Design method of damping controller based on power system, controller and power system
By optimizing the adaptive crossover mutation factor through a multi-parameter differential evolution identification model, the problem of damping controllers easily getting trapped in local optima is solved, achieving higher accuracy in model identification and better low-frequency oscillation suppression.
Patent Information
- Application Number
- CN202310856222.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-07-12
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2043-07-12
AI Technical Summary
In traditional power system-based damping controller design methods, the combination of artificial intelligence algorithms and traditional Smith predictors is prone to getting trapped in local optima, leading to model identification bias and potentially causing power system instability.
A multi-parameter differential evolution identification model is used to identify the parameters to be identified and the time delay factor in the transfer function of the controlled object online. By optimizing the adaptive crossover factor and the adaptive mutation factor, the convergence speed of the algorithm is accelerated and local optima are avoided. A damping controller is designed.
The model identification accuracy has been improved, the low-frequency oscillation suppression effect of wind power systems under multiple interference conditions has been enhanced, and the damping controller design is more reasonable.
Smart Images

Figure CN117096895B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of power system technology, and in particular to a damping controller design method, controller, and power system based on a power system. Background Technology
[0002] Power system data acquisition utilizes data network transmission technology, with an average data transmission delay approaching 100ms. However, the time delay cannot be fixed, making power system stability analysis and control more complex. In recent years, with the large-scale integration of wind power into the power system, small-disturbance stability problems have become increasingly prominent, further aggravating the impact of time delay on the wide-area damping control of the power system. Related technologies compensate for the predicted parameters output by the Smith predictor controller to eliminate the effects of time delay. Applications of the Smith predictor in handling time delay problems can be broadly categorized into three types: first, optimizing the Smith predictor structure; second, optimizing parameter performance indicators using traditional mathematical methods, such as robustness indicators, Lyapunov theory, and linear inequality theory; and third, combining artificial intelligence algorithms with traditional Smith predictors to improve their performance. Although the combination of artificial intelligence algorithms and traditional Smith predictors can achieve good fitting results for nonlinear parameters, it is generally prone to getting trapped in local optima and cannot determine the true global optimum. This leads to model identification bias, which in turn prevents the Smith predictor from fully compensating for time delays and may cause the system to become unstable. Summary of the Invention
[0003] This invention provides a damping controller design method, controller, and power system based on power systems. It addresses the shortcomings of traditional damping controller design methods based on power systems, which often fall into local optima by combining artificial intelligence algorithms with traditional Smith predictors. This results in the inability to determine the true global optimum, leading to model identification bias and potentially causing power system instability.
[0004] This invention provides a damping controller design method based on a power system, comprising:
[0005] Solve the initial controller model based on the preset power system model;
[0006] The preset power system model is used as the controlled object transfer function of the damping controller, and a time delay factor is introduced into the controlled object transfer function.
[0007] The optimized multi-parameter differential evolution identification model is used to identify the parameters to be identified in the transfer function of the controlled object and the time delay factor online to obtain the identification results, which include the identification results of the parameters of the transfer function of the controlled object and the identification results of the time delay factor.
[0008] The damping controller is designed based on the identification results.
[0009] According to the present invention, a damping controller design method based on a power system is provided, wherein the multi-parameter differential evolution identification model optimization method includes:
[0010] The parameters to be identified and the time delay factor in the transfer function of the controlled object are input into the multi-parameter differential evolution identification model to obtain the identification transfer function containing the time delay factor.
[0011] Calculate the Vinncombe distance between the controlled object transfer function and the identification transfer function;
[0012] If the Vinncombe distance does not meet the preset recognition accuracy requirement, then the parameters in the multi-parameter differential evolution recognition model are optimized.
[0013] The optimized multi-parameter differential evolution identification model is used to re-identify until the Vinncombe distance reaches the preset identification accuracy requirement, and then the optimized multi-parameter differential evolution identification model is output.
[0014] According to the damping controller design method based on a power system provided by the present invention, the calculation of the Vinncombe distance between the transfer function of the controlled object and the identification transfer function includes:
[0015] Obtain the Vinncombe distance constraint formula
[0016]
[0017] When the constraint formula is satisfied, the Vinncombe distance between the controlled object transfer function and the identification transfer function is expressed as:
[0018] maxκ(G1(e jω ), G2(e jω ))
[0019] When the constraint formula is not satisfied, the Vinncombe distance between the controlled object transfer function and the identification transfer function is expressed as 1;
[0020] In the formula, η(G2) is the number of poles in the right half-plane of G2. Let κ(G1(e) be the number of poles in the closed right half-plane of G1, and w(x) be the number of times the Nyquist curve of the transfer function encircles the origin counterclockwise. jω ),G2(e jω )) is the chord distance obtained by projecting G1 and G2 onto the unit Riemann sphere.
[0021] According to the damping controller design method based on a power system provided by the present invention, the method for determining whether the Vinncombe distance meets the preset identification accuracy requirement includes:
[0022] Under the premise of stable system output, determine whether the Vinncombe distance is within a preset error threshold;
[0023] If so, the Vinncombe distance is determined to meet the preset recognition accuracy requirement.
[0024] According to the present invention, a damping controller design method based on a power system is provided, wherein designing the damping controller based on the identification result includes:
[0025] The parameter matrix is obtained based on the preset power system model, the initial controller model, and the interference signal;
[0026] Estimate the system frequency stability margin on the error boundary based on the parameter matrix;
[0027] When the system frequency stability margin is greater than the maximum value of the Vinncombe distance, the initial controller model is used as a damping controller.
[0028] According to the present invention, a damping controller design method based on a power system, wherein designing the damping controller based on the identification result further includes:
[0029] When the system frequency stability margin is no greater than the maximum value of the Vinncombe distance, if |P|≤ρ,ρ(|P|ξ1+ξ2)<1, where P is a single-input single-output transfer function, then the damping controller is:
[0030]
[0031] In the formula: S i = (1+G i K i ) -1 , ξ1 is the difference between the updated frequency domain stability margin and the previous stability margin, ξ2 is the previous frequency domain stability margin, ρ is the maximum distance between the transfer functions of the identified model and the target model, and K is the maximum distance between them. i This is the initial controller model.
[0032] According to the present invention, a damping controller design method based on a power system is provided, wherein the multi-parameter differential evolution identification model includes an adaptive crossover factor and an adaptive mutation factor;
[0033] The adaptive crossover factor is:
[0034]
[0035] In the formula, CR min For the minimum adaptive crossover factor, CR max For the maximum adaptive crossover factor, G m G represents the maximum number of iterations and the generation number.
[0036] The adaptive mutation factor is:
[0037]
[0038] In the formula, F min F is the minimum adaptive variation factor. max G is the maximum adaptive mutation factor. m G represents the maximum number of iterations and the generation number.
[0039] According to the present invention, a damping controller design method based on a power system is provided, wherein solving the initial controller model based on a preset power system model includes:
[0040] The parameters of the preset power system model are estimated based on the online identification method for low-frequency oscillation mode noise signals in the power system.
[0041] The estimation results are tested according to the model order determination criterion. When the estimation results meet the requirements of the model order determination criterion, a reduced-order power system model is obtained.
[0042] The initial controller model is obtained by solving the closed-loop stability condition corresponding to the reduced-order power system model.
[0043] The present invention also provides a damping controller, comprising: a damping controller designed based on the power system-based damping controller design method described in any one of the preceding claims.
[0044] The present invention also provides a power system including the damping controller described above.
[0045] The present invention provides a damping controller design method, controller, and power system based on a power system. The method involves solving an initial controller model based on a preset power system model; using the preset power system model as the controlled object transfer function of the damping controller, and introducing a time delay factor into the controlled object transfer function; using an optimized multi-parameter differential evolutionary identification model to perform online identification of the parameters to be identified and the time delay factor in the controlled object transfer function, obtaining identification results, including the identification results of the controlled object transfer function parameters and the identification results of the time delay factor; and designing a damping controller based on the identification results. Using the optimized multi-parameter differential evolutionary identification model to perform online identification of the parameters to be identified and the time delay factor in the controlled object transfer function can accelerate the algorithm's convergence speed, avoid getting trapped in local optima, and achieve higher identification accuracy. The resulting damping controller design is more reasonable, and it has a better suppression effect on low-frequency oscillations in power systems containing wind power under multiple interference conditions. Attached Figure Description
[0046] To more clearly illustrate the technical solutions in this 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 some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.
[0047] Figure 1 This is one of the flowcharts illustrating the design method of a damping controller based on a power system provided by the present invention;
[0048] Figure 2 This is the second flowchart of the power system-based damping controller design method provided by the present invention;
[0049] Figure 3 This is the third flowchart illustrating the design method for damping controllers based on power systems.
[0050] Figure 4 A power system identification model diagram provided by the present invention;
[0051] Figure 5 The schematic diagram of the Smith predictive control structure provided by this invention;
[0052] Figure 6 The first-order system identification results and convergence curves provided by this invention;
[0053] Figure 7 The present invention provides step response curves of the Smith predictor under different conditions;
[0054] Figure 8 The schematic diagram of the four-machine two-zone system provided by this invention;
[0055] Figure 9 This is a diagram showing the first identification result of the adaptive multi-parameter differential evolution algorithm provided by the present invention.
[0056] Figure 10 This is a diagram showing the second identification result of the adaptive multi-parameter differential evolution algorithm parameters provided by the present invention;
[0057] Figure 11 The diagram showing the dynamic relationship between frequency domain stability margin and Vinnicombe distance provided by this invention;
[0058] Figure 12 These are rotor angle curves under different time delay conditions provided by this invention;
[0059] Figure 13 This is a comparison diagram between the least squares iterative identification method provided by this invention and the method provided by this invention.
[0060] Figure 14 This invention provides a rotor angle curve with a time delay of 100ms.
[0061] Figure 15 This is the rotor angle curve provided by the present invention when the time delay is 150ms. Detailed Implementation
[0062] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.
[0063] Figure 1 A flowchart of a power system-based damping controller design method provided in an embodiment of the present invention is shown below. Figure 1 As shown, the damping controller design method based on a power system provided in this embodiment of the invention includes:
[0064] Step 101: Solve the initial controller model based on the preset power system model;
[0065] In this embodiment of the invention, solving the initial controller model based on a preset power system model includes:
[0066] Step 1011: Estimate the parameters of the preset power system model based on the online identification method for low-frequency oscillation mode noise signals in power systems;
[0067] Step 1012: Verify the estimation results according to the model order determination criteria. When the estimation results meet the requirements of the model order determination criteria, a reduced-order power system model is obtained.
[0068] Step 1013: Solve for the initial controller model based on the closed-loop stability conditions corresponding to the reduced-order power system model.
[0069] Step 102: Use the preset power system model as the controlled object transfer function of the damping controller, and introduce a time delay factor into the controlled object transfer function;
[0070] Step 103: Use the optimized multi-parameter differential evolution identification model to identify the parameters to be identified and the time delay factor in the transfer function of the controlled object online, and obtain the identification results, including the identification results of the parameters of the transfer function of the controlled object and the identification results of the time delay factor.
[0071] In this embodiment of the invention, the multi-parameter differential evolution identification model includes an adaptive crossover factor and an adaptive mutation factor;
[0072] The adaptive crossover factor is:
[0073]
[0074] In the formula, CR min For the minimum adaptive crossover factor, CR max For the maximum adaptive crossover factor, G m G represents the maximum number of iterations and the generation number.
[0075] In the evolution process of the adaptive differential evolution algorithm, the crossover factor and the crossover rate change dynamically as the number of iterations increases. In the early stage, a larger crossover factor ensures the global variation, while in the later stage, a smaller crossover rate focuses more on the local convergence.
[0076] The adaptive mutation factor is:
[0077]
[0078] In the formula, F min F is the minimum adaptive variation factor. max G is the maximum adaptive mutation factor. m G represents the maximum number of iterations and the generation number.
[0079] In differential evolution algorithms, the settings of control parameters directly affect the convergence speed and reliability of the algorithm. Differential evolution algorithms control relatively few parameters, including population size NP, scaling factor (mutation parameter) F, and crossover rate (crossover probability parameter) CR. The choice of control parameters has a significant impact on the performance of differential evolution algorithms. The scaling factor (mutation parameter) F controls the size of the basis vector. A smaller F results in smaller compensation for mutations, improving the algorithm's development capability and making it easier to find a higher-quality optimal solution. However, this leads to a longer convergence time and makes the algorithm more prone to getting trapped in local optima. A larger F causes greater perturbation to the difference vector, which is beneficial for maintaining population diversity and improving the algorithm's search capability. However, the downside is that the algorithm tends towards random search, making it difficult to find the optimal solution.
[0080] Therefore, designing the above-mentioned adaptive crossover factor and adaptive mutation factor can accelerate the convergence speed of the model and avoid getting trapped in local optima.
[0081] Step 104: Design a damping controller based on the identification results.
[0082] Traditional nonlinear parameter fitting methods that combine artificial intelligence algorithms with Smith predictors are prone to getting trapped in local optima and cannot determine the true global optimum. This leads to model identification bias, which in turn prevents the Smith predictor from fully compensating for time delays and may cause system instability.
[0083] The damping controller design method based on a power system provided in this invention involves solving an initial controller model based on a preset power system model; using the preset power system model as the controlled object transfer function of the damping controller, and introducing a time delay factor into the controlled object transfer function; using an optimized multi-parameter differential evolution identification model to perform online identification of the parameters to be identified and the time delay factor in the controlled object transfer function, obtaining identification results, including the identification results of the controlled object transfer function parameters and the identification results of the time delay factor; and designing a damping controller based on the identification results. Using the optimized multi-parameter differential evolution identification model to perform online identification of the parameters to be identified and the time delay factor in the controlled object transfer function can accelerate the algorithm's convergence speed, avoid getting trapped in local optima, and achieve higher identification accuracy. The resulting damping controller design is more reasonable, and it has a better suppression effect on low-frequency oscillations in wind-powered power systems under multiple interference conditions.
[0084] Based on any of the above embodiments, such as Figure 2 As shown, the optimization method for the multi-parameter differential evolution identification model includes:
[0085] Step 201: Input the parameters to be identified and the time delay factor in the transfer function of the controlled object into the multi-parameter differential evolution identification model to obtain the identification transfer function containing the time delay factor;
[0086] Step 202: Calculate the Vinncombe distance between the controlled object transfer function and the identification transfer function;
[0087] In this embodiment of the invention, Vinnicombe distance refers to the distance between two frequency responses, representing a measurement of the distance between two transfer functions.
[0088] In this embodiment of the invention, calculating the Vinncombe distance between the controlled object transfer function and the identification transfer function includes:
[0089] Obtain the Vinncombe distance constraint formula
[0090]
[0091] When the above constraint formula is satisfied, the Vinncombe distance between the controlled object transfer function and the identification transfer function is expressed as:
[0092] maxκ(G1(e jω ), G2(e jω ))
[0093] When the above constraint formula is not satisfied, the Vinncombe distance between the controlled object transfer function and the identification transfer function is expressed as 1.
[0094] In the formula, η(G2) is the number of poles in the right half-plane of G2. Let κ(G1(e) be the number of poles in the closed right half-plane of G1, and w(x) be the number of times the Nyquist curve of the transfer function encircles the origin counterclockwise. jω ),G2(e jω )) is the chord distance obtained by projecting G1 and G2 onto the unit Riemann sphere.
[0095] Step 203: If the Vinncombe distance does not meet the preset recognition accuracy requirement, then optimize the parameters in the multi-parameter differential evolution recognition model.
[0096] In this embodiment of the invention, the method for determining whether the Vinncombe distance meets the preset recognition accuracy requirement includes:
[0097] Under the premise of stable system output, determine whether the Vinncombe distance is within a preset error threshold;
[0098] If so, the Vinncombe distance is determined to meet the preset recognition accuracy requirement.
[0099] When the Vinncombe distance reaches the preset recognition accuracy requirement, the controller can keep the recognized model set calm, indicating that the controller is a robust and stable controller, and therefore can be directly used as a wide-area time-delay damping controller.
[0100] Otherwise, it is determined that the Vinncombe distance does not meet the preset recognition accuracy requirement.
[0101] If the Vinncombe distance is not within the preset error threshold during the closed-loop identification process of the power system, the damping controller of the system needs to be improved, and a new method for improving the damping controller needs to be given.
[0102] In this embodiment of the invention, the preset error threshold is, for example, 0.05. If the error range is met, it means that the system can output a result that is closest to the real identification model under stable conditions. This ensures that the system can meet the frequency domain stability and makes the identification result as accurate as possible. If the error range is not met, it means that the system identification effect has a large error and it is necessary to re-identify or adjust the relevant parameters to achieve a better identification effect.
[0103] Step 204: Re-identify using the optimized multi-parameter differential evolution identification model until the Vinncombe distance reaches the preset identification accuracy requirement, and output the optimized multi-parameter differential evolution identification model.
[0104] Based on any of the above embodiments, such as Figure 3 As shown, the damping controller is designed based on the identification results, including:
[0105] Step 301: Obtain the parameter matrix based on the preset power system model, the initial controller model, and the interference signal;
[0106] Step 302: Estimate the system frequency stability margin on the error boundary based on the parameter matrix;
[0107] Step 303: When the system frequency stability margin is greater than the maximum value of the Vinncombe distance, the initial controller model is used as a damping controller.
[0108] Step 304: When the system frequency stability margin is no greater than the maximum value of the Vinncombe distance, if |P|≤ρ,ρ(|P|ξ1+ξ2)<1, where P is a single-input single-output transfer function, then the damping controller is:
[0109]
[0110] In the formula: S i = (1+G i Ki ) -1 , ξ1 is the difference between the updated frequency domain stability margin and the previous stability margin, ξ2 is the previous frequency domain stability margin, and ρ is the maximum distance between the transfer functions of the identified model and the target model.
[0111] In embodiments of the present invention, the power system identification model is as follows: Figure 4 As shown, u and y represent the input and output signals of the controlled system, respectively; e represents a random disturbance signal with zero mean and variance λ; r represents an external reference signal independent of signal e; G(q,θ) represents the forward channel controlled system model; H(q,θ) represents the filter model of the forward channel disturbance signal e; K(q,θ) represents the feedback channel controller model; and θ represents the model parameters to be identified. The closed-loop system expression can be written as:
[0112] y(t)=S(q,θ)G(q,θ)r(t)+S(q,θ)H(q)e(t)
[0113] u(t)=S(q,θ)r(t)-S(q,θ)K(q,θ)e -τ(q,θ) H(q)e(t)
[0114] It can be further written as:
[0115]
[0116] where S(q,θ)=(1+G(q,θ)K(q,θ)e -τ(q,θ) ) -1 This is a sensitivity function.
[0117] The transfer function of the time-delay system model in continuous time is shown in the following equation, which can be represented by a conventional transfer function and a time-delay term. The product of
[0118]
[0119] Where: k p It is the proportional coefficient of the transfer function, T p It is the time constant, τ p It is a time delay term.
[0120] As can be seen from the above equation, when a closed-loop system is constructed, the denominator of the closed-loop transfer function H(s) in the following equation will contain a time delay term, which will cause H(s) to have an infinite number of poles. Therefore, the system may have instability problems and it is difficult to achieve effective control.
[0121]
[0122] The Smith predictor is a type of predictive controller. The Smith predictor method can completely cancel out time-delay terms. Its control structure is as follows: Figure 5 As shown, assume the model transfer function is
[0123]
[0124] The Smith controller's controlled object transfer function is
[0125]
[0126] From the above equation, we can see that its closed-loop transfer function is:
[0127]
[0128] The characteristic equation is
[0129] 1+K(s)G(s)e -τs =0
[0130] To eliminate lag terms in the characteristic equation, the denominator cannot contain time-delay terms, requiring its closed-loop transfer function to...
[0131] 1+K(s)G o (s)=0
[0132] If the model perfectly matches the controlled object, that is, G(s) = G p In the absence of external interference, a good control effect can be obtained by designing a reasonable controller K(s). However, due to the complexity of the controlled object in the industrial field or actual environment and the difficulty in establishing a relatively accurate mathematical model, coupled with external interference, the traditional Smith control structure will be difficult to adapt to the real situation.
[0133] Differential evolution is a type of evolutionary algorithm. During initialization, NP individuals are randomly generated in a D-dimensional direction. During evolution, individuals are subjected to mutation, crossover, and selection operations. Assuming the population initialization uses G = 0, 1, 2, ..., Gmax to represent the generations, the i-th individual in the current generation is represented as:
[0134]
[0135] In the formula, D represents the dimension of the population. During initialization, the population should cover the entire search space.
[0136]
[0137] In the formula, xi(0) represents the i-th chromosome in the 0th generation of the population; rand(0,1) represents a random number uniformly distributed in the interval (0,1). A difference strategy is used to achieve individual differences. Two different individuals are randomly selected from the population, their vector difference is scaled, and then combined with the individual to be mutated to generate an intermediate.
[0138] v i (g+1)=x r1 (g)+F·[x r2 (g)-x r3 (g)],i≠r1≠r2≠r3
[0139] In the formula, F is the mutation factor; xi(g) is the i-th individual in the g-th generation population.
[0140] Crossover operations, including binomial crossover and exponential crossover; performing crossover operations between individuals on the g-th generation population xi(g) and its intermediate variant vi(g+1):
[0141]
[0142] In the formula, CR is the cross-factor.
[0143] The selection operation, in the differential algorithm, generates a population of offspring through mutation and crossover operations. Then, a one-to-one greedy selection operator compares each individual with its corresponding parent, saving the better one for the next generation. The selection operation is shown below:
[0144]
[0145] Constructing a multi-parameter differential evolution identification model based on the differential algorithm can improve the accuracy of the identification results.
[0146] This invention addresses the time-delay compensation and control problem in power systems with wind power integration and minor disturbances. It improves the adaptive differential evolution algorithm using the Vinncombe metric and frequency domain stability margin correlation theory, and applies it to power system model identification and online Smith predictor control. The system stability judgment conditions are derived, and this improvement addresses the problem of the adaptive multi-parameter differential evolution algorithm easily getting trapped in local optima during identification. A time-delay wide-area damped controller is designed for the reduced-order model of a power system with wind power integration and minor disturbances. Based on the identification results, the Smith predictor controller is improved, and improvements are made for handling disturbance signals, model-parameter mismatches, and time-delay factor mismatches.
[0147] Improved multi-parameter identification differential evolution algorithm Smith control structure, such as Figure 4 As shown, the design steps are as follows:
[0148] Step 1: Given the power system model G0 to be initially reduced in order, derive the controller model K based on the above.
[0149] Step 2: Use the above model G0 as the controlled object transfer function of the Smith predictor, and introduce the relevant lag factor e. -τs The controller model K serves as the controller for the Smith predictor, and the Smith predictor control structure is constructed.
[0150] Step 3: Apply the improved multi-parameter differential evolution algorithm described above to the time delay term e in the feedback loop. -τs Perform online identification with the item to be identified.
[0151] Step 4: Based on Vinncombe theory, calculate the relationship between the transfer function G0 of the controlled object and the transfer function identified by the differential evolution algorithm. p distance δ v (G0,G p ).
[0152] Step 5: Based on the frequency stability margin theory, determine whether the identification model has reached the identification accuracy. If it cannot reach the preset accuracy requirement, return to Step 3 above to perform model identification.
[0153] Step 6: Substitute the identification models obtained in Steps 3, 4 and 5 into the Smith predictive controller established in Step 2 to perform online real-time compensation for the time delay element.
[0154] To verify the design effectiveness of the damping controller, in this embodiment of the invention, a first-order system is used as an example to simulate and verify the designed damping controller. The system transfer function is assumed to be... The time delay factors are mismatched. The preset target time delay factor t = 50, and the time delay compensation factor t = 10. The parameters of the model to be identified in the time delay compensation stage are initialized with random parameters according to the differential evolution algorithm. Its input signal is a step signal. The parameter identification results of the improved multi-objective differential evolution algorithm are as follows: Figure 6 As shown, assume the parameters to be identified are {α1, α2, α3, τ}. Figure 6 The scatter plots in the model represent parameter changes during the identification process, and the model outputs the identification results.
[0155]
[0156] The identification error index is,
[0157]
[0158] Where N is the number of iterations for identification; y i This is the output for the i-th sample. The identification error of the identification result is 6.1221e. -8This identification process can directly identify the time delay factor without using traditional time delay reduction methods, reducing the conservatism of the system identification stage, and the identification accuracy is close to the ideal situation. It can be seen that the output of the damping controller design method provided in this embodiment is very close to the actual system.
[0159] like Figure 7 As shown, simulations are performed under three scenarios: mismatched controlled object parameters, mismatched time delay factors, and mismatched controlled object and time delay factors. Figure 7 It can be seen that the system's step response exhibits significant oscillations when there is a mismatch in time delay factors, mismatch in controlled object parameters, or mismatch between the controlled object and the time delay factors. The method provided in this embodiment closely approximates the step response under ideal system conditions. It offers better time delay compensation than the traditional Smith predictor and can address the issues of mismatched controlled object parameters and time delay factors to a certain extent.
[0160] In this embodiment of the invention, a four-machine, two-zone power system is used for simulation, such as... Figure 8 As shown, the transfer function of the reduced-order model of the power system can be expressed as:
[0161]
[0162] From the above transfer function, the initial damping controller of the system is:
[0163]
[0164] Assume the identification parameters of the target transfer function of the reduced-order power system to be identified are as follows: The transfer function of the solid system after identification can be expressed as:
[0165]
[0166] The time delay factor τ of the preset target transfer function is 0.8. The identification results of the first power system are α1 = 1, α2 = 2.995, α3 = 2, and τ = 0.8. Figure 9 As shown, the identification results for the second power system are β1 = 0.998, β2 = 4.998, β3 = 5.24, and β4 = 4.998. Figure 10 As shown. Similar to the first-order system, the identified power system model is as follows.
[0167]
[0168] The relationship between frequency domain stability margin and the dynamic change of Vinnicombe distance is shown in the figure below. Figure 11As shown, it can be seen that before the 5th identification, the frequency domain stability margin of the system is less than the Vinnicombe distance, which will not satisfy the above stabilization condition. It can be concluded that the controller is insufficient to stabilize the system. With the increase of the number of iterations, after the 5th identification, the frequency domain stability margin is greater than the Vinnicombe distance. It can be concluded that the controller can stabilize the system, and the frequency domain stability margin gradually increases. Therefore, the optimal controller can be obtained.
[0169] Rotor angle curves under different time delay conditions are as follows Figure 12 As shown, simulations were performed under four different conditions: 50ms, 100ms, 150ms, and in the absence of time delay and additional control. Through comparison, it can be seen that the method provided by the embodiments of the present invention has a significant suppression effect on low-frequency oscillations under different time delay conditions.
[0170] like Figure 13 As shown, both the least squares iterative identification method and the method provided in the embodiments of the present invention can achieve a close approximation of the real model during the identification process. The convergence curve shows that the method provided in the embodiments of the present invention has a faster approximation speed and a closer approximation of the real model during the identification process. Figure 14 , 15 The two algorithms were compared and simulated with time delays of 100ms and 150ms respectively. It can be seen that the method provided by the embodiment of the present invention has a faster convergence speed and a relatively smaller rotor angle oscillation amplitude. Compared with the iterative identification method, the method in this paper has a better suppression effect on low-frequency oscillations in time-delay power systems.
[0171] This invention provides a damping controller design method based on a power system. Addressing the impact of model identification errors and communication delays on system damping control caused by wind power integration into the power system, this invention proposes an improved multi-parameter identification differential evolution algorithm-based Smith time-delay compensator and optimal damping controller design method. The method is compared with a recursive least squares iterative identification method, and simulation verification is performed on a four-machine, two-zone system. Results show that, compared to the recursive least squares iterative identification method, the method provided in this invention has a faster convergence speed and higher identification accuracy. The corresponding damping controller design is more reasonable, and the low-frequency oscillation suppression effect is better in multi-interference scenarios involving wind power.
[0172] On the other hand, the present invention also provides a damping controller, comprising: a damping controller designed based on the power system-based damping controller design method described in any one of the above claims.
[0173] On the other hand, the present invention also provides a power system including the damping controller described above.
[0174] In this embodiment of the invention, the power system can be a low-order system or a high-order system.
[0175] The device embodiments described above are merely illustrative. 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 modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.
[0176] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.
[0177] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; 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; and these 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 damping controller design method based on a power system, characterized in that, include: Solve the initial controller model based on the preset power system model; The preset power system model is used as the controlled object transfer function of the damping controller, and a time delay factor is introduced into the controlled object transfer function. The optimized multi-parameter differential evolution identification model is used to identify the parameters to be identified in the transfer function of the controlled object and the time delay factor online to obtain the identification results, which include the identification results of the parameters of the transfer function of the controlled object and the identification results of the time delay factor. The damping controller is designed based on the identification results; The multi-parameter differential evolution identification model optimization method includes: The parameters to be identified and the time delay factor in the transfer function of the controlled object are input into the multi-parameter differential evolution identification model to obtain the identification transfer function containing the time delay factor. Calculate the Vinnicombe distance between the controlled object transfer function and the identification transfer function; If the Vinnicombe distance does not meet the preset recognition accuracy requirement, then the parameters in the multi-parameter differential evolution recognition model are optimized. The optimized multi-parameter differential evolution identification model is used to re-identify the Vinnicombe distance until the Vinnicombe distance reaches the preset identification accuracy requirement, and then the optimized multi-parameter differential evolution identification model is output.
2. The design method for a damping controller based on a power system according to claim 1, characterized in that, The calculation of the Vinnicombe distance between the controlled object transfer function and the identification transfer function includes: Obtain the Vinnicombe distance constraint formula ; When the constraint formula is satisfied, the Vinnicombe distance between the controlled object transfer function and the identification transfer function is expressed as: ; When the constraint formula is not satisfied, the Vinnicombe distance between the controlled object transfer function and the identification transfer function is expressed as 1; In the formula, = , for Number of poles in the right half-plane for Number of poles in the closed right half-plane The number of times the Nyquist curve of the transfer function encircles the origin counterclockwise. yes and The chord distance obtained by projecting onto the unit Riemann sphere.
3. The design method for a damping controller based on a power system according to claim 1, characterized in that, The method for determining whether the Vinnicombe distance meets the preset recognition accuracy requirement includes: Under the premise of stable system output, determine whether the Vinnicombe distance is within a preset error threshold; If so, the Vinnicombe distance is determined to meet the preset recognition accuracy requirement.
4. The design method for a damping controller based on a power system according to claim 1, characterized in that, The step of designing the damping controller based on the identification result includes: The parameter matrix is obtained based on the preset power system model, the initial controller model, and the interference signal; Estimate the system frequency stability margin on the error boundary based on the parameter matrix; When the system frequency stability margin is greater than the maximum value of the Vinnicombe distance, the initial controller model is used as a damping controller.
5. The design method for a damping controller based on a power system according to claim 4, characterized in that, The step of designing the damping controller based on the identification result further includes: When the system frequency stability margin is not greater than the maximum value of the Vinnicombe distance, if ,in If it is a single-input single-output transfer function, then the damping controller is: ; In the formula: , To update the difference between the frequency domain stability margin and the previous stability margin, ρ represents the previous frequency domain stability margin, and ρ is the maximum distance between the transfer functions of the identified model and the target model. This is the initial controller model.
6. The design method for a damping controller based on a power system according to claim 1, characterized in that, The multi-parameter differential evolution identification model includes an adaptive crossover factor and an adaptive mutation factor; The adaptive crossover factor is: ; In the formula, CRmin is the minimum adaptive crossover factor, CRmax is the maximum adaptive crossover factor, Gm is the maximum number of iterations, and G is the number of generations. The adaptive mutation factor is: ; In the formula, Fmin is the minimum adaptive mutation factor, Fmax is the maximum adaptive mutation factor, Gm is the maximum number of iterations, and G is the number of generations.
7. The design method for a damping controller based on a power system according to claim 1, characterized in that, The step of solving the initial controller model based on the preset power system model includes: The parameters of the preset power system model are estimated based on the online identification method for low-frequency oscillation mode noise signals in the power system. The estimation results are tested according to the model order determination criteria. When the estimation results meet the requirements of the model order determination criteria, a reduced-order power system model is obtained. The initial controller model is obtained by solving the closed-loop stability condition corresponding to the reduced-order power system model.
8. A damping controller, characterized in that, The design is based on the power system-based damping controller design method described in any one of claims 1 to 7.
9. An electric power system, characterized in that, Includes the damping controller as described in claim 8.
