to the SVG nodes. The extracted results are plotted with time series on the horizontal axis and voltage amplitude on the vertical axis to form a standard voltage response trajectory. To ensure timing consistency, the system performs the following actions during data processing: duplicate point removal (such as redundant records at the initial stage of a fault, such as 0.2s and 0.3s); time alignment processing; and multi-node voltage mean fusion (for example, applicable to multiple SVG access points).

[0127] This invention constructs a piecewise penalized loss function based on the difference between the simulated voltage response curve and the standard low-breakdown voltage curve. Taking the fault clearing time as the starting point, the voltage response is divided into four key time periods. In this example, the parameter settings for each key time period are shown in Table 1 below:

[0128] Table 1: Parameter Settings for Key Time Periods

[0129]

[0130] The penalty function is shown below:

[0131]

[0132] in, For the first Each node at time... The voltage response; For the first The minimum voltage threshold of the segment, This represents the time interval corresponding to this segment.

[0133] In this invention, the penalty weights at each stage are set in an increasing trend, a design that corresponds to the gradual increase in the voltage threshold. This reflects a dual constraint on both the voltage transient recovery speed and the steady-state maintenance capability. This setting guides the optimization algorithm to focus on the recovery performance in the mid-to-late stages, thereby meeting the compliance requirements of power systems for voltage stability.

[0134] The fitness function applies a weighted penalty to all moments when the voltage is below the low-pass standard curve. The earlier, deeper, and longer the deviation, the higher the loss will be, which will serve as the optimization objective value for particle swarm optimization.

[0135] The system performs an iterative particle swarm search based on the fitness function described above. In each round, the particle's position and velocity are updated using the following formula:

[0136]

[0137]

[0138] in, and The first The particle in the first The speed and position of the generation; This represents the optimal position in the individual particle's history. This is the globally optimal position for the entire population; Inertia factor; , As a learning factor, , It is a random number between [0,1].

[0139] After each simulation, the system automatically records the particle fitness value and updates the historical best position until the convergence condition or the maximum number of iterations is reached.

[0140] The optimal results for each round are exported as an Excel file. This file may include: the optimized parameter set, the SVG node voltage response curve, the corresponding low-throughput standard curve, and the voltage deviation and total loss value (fitness value) for each segment.

[0141] The final output set of optimal control parameters can be directly used for setting the field SVG controller. It can also serve as initial values ​​for different operating modes (such as load variation scenarios and power plant output variation scenarios) for the rapid generation of multi-condition control strategies. These changes in operating modes are typically reflected in differences in system load levels or renewable energy output, such as regional load peaks and troughs, and power plant output variations.

[0142] Reference Figure 4 This diagram illustrates a comparison between the voltage response curve and the low-breakdown reference curve in this example.

[0143] The voltage response curve is the simulated voltage change trajectory of the SVG node. The low-through reference curve is a standard comparison curve constructed from multiple segmented thresholds. This invention uses the deviation between the voltage response and the standard curve as the fitness function input, penalizing all regions below the standard curve, thereby guiding the optimization algorithm to output a more compliant set of control parameters. Figure 4 In the graph, the horizontal axis represents time (in cycles), and the vertical axis represents voltage amplitude (in pu). Figure 4 The image shows the voltage recovery trajectory of the system after the disturbance was cleared, as well as the low-pass standard comparison curve (step line) constructed from the voltage thresholds of each time period. Figure 4 It visually demonstrates the triggering conditions and effects of the penalty mechanism in fitness function evaluation.

[0144] The method provided by this invention achieves fully automated tuning of SVG control parameters, thereby eliminating subjective errors caused by manual settings. On the one hand, by using low-voltage compliance as the core indicator, the engineering interpretability of the optimization objective is improved. On the other hand, based on an integrated simulation-extraction-evaluation-iteration process, the controllability and response speed of system voltage disturbance recovery are enhanced. The technical solution proposed in this invention is applicable to various power grid scenarios such as new energy access systems and urban load centers, and has good promotional value and practical application prospects.

[0145] Reference Figure 5 The diagram illustrates a structural block diagram of an SVG parameter dynamic tuning device provided in an embodiment of the present invention, which may specifically include:

[0146] The parameter set determination unit 501 is used to construct the control model of the SVG and determine the parameter set to be optimized of the control model;

[0147] The fitness function construction unit 502 is used to consider the voltage response of the control model during transient simulation and construct a fitness function based on voltage recovery piecewise penalty.

[0148] The parameter optimization unit 503 is used to optimize the parameter set to be optimized by particle swarm optimization based on the fitness function to obtain the optimal parameter set.

[0149] The parameter set application unit 504 is used to use the optimal parameter set as the control parameter set of the SVG in actual application.

[0150] In one alternative embodiment, the fitness function construction unit 502 includes:

[0151] The critical time period division unit is used to consider the voltage response of the control model during transient simulation and divide the voltage response into four critical time periods.

[0152] The segmented parameter setting unit is used to set the minimum voltage threshold for each of the key time periods and to set the segmented penalty weight for each of the minimum voltage thresholds.

[0153] The fitness function construction sub-unit is used to construct a fitness function based on voltage recovery segmented penalty according to the minimum voltage threshold and segmented penalty weight of each key time period and the voltage response.

[0154] In one optional embodiment, the minimum voltage threshold and segmented penalty weight for each of the key time periods are respectively set as follows:

[0155] First period Corresponding to the first minimum voltage threshold First segment penalty weight ;

[0156] Second period Corresponding to the second lowest voltage threshold Second segment penalty weight ;

[0157] Third period Corresponding to the third lowest voltage threshold The third segment penalty weight ;

[0158] Fourth period Corresponding to the fourth lowest voltage threshold The fourth segment penalty weight ;

[0159] The time unit for each period is a cycle (Zodiac); This refers to the time required to clear the fault. Indicates the time.

[0160] In one optional embodiment, the parameter optimization unit 503 is specifically used for:

[0161] The set of parameters to be optimized is updated iteratively through particle swarm optimization. In each iteration:

[0162] A transient simulation was performed on the control model, and the voltage time series of each node was recorded.

[0163] Based on the voltage time series of each node, a voltage response curve is generated;

[0164] Based on the fitness function, the voltage response curve and the low-breakdown reference curve are time-aligned and error is calculated to complete the fitness evaluation and generate multiple fitness values ​​for the current iteration round.

[0165] After the current iteration ends, save the simulation results when the current fitness value is optimal;

[0166] Through continuous iterative optimization, when the maximum number of iterations is reached, or when the preset low-pass requirement is met, the final optimal parameter set is output.

[0167] In one optional embodiment, the simulation results include the voltage response curve, low-breakdown reference curve, and optimal fitness value generated when the control model performs transient simulation; the device further includes:

[0168] The simulation results output unit is used to output and save the voltage response curve, low-breakdown reference curve and optimal fitness value after each iteration in a structured table format.

[0169] In one optional embodiment, the parameter set determination unit 501 includes:

[0170] The variable selection unit is used to construct the control model of the SVG and select multiple key control parameters in the control model as variables to be optimized.

[0171] The parameter search space construction unit is used to set upper and lower bounds for each of the variables to be optimized and construct the parameter search space.

[0172] The variable integration unit is used to integrate the various variables to be optimized located in the parameter search space as the set of parameters to be optimized for the control model.

[0173] In one optional embodiment, the variables to be optimized in the set of parameters to be optimized include the characteristic time constant of the filtering and measurement circuit, the lead time constant of the first feedforward control, the lag time constant of the first feedforward control, the lead time constant of the second feedforward control, the lag time constant of the second feedforward control, the time constant of the proportional element, the gain of the proportional element, the gain of the integral element, and the slope of the voltage-current characteristic curve.

[0174] As the device embodiment is basically similar to the method embodiment, it is described in a relatively simple way. For relevant details, please refer to the description of the method embodiment above.

[0175] It should be noted that, in order to enable those skilled in the art to better distinguish data of the same type but with different actual meanings, the embodiments of the present invention use terms such as "first" and "second" to distinguish and describe some technical features. The terms "first" and "second" are used only for data differentiation and have no other special meanings. It is understood that the present invention does not impose any limitations on them.

[0176] This invention also provides an electronic device, which includes a processor and a memory:

[0177] The memory is used to store program code and transfer the program code to the processor;

[0178] The processor is used to execute the SVG parameter dynamic tuning method of any embodiment of the present invention according to the instructions in the program code.

[0179] This invention also provides a computer-readable storage medium for storing program code for executing the SVG parameter dynamic tuning method of any embodiment of this invention.

[0180] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0181] In the embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between devices or units through some interfaces, and may be electrical, mechanical, or other forms.

[0182] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0183] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0184] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or 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, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0185] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for dynamically tuning SVG parameters, characterized in that, include: Construct a control model for the SVG and determine the set of parameters to be optimized for the control model; Considering the voltage response of the control model during transient simulation, a fitness function based on voltage recovery piecewise penalty is constructed. Based on the fitness function, the optimal parameter set is obtained by optimizing the parameter set to be optimized through particle swarm optimization. The optimal parameter set is used as the control parameter set for the SVG in practical applications.

2. The SVG parameter dynamic tuning method according to claim 1, characterized in that, The process considers the voltage response of the control model during transient simulation and constructs a fitness function based on voltage recovery piecewise penalty, including: Considering the voltage response of the control model during transient simulation, the voltage response is divided into four key time periods; A minimum voltage threshold is set for each of the key time periods, and a segmented penalty weight is set for each of the minimum voltage thresholds. Based on the minimum voltage threshold and segmented penalty weights for each critical time period, and combined with the voltage response, a fitness function based on voltage recovery segmented penalty is constructed.

3. The SVG parameter dynamic tuning method according to claim 2, characterized in that, The minimum voltage threshold and segmented penalty weight for each of the aforementioned key time periods are set as follows: First period Corresponding to the first minimum voltage threshold First segment penalty weight ; Second period Corresponding to the second lowest voltage threshold Second segment penalty weight ; Third period Corresponding to the third lowest voltage threshold The third segment penalty weight ; Fourth period Corresponding to the fourth lowest voltage threshold The fourth segment penalty weight ; The time unit for each period is a cycle (Zodiac); This refers to the time required to clear the fault. Indicates the time.

4. The SVG parameter dynamic tuning method according to claim 1, characterized in that, The step of optimizing the parameter set based on the fitness function using particle swarm optimization to obtain the optimal parameter set includes: The set of parameters to be optimized is updated iteratively through particle swarm optimization. In each iteration: A transient simulation was performed on the control model, and the voltage time series of each node was recorded. Based on the voltage time series of each node, a voltage response curve is generated; Based on the fitness function, the voltage response curve and the low-breakdown reference curve are time-aligned and error is calculated to complete the fitness evaluation and generate multiple fitness values ​​for the current iteration round. After the current iteration ends, save the simulation results when the current fitness value is optimal; Through continuous iterative optimization, when the maximum number of iterations is reached, or when the preset low-pass requirement is met, the final optimal parameter set is output.

5. The SVG parameter dynamic tuning method according to claim 4, characterized in that, The simulation results include the voltage response curve, low-breakdown reference curve, and optimal fitness value generated when the control model performs transient simulation; the method further includes: Output and save the voltage response curve, low-breakdown reference curve, and optimal fitness value after each iteration in a structured table format.

6. The method for dynamically tuning SVG parameters according to any one of claims 1 to 5, characterized in that, The construction of the SVG control model and the determination of the set of parameters to be optimized for the control model include: A control model for the SVG is constructed, and several key control parameters in the control model are selected as variables to be optimized. Set upper and lower bounds for each of the variables to be optimized, and construct a parameter search space; The various variables to be optimized located in the parameter search space are integrated as the set of parameters to be optimized for the control model.

7. The SVG parameter dynamic tuning method according to claim 6, characterized in that, The variables to be optimized in the set of parameters to be optimized include the characteristic time constant of the filtering and measurement circuit, the lead time constant of the first feedforward control, the lag time constant of the first feedforward control, the lead time constant of the second feedforward control, the lag time constant of the second feedforward control, the time constant of the proportional element, the gain of the proportional element, the gain of the integral element, and the slope of the voltage-current characteristic curve.

8. An SVG parameter dynamic tuning device, characterized in that, include: The parameter set determination unit is used to construct the control model of the SVG and determine the parameter set to be optimized of the control model; The fitness function construction unit is used to consider the voltage response of the control model during transient simulation and construct a fitness function based on voltage recovery piecewise penalty. The parameter optimization unit is used to optimize the parameter set to be optimized based on the fitness function and through particle swarm optimization to obtain the optimal parameter set. The parameter set application unit is used to use the optimal parameter set as the control parameter set of the SVG in actual application.

9. An electronic device, characterized in that, The device includes a processor and a memory: The memory is used to store program code and transmit the program code to the processor; The processor is used to execute the SVG parameter dynamic tuning method according to any one of claims 1-7 according to the instructions in the program code.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store program code for executing the SVG parameter dynamic tuning method according to any one of claims 1-7.

Images