Photovoltaic generator set control parameter identification method based on whale optimization algorithm

Abstract

The invention provides a photovoltaic generator set control parameter identification method based on a whale optimization algorithm, and belongs to the technical field of photovoltaic generator set control. Comprising the following steps: S1, carrying out a multi-working-condition test based on a hardware-in-the-loop simulation platform, and recording actually measured data; s2, initializing a to-be-identified parameter by using the measured data, updating the to-be-identified parameter to the identification model, and performing iterative optimization on the to-be-identified parameter based on a whale optimization algorithm to obtain an identification data set; s3, screening the identification data set by using a good point set, and locking key parameters by using correlation analysis; and S4, using Tent mapping to construct a predation strategy of the whale group, and performing step-by-step identification on the key parameters and other parameters to obtain a final control parameter identification result. According to the method, the identification parameters are optimized based on the whale optimization algorithm, the parameter identification efficiency is improved, the basic algorithm is improved through the good point set and Tent mapping, and the adaptability of the control parameters to different working conditions is improved on the basis of improving the identification accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic generator set control, and in particular to a photovoltaic generator set control parameter identification method based on a whale optimization algorithm. Background Art

[0002] With the rapid advancement of the global economy and the continuous increase in human demand for energy, the depletion of traditional energy has become a reality that cannot be ignored. In this context, clean and renewable new energy has become the important task of changing the existing energy structure and promoting sustainable development. Solar energy stands out among many new energy sources with its unique advantages and shows great development potential. However, the rise of photovoltaic grid-connected power generation has also brought many challenges to the power grid system. Problems in power quality, stability and reliability, and power grid efficiency have followed one after another. At the same time, photovoltaic systems may also cause island effects, which have a profound impact on distribution networks.

[0003] In order to effectively deal with these potential problems, a more sophisticated model must be established for photovoltaic generators. The accuracy of the model depends on the accuracy of the parameters. Therefore, parameter identification technology is the basis for ensuring the accuracy of the photovoltaic generator model and the key to ensuring the safe and stable operation of the power grid. Traditional parameter identification methods are often accompanied by high complexity, resulting in low parameter identification efficiency; at the same time, there are certain limitations in the correlation analysis between control parameters and observed quantities, resulting in low accuracy of the identification results and lack of sufficient adaptability. Summary of the invention

[0004] In view of this, the present invention provides a photovoltaic generator control parameter identification method based on the whale optimization algorithm to solve the technical problems of low accuracy, low identification efficiency and lack of sufficient adaptability of traditional parameter identification methods.

[0005] The technical solution adopted by the present invention to solve the technical problem is:

[0006] A method for identifying control parameters of a photovoltaic generator set based on a whale optimization algorithm comprises the following steps:

[0007] S1. Carry out multi-condition tests based on the hardware-in-the-loop simulation platform and record the measured data;

[0008] S2. Initialize the parameters to be identified using the measured data, update the parameters to be identified to the identification model, and iteratively optimize the parameters to be identified based on the whale optimization algorithm to obtain an identification data set;

[0009] S3, use the best point set to screen the identification data set, and use correlation analysis to lock the key parameters;

[0010] S4. Use Tent mapping to construct a multi-whale predation strategy, identify key parameters and other parameters step by step, and obtain the final control parameter identification results.

[0011] Preferably, in step S1, carrying out multi-condition testing based on a hardware-in-the-loop simulation platform includes the following steps:

[0012] S11. Build a hardware-in-the-loop simulation platform for photovoltaic generators based on the RT-LAB real-time simulation platform, and perform hardware-in-the-loop simulation on real photovoltaic controllers;

[0013] S12. Perform low voltage ride-through test of three-phase short circuit fault under different conditions and record the measured data;

[0014] S13. Based on the control strategy and grid-connected structure of the photovoltaic grid-connected inverter, analyze and determine the parameters to be identified and the observed quantities related to the parameters to be identified.

[0015] Preferably, the parameter to be identified is the current inner loop PI parameter k dP , k dI , k qP , k qI , and the voltage outer loop PI parameter k uP , k uI , the observed quantity is u dc 、i d 、i q , P and Q.

[0016] Preferably, in step S2, the parameters to be identified are iteratively optimized based on the whale optimization algorithm to obtain an identification data set, including the following steps:

[0017] S21, preliminarily identifying the parameters to be identified based on the whale optimization algorithm, and searching for the parameters to be identified that meet the working condition observation quantity;

[0018] S22. Establish an identification data set according to the parameters to be identified and the observed quantities.

[0019] Preferably, the preliminary identification of the parameters to be identified based on the whale optimization algorithm includes completing preliminary identification according to three major behaviors of whales: encircling predation, spiral predation, and searching for prey:

[0020] Encirclement and predation:

[0021] When p'<0.5, the whale will tend to adopt the method of encirclement hunting, and the whale n will adjust the position X according to the current best whale position (X)* n , the mathematical formula for its position update can be expressed as:

[0022] X n t+1 =(Xt )*-AD

[0023] Where, X n t is the position of individual n at time t, X n t+1 is the position of individual n at time t+1, (X t )* is the optimal predation position of the whale group at time t, and D is the individual position to the current optimal position (X t )* approximation iteration distance, A and C are optimization coefficients; by adjusting A and C, the corresponding position of the current optimal solution can be reached, and the calculation formula is:

[0024]

[0025] Where R1 and R2 are random numbers in the range [0,1], and T is the maximum time;

[0026] Spiral Predation:

[0027] If p'≥0.5, the whale adopts spiral predation, and its position update method is:

[0028] X n t+1 =X n t +((X t )*-X n t ) bl cos(2πl)

[0029] In the formula, b is a constant that affects the spiral shape, and l is a random number in the range [-1,1];

[0030] Searching for prey:

[0031] X n t+1 =X n,rand t -AD

[0032] D=|CX n,rand t -X n,rand t

[0033] Where, X n,rand t For individual X n A random position at time t.

[0034] Preferably, in step S3, the step of screening the identification data set using the best point set and locking the key parameters using correlation analysis includes the following steps:

[0035] S31, using the good point set to update the recognition data set;

[0036] S32. Use perturbation analysis method to extract key parameters.

[0037] Preferably, the good point set is:

[0038] The dimension of the search space is J, and the set of good points corresponding to the whale group is Y[Y1,…,Y n ,…,Y N ], where Y n It can be expressed as:

[0039] Y n =[Y n1 ,...,Y nj ,...,Y nJ ]

[0040] Y nj =2ncos(2πj / p)

[0041] (p-3) / 2≥J

[0042] Wherein, j (j=1, .., J) represents the dimension, and p is the smallest prime number that satisfies the third formula.

[0043] Preferably, in step S3, extracting key parameters by using the perturbation analysis method includes: setting a random perturbation range β∈(0,1), before and after the observed quantity perturbation, recording the hunger values ​​of individual n as F n and F n ', the key parameters are determined based on:

[0044]

[0045] Preferably, in step S4, using tent mapping to construct a multi-whale predation strategy, identifying key parameters and other parameters step by step, and obtaining the final control parameter identification result includes the following steps:

[0046] S41, construct multiple whale groups using Tent mapping and identify control parameters step by step;

[0047] S42. Record the hunger value of the whale group at time t as F t , the hunger value corresponding to the current optimal identification result (X)* is F*, and it is determined whether F*≥F t ;

[0048] S43. If satisfied, then F*=F t ,(X)*=(X t )*;

[0049] S44, if not satisfied, then F*, (X)* remain unchanged;

[0050] S45, determining whether the number of iterations reaches T; if the number of iterations does not reach T, returning to step S41;

[0051] S46. If the number of iterations reaches T, then output the K corresponding to F* in the multi-whale group n As the control parameter identification result.

[0052] Preferably, the Tent mapping model is as follows, where ε is generally set to 0.5:

[0053]

[0054] (X t )* and (X)* are compared using the following formula to achieve whale group update:

[0055]

[0056] Compared with the prior art, the present invention has the following beneficial effects:

[0057] The present invention is based on a hardware-in-the-loop simulation platform and carries out multi-condition tests, recording the measured data of the real controller under different disturbances as the identification basis and verification basis of the control parameters. Then, based on the whale optimization algorithm, the identification model is used to obtain the observed quantity-control parameter data set of the photovoltaic generator set, which effectively improves the parameter identification efficiency. Then, the identification data set is screened by the good point set to ensure that the preliminary identification results of the parameters are evenly distributed in the search space, which is conducive to the rapid convergence of the identification results; the key parameters affecting the dynamic response characteristics of the photovoltaic generator set are obtained by correlation analysis, and the identification is prioritized, which can reduce the workload of the optimization algorithm, thereby improving the identification accuracy. Finally, the Tent mapping is used to construct a multi-whale predation method to simulate the parallel predation behavior of multiple whale groups in different search spaces, thereby improving the global search ability of the algorithm, avoiding the identification results from falling into the local optimum, thereby improving the adaptability of the identification results to multiple working conditions, and then accurately identifying the control parameters of the grid-connected inverter on the basis of fully considering the correlation of the dynamic response of the photovoltaic generator set. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A flow chart of a method of the present invention.

[0059] Figure 2 This is the double closed-loop control block diagram of the grid-connected inverter of the photovoltaic generator set.

[0060] Figure 3 This is another method flow chart of the present invention.

[0061] Figure 4 Schematic diagram of the individual predation methods of the whale optimization algorithm.

[0062] Figure 5 For photovoltaic generators, the voltage drops to 0.75U N Comparison chart of the measured value and the identified value of the working condition.

[0063] Figure 6 For photovoltaic generators, the voltage drops to 0.50U N Comparison chart of the measured value and the identified value of the working condition.

[0064] Figure 7 For photovoltaic generators, the voltage drops to 0.20U N Comparison chart of the measured value and the identified value of the working condition. DETAILED DESCRIPTION

[0065] The technical solutions and technical effects of the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings of the present invention.

[0066] Please see Figure 1 , a control parameter identification method for a photovoltaic generator set based on a whale optimization algorithm, comprising the following steps:

[0067] S1. Carry out multi-condition tests based on the hardware-in-the-loop simulation platform and record the measured data;

[0068] S2. Initialize the parameters to be identified using the measured data, update the parameters to be identified to the identification model, and iteratively optimize the parameters to be identified based on the whale optimization algorithm to obtain an identification data set;

[0069] S3, use the best point set to screen the identification data set, and use correlation analysis to lock the key parameters;

[0070] S4. Use Tent mapping to construct a multi-whale predation strategy, identify key parameters and other parameters step by step, and obtain the final control parameter identification results.

[0071] The present invention is based on a hardware-in-the-loop simulation platform and carries out multi-condition tests, recording the measured data of the real controller under different disturbances as the identification basis and verification basis of the control parameters. Then, based on the whale optimization algorithm, the identification model is used to obtain the observed quantity-control parameter data set of the photovoltaic generator set, which effectively improves the parameter identification efficiency. Then, the identification data set is screened by the good point set to ensure that the preliminary identification results of the parameters are evenly distributed in the search space, which is conducive to the rapid convergence of the identification results; the key parameters affecting the dynamic response characteristics of the photovoltaic generator set are obtained by correlation analysis, and the identification is prioritized, which can reduce the workload of the optimization algorithm, thereby improving the identification accuracy. Finally, the Tent mapping is used to construct a multi-whale predation method to simulate the parallel predation behavior of multiple whale groups in different search spaces, thereby improving the global search ability of the algorithm, avoiding the identification results from falling into the local optimum, thereby improving the adaptability of the identification results to multiple working conditions, and then accurately identifying the control parameters of the grid-connected inverter on the basis of fully considering the correlation of the dynamic response of the photovoltaic generator set.

[0072] Specifically, in step S1, a photovoltaic power generation group hardware-in-the-loop simulation platform is built based on RT-LAB and multiple operating condition tests are carried out. The measured data of the real controller under different disturbances are recorded, and the observation identification parameters are obtained according to the measured data, which are used as the identification basis and verification basis of the subsequent control parameters. In step S2, the measured data are used to initialize the parameters to be identified, and the parameters are updated to the MATLAB / Simulink identification model for iterative optimization. During each iteration, the whale optimization algorithm (WOA) can optimize the parameters using the output response of the identification model, so as to find a set of parameter results that minimizes the observation error, which is used to establish the training data set. Among them, the operating condition observation u dc 、i d 、i q , P and Q are input features, and the parameter to be identified is the target value.

[0073] Each set of data in the identification data set is a parameter value and a corresponding observation, including the identification error under the parameter. The range of each parameter value is determined based on several sets of data with small identification errors. Using the good point set in step three, multiple populations (i.e., parameters to be identified) can be generated based on this range as the basis of the algorithm in step four. At the same time, the correlation analysis in step three can determine the parameters with high sensitivity to the observation, and they are identified first in step four, thereby improving the efficiency of parameter identification.

[0074] In the specific implementation process, in step S1, carrying out multi-condition testing based on the hardware-in-the-loop simulation platform includes the following steps:

[0075] S11. Build a hardware-in-the-loop simulation platform for photovoltaic generators based on the RT-LAB real-time simulation platform, and perform hardware-in-the-loop simulation on real photovoltaic controllers;

[0076] S12. Conduct low voltage ride-through test of three-phase short circuit fault under different conditions and record the measured data; specifically, the voltage drop degree and initial active power are different.

[0077] S13. Based on the control strategy and grid-connected structure of the photovoltaic grid-connected inverter, analyze and determine the parameters to be identified and the observed quantities related to the parameters to be identified.

[0078] Specifically, hardware-in-the-loop simulation of the real photovoltaic controller is performed based on the RT-LAB real-time simulation platform, and low voltage ride-through test of three-phase short-circuit fault is required under different conditions, where different conditions include different voltage drop levels and different initial active power sizes. Hardware-in-the-loop simulation technology can reflect the control strategy of the real photovoltaic inverter, and at the same time take into account the authenticity of full hardware simulation and the controllability of full digital simulation. It has rich test scenarios and can fully and accurately reflect the grid-connected characteristics of the inverter, and is not affected by weather factors and the external environment.

[0079] The photovoltaic power generation system is mainly composed of a DC-DC boost circuit, a DC support capacitor, a three-phase inverter and a filter circuit. Among them, the maximum power point tracking control of the photovoltaic array mainly depends on the front-stage boost link in the system, and the back-stage DC-AC link, as a key step in photovoltaic power generation grid connection, should meet the frequency and phase requirements required for grid connection to ensure stable grid connection capability and anti-interference performance in a complex power grid environment.

[0080] Please see Figure 2 The photovoltaic grid-connected inverter adopts a dual closed-loop control strategy. In the synchronous rotating coordinate system, the power decoupling control of the photovoltaic unit can be realized through the grid voltage. The voltage outer loop equation and the grid-side voltage dq axis component V dg and V qg It can be expressed as formula (1) to (3), where i q Generally set to 0.

[0081]

[0082] In the formula, u dc and u dc *respectively DC voltage and its reference value, i d 、i q 、i d * and i q * are the dq axis components of the inverter output current and the reference value, respectively, k uP and k uI They are the voltage outer loop PI parameters; V d、V q are the inverter output voltage, V d *、V q * are the reference values ​​of the inverter output voltage, L is the filter inductance, w is the phase-locked loop angular frequency, k dP , k dI , k qP and k qI are the current inner loop PI parameters respectively.

[0083] From the grid-connected structure, we can see that formula (4) can be further expressed as:

[0084]

[0085] From equation (1) and equation (4), we can get d It can be expressed as:

[0086]

[0087] It can be seen that the magnitude of active power P and reactive power Q is respectively determined by i d and i q Influence.

[0088] During normal operation, the photovoltaic inverter gives priority to ensuring active power output to meet power supply needs. During a grid failure, the grid connection point voltage drops. To ensure power supply reliability, the photovoltaic inverter needs to adjust reactive power to support voltage recovery. Therefore, setting an accurate k dP , k dI , k qP , k qI , k uP , k uI It is the key to maintain grid stability. Since the control parameters cannot be obtained directly from the nameplate and the system model is highly nonlinear, it is necessary to use intelligent algorithms for identification and then establish an accurate simulation analysis model. Therefore, the parameter to be identified is the current inner loop PI parameter k dP , k dI , k qP , k qI , and the voltage outer loop PI parameter k uP , k uI , will u dc 、i d 、i q , P and Q are taken as observations to analyze the correlation of control parameters on the dynamic response of photovoltaic generator sets.

[0089] For further information, see Figure 3 In step S2, the parameters to be identified are iteratively optimized based on the whale optimization algorithm to obtain an identification data set, including the following steps:

[0090] S21. Preliminary identification of the parameters to be identified is performed based on the whale optimization algorithm, and the parameters to be identified that meet the working condition observation quantity are searched;

[0091] S22. Establish an identification data set according to the parameters to be identified and the observed quantities.

[0092] Among them, the Matlab / Simulink identification model can fill in the parameters to be identified. The model will output the observed value from the simulation. After comparing it with the measured value, an error will be obtained. The whale optimization algorithm is used to optimize this error so that the parameters are constantly approaching the optimal value. Therefore, during the first comparison, a value needs to be filled in the model as the initial value (initialization) of the parameter to be identified. The initial value is generally selected based on the empirical value. Then, the parameters to be identified are initially identified based on the whale optimization algorithm. The parameter range is locked through multiple comparisons to establish a data set to be identified. For example, when the parameters to be identified are [a, b, c] and [d, e, f], the corresponding errors are the same, and the final result cannot be determined. The purpose of the preliminary identification is to lock the parameter range. For example, the above results show that the parameter ranges are [a, d][b, e][c, f].

[0093] In some implementations, based on the WOA algorithm, the observed quantity X and the parameter to be identified K are expressed as shown in equations (6) to (7):

[0094]

[0095] Where N is the number of whales, X is n is the location information of individual n, K n Can be used as individual n hunger value F n Parameters used to calculate X n The identification errors of each observation in are denoted as u dc,e 、i d,e 、i q,e , P e and Q e . Let the number of multiple working conditions be G, then the whale group error F and the individual error F n The expression is:

[0096]

[0097] Furthermore, the preliminary identification of the parameters to be identified based on the whale optimization algorithm includes completing preliminary identification according to the three major behaviors of whales: encirclement predation, spiral predation and prey search:

[0098] Encirclement and predation:

[0099] Whales rely on two main strategies in the process of foraging: encirclement and spiral. The choice of these two methods is determined by probability p'. Specifically, when p'<0.5, the whale will tend to adopt the encirclement method, and the whale n will adjust the position X according to the current best whale position (X)*. n , the mathematical formula for its position update can be expressed as:

[0100] X n t+1 =(X t )*-AD (9)

[0101] Where, X n t is the position of individual n at time t, X n t+1 is the position of individual n at time t+1, (X t )* is the optimal predation position of the whale group at time t, and D is the individual position to the current optimal position (X t )* approximation iteration distance, A and C are optimization coefficients; Figure 4 It is shown that when dealing with two-dimensional problems, formula (9) can reach different positions around the current optimal solution by adjusting A and C, where the calculation formulas of A and C are:

[0102]

[0103] Where R1 and R2 are random numbers in the range [0,1], and T is the maximum time;

[0104] Spiral Predation:

[0105] If p'≥0.5, the whale adopts spiral predation, and its position update method is:

[0106] X n t+1 =X n t +((X t )*-X n t ) bl cos(2πl) (11)

[0107] In the formula, b is a constant that affects the spiral shape, and l is a random number in the range [-1,1];

[0108] Searching for prey: When searching for prey, whales need to decide whether to enable the random search strategy based on the size of A. When A ≥ 1, the whale group will randomly select a search agent as (X)*, and use the whale position represented by the agent as a benchmark to adjust the position of the fish group, thereby enabling the whale group to break through the current search limitations and explore possible better prey areas, effectively enhancing the global search efficiency of the algorithm. The search model reflecting this process is given by equations (12) to (13):

[0109] X n t+1 =X n,rand t -AD (12)

[0110] D=|CX n,rand t -X n,rand t | (13)

[0111] Where, X n,rand t For individual X n A random position at time t.

[0112] In the specific implementation process, in step S3, the identification data set is screened using the good point set, and the key parameters are locked using correlation analysis, including the following steps:

[0113] S31, using the good point set to update the recognition data set;

[0114] S32. Extract key parameters using perturbation analysis. Correlation analysis is a perturbation analysis method based on observed quantities. Parameters with high sensitivity to errors caused by disturbances are considered key parameters and are identified first.

[0115] The whale's starting optimization position plays a vital role in the algorithm. It is not only related to the accuracy of the solution, but also directly affects the algorithm's running speed. The good point set initialization has a high uniformity of distribution in space, which has become the preferred choice for improving the algorithm's optimization performance. Therefore, this strategy is used to update the identification data set and then select the most valuable control parameter training values. The good point set can improve the individual initialization position of the algorithm, making it more evenly distributed in the search space, avoiding the uneven distribution problem caused by random initialization, and improving the algorithm's utilization of the entire search space.

[0116] The dimension of the search space is J, and the set of good points corresponding to the whale group is Y[Y1,…,Y n ,…,Y N ], where Y n It can be expressed as:

[0117] Y n =[Yn1 ,...,Y nj ,...,Y nJ ] (15)

[0118] Y nj =2ncos(2πj / p) (16)

[0119] (p-3) / 2≥J (17)

[0120] In the formula, j (j=1, .., J) represents the dimension, and p is the smallest prime number that satisfies formula (17).

[0121] In some embodiments, in step S3, extracting key parameters using the perturbation analysis method includes: setting a random perturbation range β∈(0,1). To test the influence of the observed quantity on the parameters to be identified, setting the random perturbation range can significantly improve the interpretability of the variable correlation analysis stage, thereby extracting key parameters. Before and after the observed quantity perturbation, the hunger values ​​of individual n are respectively F n and F n ', the key parameters are determined based on:

[0122]

[0123] That is, under large disturbance conditions, n K n It is the key parameter that needs to be identified first, otherwise it will be identified as a non-key parameter in the second stage.

[0124] In some embodiments, in step S4, using tent mapping to construct a multi-whale predation strategy, identifying key parameters and other parameters step by step, and obtaining the final control parameter identification result includes the following steps:

[0125] S41, construct multiple whale groups using Tent mapping and identify control parameters step by step;

[0126] S42. Record the hunger value of the whale group at time t as F t , the hunger value corresponding to the current optimal identification result (X)* is F*, and it is determined whether F*≥F t ;

[0127] S43. If satisfied, then F*=F t ,(X)*=(X t )*; if the hunger value corresponding to the current optimal identification result (X)* is greater than or equal to the hunger value at time t, F t , indicating that the hunger value at time t is the smallest, that is, the observation error is the smallest, then the optimal identification result is updated to (X t )*, optimal identification result (X t)*The corresponding hunger value is updated to F t ;

[0128] S44, if not satisfied, then F*, (X)* remain unchanged;

[0129] S45, determining whether the number of iterations reaches T; if the number of iterations does not reach T, returning to step S41;

[0130] S46. If the number of iterations reaches T, then output the K corresponding to F* in the multi-whale group n As the control parameter identification result.

[0131] Through Tent mapping, multiple initial whale groups with different optimization coefficients are generated. Each whale group will represent a set of potential optimal solutions for key parameters. When the key parameters are determined, other parameters are further identified. This method can not only effectively balance global and local searches, but also improve identification efficiency by prioritizing the identification of key parameters, thereby improving the adaptability of control parameters to different working conditions.

[0132] In some embodiments, the Tent mapping model is as shown in equation (18), where ε is generally set to 0.5.

[0133]

[0134] (X t )* and (X)* are compared using formula (19) to achieve whale group update.

[0135]

[0136] The present invention proposes a multi-operating condition control parameter identification method for photovoltaic power generation groups based on the whale optimization algorithm. The perturbation analysis method is used to screen the key parameters affecting the observed quantities, the interpretability of the correlation is enhanced, and the basic algorithm is improved through strategies such as the best point set and Tent mapping. The control parameters of the grid-connected inverter can be accurately identified on the basis of fully considering the correlation of the dynamic response of the photovoltaic power generation group.

[0137] Embodiment 1:

[0138] Voltage drop 0.75U based on hardware-in-the-loop simulation platform N The low voltage ride-through test of the three-phase short-circuit fault is obtained. d 、i q At the same time, the control parameters obtained by the photovoltaic power generation group control parameter identification method based on the whale optimization algorithm of the present invention are substituted into the corresponding identification model to calculate i d 、i q Identification value, plotted as Figure 5 The photovoltaic generator set shown in the figure has a voltage drop of 0.75UN Comparison chart of the measured value and the identified value of the working condition.

[0139] Embodiment 2:

[0140] Voltage drop 0.50U based on hardware-in-the-loop simulation platform N The low voltage ride-through test of the three-phase short-circuit fault is obtained. d 、i q At the same time, the control parameters obtained by the photovoltaic power generation group control parameter identification method based on the whale optimization algorithm of the present invention are substituted into the corresponding identification model to calculate i d 、i q Identification value, plotted as Figure 6 The photovoltaic generator set shown in the figure has a voltage drop of 0.50U N Comparison chart of the measured value and the identified value of the working condition.

[0141] Embodiment 3:

[0142] Voltage drop 0.20U based on hardware-in-the-loop simulation platform N The low voltage ride-through test of the three-phase short-circuit fault is obtained. d 、i q At the same time, the control parameters obtained by the photovoltaic power generation group control parameter identification method based on the whale optimization algorithm of the present invention are substituted into the corresponding identification model to calculate i d 、i q Identification value, plotted as Figure 7 The photovoltaic generator set shown in the figure has a voltage drop of 0.20U N Comparison chart of the measured value and the identified value of the working condition.

[0143] pass Figure 5 to Figure 7 It can be seen that the photovoltaic generator control parameter identification method based on the whale optimization algorithm of the present invention can accurately identify the grid-connected inverter control parameters under the condition of considering the correlation of the dynamic response of the photovoltaic generator.

[0144] What is disclosed above is only a preferred embodiment of the present invention, which certainly cannot be used to limit the scope of rights of the present invention. A person skilled in the art can understand that all or part of the processes of the above embodiments and equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims

1. A method for identifying control parameters of a photovoltaic generator set based on a whale optimization algorithm, characterized in that: The following steps are involved: S1. Carry out multi-condition tests based on the hardware-in-the-loop simulation platform and record the measured data; S2. Initialize the parameters to be identified using the measured data, update the parameters to be identified to the identification model, and iteratively optimize the parameters to be identified based on the whale optimization algorithm to obtain an identification data set; S3, use the best point set to screen the identification data set, and use correlation analysis to lock the key parameters; S4. Use Tent mapping to construct a multi-whale predation strategy, identify key parameters and other parameters step by step, and obtain the final control parameter identification results.

2. The photovoltaic generator control parameter identification method based on the whale optimization algorithm according to claim 1 is characterized in that: In step S1, carrying out multi-condition testing based on the hardware-in-the-loop simulation platform includes the following steps: S11. Build a hardware-in-the-loop simulation platform for photovoltaic generators based on the RT-LAB real-time simulation platform, and perform hardware-in-the-loop simulation on real photovoltaic controllers; S12. Perform low voltage ride-through test of three-phase short circuit fault under different conditions and record the measured data; S13, based on the control strategy and grid-connected structure of the photovoltaic grid-connected inverter, analyze and determine the parameters to be identified, and Observations related to the parameters to be identified.

3. The photovoltaic generator control parameter identification method based on the whale optimization algorithm according to claim 2 is characterized in that: The parameter to be identified is the current inner loop PI parameter k dP , k dI , k qP , k qI , and the voltage outer loop PI parameter k uP , k uI , the observed quantity is u dc 、i d 、i q , P and Q.

4. The photovoltaic generator control parameter identification method based on the whale optimization algorithm according to claim 1 is characterized in that: In step S2, the parameters to be identified are iteratively optimized based on the whale optimization algorithm to obtain an identification data set, including the following steps: S21, preliminarily identifying the parameters to be identified based on the whale optimization algorithm, and searching for the parameters to be identified that meet the working condition observation quantity; S22. Establish an identification data set according to the parameters to be identified and the observed quantities.

5. The photovoltaic generator control parameter identification method based on the whale optimization algorithm according to claim 4 is characterized in that: The preliminary identification of the parameters to be identified based on the whale optimization algorithm includes completing the preliminary identification based on the three major behaviors of whales: encirclement predation, spiral predation and prey search: Encirclement and predation: When p'<0.5, the whale will tend to adopt the method of encirclement hunting, and the whale n will adjust the position X according to the current best whale position (X)* n , the mathematical formula for its position update can be expressed as: X n t+1 =(X t )*-AD In the formula, X n t is the position of individual n at time t, X n t+1 is the position of individual n at time t+1, (X t )* is the optimal predation position of the whale group at time t, and D is the individual position to the current optimal position (X t )* approximation iteration distance, A and C are optimization coefficients; by adjusting A and C, the corresponding position of the current optimal solution can be reached, and the calculation formula is: Where R1 and R2 are random numbers in the range [0,1], and T is the maximum time; Spiral Predation: If p'≥0.5, the whale adopts spiral predation, and its position update method is: X n t+1 =X n t +((X t )*-X n t )e bl cos(2πl) In the formula, b is a constant that affects the spiral shape, and l is a random number in the range [-1,1]; Searching for prey: X n t+1 =X n,rand t -AD D=|CX n,rand t -X n,rand t | Where, X n,rand t For individual X n A random position at time t.

6. The photovoltaic generator control parameter identification method based on the whale optimization algorithm according to claim 1 is characterized in that: In step S3, the method of using the good point set to screen the identification data set and using correlation analysis to lock the key parameters includes the following steps: S31, using the good point set to update the recognition data set; S32. Use perturbation analysis method to extract key parameters.

7. The photovoltaic generator control parameter identification method based on the whale optimization algorithm according to claim 6 is characterized in that: The good point set is: The dimension of the search space is J, and the set of good points corresponding to the whale group is Y[Y1,…,Y n ,…,Y N ], where Y n It can be expressed as: AND n =[And n1 ,...,AND nj ,...,AND nJ ] Y nj =2ncos(2πj / p) (p-3) / 2≥J Wherein, j (j=1, .., J) represents the dimension, and p is the smallest prime number that satisfies the third formula.

8. The photovoltaic generator control parameter identification method based on the whale optimization algorithm according to claim 7 is characterized in that: The key parameters extracted by perturbation analysis method include: setting the random perturbation range β∈(0,1), before and after the observation perturbation, the hunger values ​​of individual n are F n and F n ', the key parameters are determined based on:

9. The photovoltaic generator control parameter identification method based on the whale optimization algorithm according to claim 1 is characterized in that: In step S4, tent mapping is used to construct a multi-whale predation strategy, and key parameters and other parameters are identified step by step to obtain the final control parameter identification results, including the following steps: S41, construct multiple whale groups using Tent mapping and identify control parameters step by step; S42. Record the hunger value of the whale group at time t as F t , the hunger value corresponding to the current optimal identification result (X)* is F*, and it is determined whether F*≥F t ; S43. If satisfied, then F*=F t ,(X)*=(X t )*; S44, if not satisfied, then F*, (X)* remain unchanged; S45, determining whether the number of iterations reaches T; if the number of iterations does not reach T, returning to step S41; S46. If the number of iterations reaches T, then output the K corresponding to F* in the multi-whale group n As the control parameter identification result.

10. The photovoltaic generator control parameter identification method based on the whale optimization algorithm according to claim 9 is characterized in that: The Tent mapping model is shown in the following formula, where ε is generally set to 0.5: (X t )* and (X)* are compared using the following formula to achieve whale group update:

Images