Low voltage ride through characteristic identification method of optical storage hybrid inverter

Abstract

The invention provides a low voltage ride through (LVRT) characteristic identification method of an optical storage hybrid inverter. The method comprises the following steps: establishing an LVRT mathematical model of the optical storage hybrid inverter; determining to-be-identified parameters of the LVRT mathematical model, and performing trajectory sensitivity analysis; collecting actual measurement data for the to-be-identified parameters; identifying the to-be-identified parameters by adopting an improved African vulture optimization algorithm; performing error calculation on the identification result, and outputting an optimal identification result; and according to the output optimal identification result, performing parameter updating on the to-be-identified parameter. According to the improved African vulture optimization algorithm based on multi-strategy fusion, the global search capability is improved, and the optimization precision and the optimization stability are enhanced. According to the LVRT characteristic identification method based on the improved algorithm, the precision and efficiency of parameter identification are remarkably improved, the grid-connected stability of the optical storage hybrid inverter during the power grid fault period is greatly enhanced, and the potential influence of adverse factors such as the islanding effect on the power grid is effectively reduced.

Description

Technical Field

[0001] The present invention belongs to the field of control of photovoltaic-storage hybrid inverters, and particularly relates to a method for identifying the low-voltage ride-through characteristics of a photovoltaic-storage hybrid inverter. Background Art

[0002] Under the background of the profound transformation of the global energy pattern today, the development and utilization of renewable energy have become the key measures to achieve sustainable development. The photovoltaic-storage hybrid power generation system has emerged as an important part of building a smart grid, improving energy utilization efficiency and reliability. As the core component of the photovoltaic-storage hybrid power generation system, the accuracy of the model of the photovoltaic-storage hybrid inverter depends on the accuracy of the obtained control parameters. The low-voltage ride-through (LVRT) control parameters, as part of the inverter control parameters, are an important link to ensure the continuous operation of the photovoltaic system. During a power grid fault, if the LVRT control parameter deviation is large and the inverter cannot provide the correct reactive power, it may cause the inverter to malfunction and shut down, resulting in overcurrent on the grid-connected side, overvoltage on the DC side, and even large-scale disconnection from the grid and serious transfer of system power flow. Therefore, accurately obtaining the LVRT control parameters is crucial.

[0003] Currently, the main means to accurately obtain the simulation parameters of new energy systems is to use appropriate algorithms and identify the parameters of the model using measured data. The commonly used methods for obtaining parameters mainly include the theoretical analysis method and the system identification method. In traditional model-based identification methods, in addition to the difficulty of establishing an accurate mathematical model, there is also the problem that it is difficult to accurately determine the model parameters. Moreover, with the continuous development of photovoltaic-storage hybrid inverter technology, new topological structures and control strategies are constantly emerging, further increasing the difficulty of establishing an accurate mathematical model. Although data-driven identification methods can avoid establishing complex mathematical models to a certain extent, the acquisition of a large amount of experimental data is indeed costly. In recent years, identification methods based on artificial intelligence, such as neural networks and support vector machines, have emerged. However, neural networks require a large amount of training data and long training time, and the results are poorly interpretable. Support vector machines are inefficient in processing large-scale data, and the parameter selection has a great impact on the results. In addition, some hybrid identification methods have also been proposed, attempting to combine the advantages of different methods to improve the accuracy and reliability of identification. However, these hybrid methods are often relatively complex, difficult to implement, and their effectiveness and stability need to be further verified in practical applications.

[0004] To address the above problems, the present invention proposes a method for identifying the LVRT characteristics of a photovoltaic-storage hybrid inverter based on an improved African vulture optimization algorithm. By improving the search accuracy and convergence speed of the algorithm and enhancing its ability to handle high-dimensional problems, accurate and efficient identification of the LVRT characteristics of the photovoltaic-storage hybrid inverter is achieved, providing strong technical support for the stable operation of the photovoltaic-storage hybrid system. Summary of the Invention

[0005] The technical problem to be solved by the present invention is: In order to overcome the deficiencies in the prior art, the present invention provides a method for identifying the low-voltage ride-through characteristics of a photovoltaic-storage hybrid inverter. This method conducts sensitivity analysis on the parameters to be identified, uses an improved African vulture optimization algorithm to identify the LVRT transient control parameters of the photovoltaic-storage hybrid inverter, and introduces an adaptive trust coefficient in the global exploration stage. After error analysis of multiple groups of identification results, the optimal value is output. This method significantly improves the performance stability and response speed of the photovoltaic-storage hybrid inverter under LVRT conditions by accurately identifying the transient control parameters.

[0006] The technical solution adopted by the present invention to solve its technical problems is: A method for identifying the low-voltage ride-through characteristics of a photovoltaic-storage hybrid inverter, comprising the following steps:

[0007] S1: Establish an LVRT mathematical model of the photovoltaic-storage hybrid inverter;

[0008] S2: Determine the parameters to be identified in the LVRT mathematical model and conduct trajectory sensitivity analysis;

[0009] S3: Collect measured data for the parameters to be identified;

[0010] S4: Use an improved African vulture optimization algorithm to identify the parameters to be identified;

[0011] S5: Calculate the error of the identification result and output the optimal identification result;

[0012] S6: Update the parameters to be identified according to the output optimal identification result.

[0013] Further, the specific process of establishing the LVRT mathematical model in step S1 is as follows:

[0014] Implement low-voltage transient control on the photovoltaic-storage hybrid inverter, convert the controlled quantity of the reactive power link into the grid-side voltage, and use the reactive power priority control strategy for the grid-side current-limiting link. Then, during the fault duration, the reference values of the active current and reactive current are as follows:

[0015]

[0016] Among them, U termis the per-unit value of the three-phase voltage phasor on the AC side of the photovoltaic-storage hybrid inverter; I d_ref is the reference value of the active current; I q_ref is the reference value of the reactive current; I d_cmd is the active current output by the photovoltaic-storage hybrid inverter; I q_cmd is the reactive current output by the photovoltaic-storage hybrid inverter; K q is the reactive current support coefficient of the LVRT mathematical model; K q(zero) is the zero-voltage crossing reactive current support coefficient; I N is the rated current;

[0017] In the power recovery stage after the fault, establish a mathematical model of the current reference value in the dq coordinate system:

[0018] I d_ref = k id (t - t1)+I d0 , t1 ≤ t < t2 (3)

[0019] I q_ref = k iq (t - t1)+I q0 , t1 ≤ t < t2 (4)

[0020] Among them, k id is the active power recovery rate; k iq is the reactive power recovery rate; I d0 , I q0 are respectively the initial values of the active current and the reactive current of the photovoltaic-storage hybrid inverter at the end of the fault; t1 is the fault clearing time; t2 is the power recovery completion time;

[0021] The above formulas (1), (2), (3) and (4) together constitute the LVRT mathematical model of the photovoltaic-storage hybrid inverter.

[0022] Furthermore, in step S2, determine the parameters to be identified in the LVRT mathematical model and perform trajectory sensitivity analysis. The specific process is as follows:

[0023] According to the LVRT mathematical model determined in step S1, determine K q , k id , k iq , I d0 , I q0 as the parameters to be identified, which jointly determine the dynamic characteristics of the output power after the low-voltage ride-through; and calculate the trajectory sensitivity of the parameters to be identified to analyze the influence degree of the parameters to be identified on the output characteristics of the photovoltaic-storage hybrid inverter. The trajectory sensitivity formula is as follows:

[0024]

[0025] Wherein, is the trajectory sensitivity of parameter θ i ; Δθ i is the perturbation of the i-th parameter; θ i0 is the given value of parameter θ i ; y is one of the observed quantities; m is the number of parameters to be identified, t is the time parameter, and i = 1, 2... m.

[0026] Furthermore, in step S3, the measured data for the parameters to be identified is collected. The specific process is as follows: The hybrid photovoltaic and energy storage inverter adopts direct power control. By changing the active power command value and setting different voltage perturbations, electrical data under multiple working conditions is obtained as the measured data; the electrical data collected under multiple working conditions is the fundamental wave component U of the grid-side voltage and the active power P.

[0027] Furthermore, in step S4, the improved African vulture optimization algorithm is used to identify the parameters to be identified. The specific process is as follows:

[0028] Taking the parameters to be identified K q , k id , k iq , I d0 , I q0 as the position vector of the African vulture population, and the weighted absolute value of the power difference between the fault period and the steady state period under different working conditions is used as the fitness function for identifying the LVRT mathematical model parameters. The specific steps are as follows:

[0029] S4.1: Initialize the population through the Tent-Logistic-Cosine chaotic mapping and the hybrid opposition-based learning strategy:

[0030] Set the number B of the African vulture population to be large enough, preferably taking the value of 50, and the maximum number of iterations J is preferably taken as 100. Substitute the parameters to be identified K q , k id , k iq , I d0 , I q0 into the Tent-Logistic-Cosine chaotic mapping formula in turn. The formula is as follows:

[0031]

[0032] Wherein, r is the control parameter, preferably taking the value of 0.35; X j , X j+1 are the position vectors of the j-th and (j + 1)-th iterations of the African vulture individual, j is the current iteration number, and the value range is [1, J]; perform hybrid opposition-based learning on the position of the African vulture individual after chaotic mapping. The hybrid opposition-based learning formula is as follows:

[0033]

[0034] Among them, MOX j is the mixed opposition number of X j , OX j is the traditional opposition number of X j , lb and ub respectively represent the upper and lower boundaries of the search space, taking 0 and 1, rand1 represents a random number between [0, 1], and as an optimization, the value is 0.3. The distance parameter k changes with the number of iterations, and the calculation formula is:

[0035]

[0036] Among them, k max and k min are the maximum and minimum values of the distance parameter, and J is the maximum number of iterations.

[0037] S4.2: Calculate the fitness value f j of each African vulture in the population. The expression of the fitness value function is as follows:

[0038]

[0039] Among them, P max and P min are respectively the maximum and minimum active power values of the measured data; A i is the absolute value of the difference between the active power and the reactive power during the fault period and the steady state period; C is the number of different working conditions, i = 1, 2... C, and a1 is the fitness coefficient.

[0040] S4.3: Group the African vulture population after initialization

[0041] According to the fitness value of the African vultures, they are divided into 3 groups. The African vulture corresponding to the optimal solution is assigned to the first group, the sub-optimal solution is placed in the second group, and the remaining African vultures are in the third group; the individuals in the third group in the population approach the two optimal individuals in the first two groups, and the behavior is shown in the following formula:

[0042]

[0043]

[0044] Among them, BestMOX 1j and BestMOX 2j respectively represent the position vectors of the optimal vulture and the sub-optimal vulture at the j-th iteration; R j is one of the best vultures at the j-th iteration; L1 takes the value of 0.5; f1 and f2 represent the fitness values of the vultures in the first group and the second group; n represents the total number of vultures in the first group and the second group; p is the grouping coefficient.

[0045] S4.4: African vulture group enters the global exploration stage

[0046] ① Establish the African vulture starvation rate model

[0047] Starving African vultures become particularly aggressive, so they will approach vultures with food. Based on the above behavior, the following mathematical model can be constructed:

[0048]

[0049] where, F represents the starvation rate; g represents the starvation weight; L2 is a random parameter in the interval (0, 1), and preferably takes the value of 0.6 at this time; z represents the starvation coefficient, when the value of z is less than 0, it means starvation, and when it is greater than 0, it tends to be satiated, and preferably takes the value of -1 at this time; h is a random parameter in the interval (-2, 2), and preferably takes the value of 1.5 at this time; ω represents the exploration rate, and its value is j / J; by controlling the dynamic decrease of the starvation rate F, the balance ratio of global exploration and local exploitation in the algorithm iteration process is regulated.

[0050] ② Global exploration stage

[0051] The present invention introduces an adaptive trust coefficient δ j , which inhibits other vultures from approaching the current optimal vulture too early in the early stage of the algorithm, avoids premature local exploitation, and at the same time increases the population diversity in the early stage of the algorithm. When |F|≥1, the algorithm enters the exploration stage, and the vultures randomly search different regions through the following formula:

[0052]

[0053] D j =|L4·R j -MOX j | (16)

[0054]

[0055] where, D j represents the exploration coefficient; L3 and L4 are parameters between the intervals (0, 2), and preferably both take the value of 0.7 at this time; P j represents the position of the vulture after update; e is the natural constant; rand2 and rand3 take the values of 0.2 and 0.3 respectively; the adaptive trust coefficient δ j will start from close to 0 and grow non-linearly with the increase of the number of iterations until it reaches 1 when the algorithm enters the global exploration stage, so that IAVOA has stronger exploration ability in the global exploration stage, thus achieving the balance of exploration and exploitation.

[0056] S4.5: Local exploration stage

[0057] When the value of |F| is between 0.5 and 1, IAVOA enters the first stage of the local exploration phase; in the first stage, two different rotation flight and siege strategies are executed, and the choice of strategy is based on the vulture position P1 after the first iteration. The specific expression is as follows:

[0058]

[0059]

[0060] where both rand4 and rand5 are numbers between (0, 1). Preferably, the values are 0.2 and 0.8 respectively; P j+1 and P j represent the vulture positions after the (j + 1)-th and j-th iterations respectively; S1 and S2 represent the spiral flight modes of the vultures.

[0061] When the value of |F| is less than 0.5, this stage of the algorithm is executed. The actions of two vultures gather several types of vultures on the food source and launch a siege and aggressive struggle for food. Different strategies are selected according to the vulture position P2 after the second iteration. The specific process is as follows:

[0062]

[0063] where A1 is the optimal aggression coefficient and A2 is the sub-optimal aggression coefficient; the Levyflight parameter is used to improve the effectiveness of the algorithm. This parameter is calculated from the Levy flight mode. μ and v are random numbers in the interval (0, 1). Preferably, the values are 0.9 and 0.1 respectively, β is the fixed numerical parameter 1.5, and σ is the flight coefficient.

[0064] Furthermore, in step S5, an error calculation is performed on the identification result, and the optimal identification result is output. The specific process is as follows:

[0065] The position parameter after vulture exploration is used as the output optimal identification result, and the voltage disturbance process is divided into three intervals A, B, and C. The error calculation formula for each interval is as follows:

[0066]

[0067] where E z is the error of each interval, Z is the three intervals A, B, and C, is the per-unit value of active power, and K end and K start are the first and last serial numbers of the measured data within the calculated error interval respectively.

[0068] The upper limit of the weighted average error of the voltage in three error intervals is 0.05, and the upper limit of the error of the reactive power is 0.15. Substitute the optimal identification result and the measured data into Equation (24) for error calculation. If it is within the allowable error range, directly output the optimal identification result; otherwise, update the position of the vulture individual, and determine whether the current iteration number is greater than the maximum iteration number. If not, return to step S4.2; if so, output the optimal identification result.

[0069] Further, in step S6, according to the output optimal identification result, parameter update is performed on the parameter to be identified. The specific process is as follows: The optimal identification parameter value is stored in the DSP control module. After completing the parameter update operation, the system restart is triggered, which greatly enhances the grid connection stability of the photovoltaic-storage hybrid inverter during grid faults and effectively reduces the potential impact of adverse factors such as islanding effect on the grid.

[0070] Further, the photovoltaic-storage hybrid inverter includes an MPPT controller, a bidirectional rectifier-inverter circuit, a bidirectional charge-discharge circuit, a DSP control module, a monitoring module, and a touch screen. The photovoltaic-storage hybrid inverter can perform maximum power tracking on the direct current generated by the photovoltaic module (i.e., photovoltaic panel) through the MPPT controller, and then invert it into alternating current through the bidirectional rectifier-inverter circuit to supply power to the load and the grid; or perform maximum power tracking on the direct current generated by the photovoltaic module through the MPPT controller, and then transfer it into the energy storage battery through the bidirectional charge-discharge circuit for storage. When needed, the electric energy in the energy storage battery is supplied to the load and the grid through the bidirectional charge-discharge circuit and the bidirectional rectifier-inverter circuit. At the same time, the electric energy of the grid can also be rectified through the bidirectional rectifier-inverter circuit and then charge the energy storage battery through the bidirectional charge-discharge circuit. During the whole process, the data of the entire system can be communicated and displayed on the touch screen through the DSP control module and the monitoring module.

[0071] Further, the photovoltaic module adopts a direct current power supply photovoltaic module, and the bidirectional rectifier-inverter circuit includes a three-phase full-bridge converter, a filter inductor L f , a capacitor C f and a static transfer switch STS (static transfer switch, STS). The load is a local load R Load , and the grid includes a grid-side inductor L g , a grid-side resistor R g and the grid Grid. The bidirectional charge-discharge circuit adopts a bidirectional DC-DC converter. When the static transfer switch STS is closed, the photovoltaic-storage hybrid inverter is connected to the main grid, that is, it is connected to the grid Grid through the grid-side inductor L g , the grid-side resistor R g . When STS is disconnected, the system operates in islanding mode.

[0072] The beneficial effects of the present invention are as follows: A method for identifying the low-voltage ride-through characteristics of a photovoltaic-storage hybrid inverter provided by the present invention is improved on the basis of the traditional African vulture optimization algorithm, and an improved African vulture optimization algorithm with multi-strategy fusion (Improved African Vulture Optimization Algorithm, IAVOA) is proposed, which improves the global search ability and enhances the optimization accuracy and optimization stability. Combining the improved algorithm, a method for identifying the LVRT characteristics of a photovoltaic-storage hybrid inverter is proposed, which significantly improves the accuracy and efficiency of parameter identification, thereby greatly enhancing the grid connection stability of the photovoltaic-storage hybrid inverter during grid faults and effectively reducing the potential impact of adverse factors such as islanding effect on the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0074] Figure 1 It is a power generation framework diagram of a photovoltaic-storage hybrid inverter.

[0075] Figure 2 It is a control structure diagram of a photovoltaic-storage hybrid inverter.

[0076] Figure 3 It is a flowchart of parameter identification.

[0077] Figure 4 It is a flowchart of the improved African vulture optimization algorithm with multi-strategy fusion. SPECIFIC EMBODIMENTS

[0078] Now, the present invention will be described in detail with reference to the accompanying drawings. This figure is a simplified schematic diagram, which only illustrates the basic structure of the present invention in a schematic manner, so it only shows the components related to the present invention.

[0079] As Figure 1 shown, the photovoltaic-storage hybrid inverter includes an MPPT controller, a bidirectional rectifier-inverter circuit, a bidirectional charge-discharge circuit, a DSP control module, a monitoring module, and a touch screen. The photovoltaic-storage hybrid inverter can perform maximum power tracking on the direct current generated by the photovoltaic module (i.e., the photovoltaic panel) through the MPPT controller, and then invert it into alternating current through the bidirectional rectifier-inverter circuit to supply power to the load and the power grid; it can also perform maximum power tracking on the direct current generated by the photovoltaic panel through the MPPT controller, and then transfer it into the energy storage battery through the bidirectional charge-discharge circuit for storage, and then supply the power in the energy storage battery to the load and the power grid through the bidirectional charge-discharge circuit and the bidirectional rectifier-inverter circuit when needed. At the same time, the power of the power grid can also be rectified through the bidirectional rectifier-inverter circuit and then charge the energy storage battery through the bidirectional charge-discharge circuit. During the whole process, the data of the entire system can be communicated through the DSP and the detection module and displayed on the touch screen.

[0080] AsFigure 2 As shown, the hybrid PV and energy storage inverter mainly consists of a PV module, a three-phase full-bridge converter, a filter inductor L f and a capacitor C f , a local load R Load , a static transfer switch (STS), a grid-side inductor L g , a grid-side resistor R g and a power grid Grid, etc.

[0081] The PV module uses a DC power PV module. The bidirectional rectifier and inverter circuit includes a three-phase full-bridge converter, a filter inductor L f , a capacitor C f and a static transfer switch STS (static transfer switch), the load is a local load R Load , the power grid includes a grid-side inductor L g , a grid-side resistor R g and a power grid Grid. The bidirectional charge and discharge circuit uses a bidirectional DC-DC converter.

[0082] When the STS is closed, the hybrid PV and energy storage inverter is connected to the main power grid, and the inverter system operates in the grid-connected mode; when the STS is open, the system operates in the island mode. The specific process is as follows:

[0083] First, the DC power PV module generates direct current. During this process, the MPPT technology is used to ensure that the DC power PV module operates at the maximum power point to improve the power generation efficiency. Subsequently, the generated direct current enters the Boost converter for boosting operation to raise the voltage to a level suitable for the inverter to work. Then, the boosted direct current is input to the hybrid PV and energy storage inverter, which converts the direct current into alternating current and ensures the quality of the output AC waveform through a specific control algorithm. The converted alternating current will pass through a filter composed of a filter inductor L f and a capacitor C f for filtering to reduce harmonics. The filtered alternating current is connected to the power grid through the static transfer switch STS. The grid-side inductor L g and the grid-side resistor R gRepresents the equivalent impedance of the power grid. When there is an excess of power generated by the DC power source photovoltaic modules, the excess electrical energy can be stored in the battery through the bidirectional DC–DC converter. This bidirectional DC-DC converter can achieve bidirectional power transmission, that is, the charging and discharging functions. The entire system is equipped with a power calculation module and a voltage-current double closed-loop control module. Among them, the power calculation module is used to accurately calculate the output power of the inverter to ensure that it meets the set requirements; the voltage-current double closed-loop control module uses a feedback mechanism to ensure that the output voltage and current of the inverter are stable and comply with the power grid specifications. In addition, the alternating current output by the inverter can also be directly supplied to the load to ensure a stable power supply for the load.

[0084] The present invention provides a method for identifying the low voltage ride-through characteristics of a photovoltaic-storage hybrid inverter, as Figure 3 shown in the LVRT parameter identification flowchart of the photovoltaic energy storage inverter of this embodiment. The specific steps are as follows:

[0085] S1: Establish an LVRT mathematical model of the photovoltaic-storage hybrid inverter.

[0086] First, implement low-voltage transient control on the photovoltaic-storage hybrid inverter, convert the controlled quantity of the reactive power link into the grid-side voltage, and use the reactive power priority control strategy for the grid-side current limiting link. Then, during the fault duration, the reference values of the active current and reactive current are as follows:

[0087]

[0088] Among them, U term is the per-unit value of the three-phase voltage phasor on the AC side of the photovoltaic-storage hybrid inverter; I d_ref is the reference value of the active current; I q_ref is the reference value of the reactive current; I d_cmd is the output active current of the photovoltaic-storage hybrid inverter; I q_cmd is the output reactive current of the photovoltaic-storage hybrid inverter; K q is the LVRT reactive current support coefficient; K q(zero) is the zero-voltage crossing reactive current support coefficient; I N is the rated current.

[0089] During the power recovery stage after the fault, establish a mathematical model of the current reference value in the dq coordinate system:

[0090] I d_ref = k id (t - t1)+I d0 , t1 ≤ t < t2 (3)

[0091] I q_ref = k iq (t - t1)+I q0 , t1 ≤ t < t2 (4)

[0092] where k id is the active power recovery rate; k iq is the reactive power recovery rate; I d0 , I q0 are respectively the initial values of the active current and the reactive current of the hybrid PV and energy storage inverter at the end of the fault; t1 is the fault clearing time; t2 is the time when the power recovery is completed.

[0093] The above formulas (1), (2), (3) and (4) together constitute the LVRT mathematical model of the hybrid PV and energy storage inverter.

[0094] S2: Determine the parameters to be identified in the LVRT mathematical model and perform trajectory sensitivity analysis.

[0095] According to the LVRT mathematical model determined in step S1, determine K q , k id , k iq , I d0 , I q0 as the parameters to be identified, which jointly determine the dynamic characteristics of the output power after the low voltage ride through. And calculate the trajectory sensitivity of the parameters to be identified to analyze the influence degree of the parameters to be identified on the output characteristics of the hybrid PV and energy storage inverter. The trajectory sensitivity formula is as follows:

[0096]

[0097] where is the trajectory sensitivity of the parameter θ i ; Δθ i is the perturbation of the i-th parameter; θ i0 is the given value of the parameter θ i ; y is one of the observed quantities; m is the number of parameters to be identified; t is the time parameter, i = 1, 2... m.

[0098] S3: Collect the measured data for the parameters to be identified.

[0099] The hybrid PV and energy storage inverter is controlled by direct power control. By changing the active power command value and setting different voltage perturbations, a set of electrical data is obtained as the measured data; the electrical data collected under multiple working conditions are the fundamental component U of the grid-side voltage and the active power P. In this embodiment, by changing the active power command value and setting voltage perturbations, 4 sets of working condition data are obtained. The voltage perturbations are set by the voltage perturbation generating device, which are 0.1U n , 0.4U n , 0.7U n , 0.8U n , where U nis the voltage rating. The active power command values are set to 0.2P N , 0.5P N , 0.6P N , 0.8P N , where P N is the rated active power. The electrical data collected under 4 sets of working conditions are the fundamental component U of the grid-side voltage and the active power P.

[0100] S4: Use the improved African vulture optimization algorithm to identify the parameters to be identified.

[0101] Take the parameters to be identified K q , k id , k iq , I d0 , I q0 as the position vector of the African vulture population, and the weighted absolute value of the power difference during the fault period and the steady state period under different working conditions is used as the fitness function for identifying the LVRT parameters. The specific algorithm steps are as follows:

[0102] S4.1: Initialize the population through Tent-Logistic-Cosine chaotic mapping and hybrid opposition-based learning strategy:

[0103] Set the number B of the African vulture population to be large enough. In this embodiment, B is taken as 50, and the maximum number of iterations J is set to 100. Substitute the parameters to be identified K q , k id , k iq , I d0 , I q0 into the Tent-Logistic-Cosine chaotic mapping formula in turn, as shown below:

[0104]

[0105] where r is the control parameter, with a value of 0.35; X j , X j+1 are the position vectors of the jth and (j + 1)th iterations of the African vulture individual, and j is the current iteration number. Since the maximum number of iterations J is 100, the value range of j is [1, 100]. Perform hybrid opposition-based learning on the position of the African vulture individual after chaotic mapping. The hybrid opposition-based learning formula is as follows:

[0106]

[0107] where MOX j is the hybrid opposition number of X j , and OX j is the opposition number of X jThe traditional opposition number, where lb and ub represent the lower and upper bounds of the search space, taking 0 and 1 respectively, rand1 represents a random number between [0, 1], with a value of 0.3 in this embodiment, and the distance parameter k changes with the number of iterations. The calculation formula is as follows:

[0108]

[0109] where k max and k min are the maximum and minimum values of the distance parameter, j is the current number of iterations, and J is the maximum number of iterations.

[0110] S4.2: Calculate the fitness value f j of each African vulture in the population. The expression of the fitness value function is as follows:

[0111]

[0112] where P max and P min are the maximum and minimum active power values of the measured data respectively; A i is the absolute value of the difference between the active power and reactive power during the fault period and the steady state period; C is the number of different working conditions, i = 1, 2... C, and a1 is the fitness coefficient.

[0113] S4.3: Group the initialized African vulture population

[0114] Group them into 3 groups according to the fitness values of the African vultures. The African vulture corresponding to the optimal solution is assigned to the first group, the sub-optimal solution is placed in the second group, and the remaining African vultures are in the third group. The individuals in the third group of the population approach the two optimal individuals in the first two groups. The behavior is shown in the following formula:

[0115]

[0116] where BestMOX 1j and BestMOX 2j represent the position vectors of the optimal vulture and the sub-optimal vulture at the j-th iteration respectively; R j is one of the best vultures at the j-th iteration; L1 takes a value of 0.5; f1 and f2 represent the fitness values of the vultures in the first and second groups; n represents the total number of vultures in the first and second groups; p is the grouping coefficient.

[0117] S4.4: The African vulture population enters the global exploration stage

[0118] ① Establish an African vulture starvation rate model

[0119] Hungry African vultures become particularly aggressive, so they will approach the vultures with food. Based on the above behavior, the following mathematical model can be constructed:

[0120]

[0121] Among them, F represents the starvation rate; g represents the starvation weight; L2 is a random parameter in the interval (0, 1), and in this embodiment, the value is 0.6 at this time; z represents the starvation coefficient. When the value of z is less than 0, it represents starvation, and when it is greater than 0, it tends to satiety. In this embodiment, the value is -1 at this time; h is a random parameter in the interval (-2, 2), and in this embodiment, the value is 1.5 at this time; ω represents the exploration rate, and the value is j / J. By controlling the dynamic decrease of the starvation rate F, the balance ratio of global exploration and local exploitation in the algorithm iteration process is regulated.

[0122] ② Global exploration stage

[0123] The present invention introduces an adaptive trust coefficient δ j , which inhibits other vultures from approaching the current optimal vulture too early in the early stage of the algorithm, avoids premature local exploitation, and at the same time increases the population diversity in the early stage of the algorithm. When |F|≥1, the algorithm enters the exploration stage, and the vulture randomly searches different regions through the following formula:

[0124]

[0125] D j = |L4·R j -MOX j | (16)

[0126]

[0127] Among them, D j represents the exploration coefficient; L3 and L4 are parameters between the intervals (0, 2), and in this embodiment, both take 0.7 at this time; P j represents the position of the vulture after update; e is the natural constant; rand2 and rand3 take values of 0.2 and 0.3 respectively. The adaptive trust coefficient δ j will start from close to 0 and grow non-linearly with the increase of the number of iterations until it reaches 1 when the algorithm enters the global exploration stage, so that IAVOA has stronger exploration ability in the global exploration stage, thus achieving the balance of exploration and exploitation.

[0128] S4.5: Local exploration stage

[0129] When the value of |F| is between 0.5 and 1, IAVOA enters the first stage of the local exploration stage. In the first stage, two different rotation flight and siege strategies are executed. The selection of the strategy is based on the vulture position P1 after the first iteration. The specific expression is:

[0130]

[0131] Among them, both rand4 and rand5 are numbers between (0, 1). In this embodiment, the values are 0.2 and 0.8; P j+1 and P j respectively represent the vulture positions after the (j + 1)-th and j-th iterations; S1 and S2 represent the spiral flight modes of the vultures.

[0132] When the value of |F| is less than 0.5, this stage of the algorithm is executed. The actions of the two vultures gather several types of vultures at the food source and launch a siege and an aggressive struggle for food. Different strategies are selected according to the vulture position P2 after the second iteration. The specific process is as follows:

[0133]

[0134] Among them, A1 is the optimal aggression coefficient, and A2 is the sub-optimal aggression coefficient; the Levyflight parameter is used to improve the effectiveness of the algorithm. This parameter is calculated from the Levy flight mode. μ and v are random numbers in the interval (0, 1). In this embodiment, the values are 0.9 and 0.1, β is the fixed numerical parameter 1.5, and σ is the flight coefficient.

[0135] S5: Calculate the error of the identification result and output the optimal identification result.

[0136] The position parameter of the vulture after exploration is used as the output optimal identification result, and the voltage disturbance process is divided into three intervals A, B, and C. The error calculation formula for each interval is as follows:

[0137]

[0138] Among them, E z is the error of each interval, Z is the three intervals A, B, and C, is the per-unit value of the active power, K end and K start are respectively the first and last sequence numbers of the measured data within the calculated error interval.

[0139] The upper limit of the weighted average error of the voltage in the three error intervals is 0.05, and the upper limit of the error of the reactive power is 0.15. Substitute the optimal identification result and the measured data into Equation (24) to calculate the error. If it is within the error tolerance range, directly output the optimal identification result. Otherwise, update the vulture individual position and determine whether the current iteration number is greater than the maximum iteration number. If not, return to step S4.2. If so, output the optimal identification result.

[0140] S6: Update the parameter to be identified according to the output optimal identification result.

[0141] The optimal identification parameter values are stored in the DSP chip. After completing the parameter update operation, the system restart is triggered, which greatly enhances the grid connection stability of the photovoltaic energy storage hybrid inverter during grid faults and effectively reduces the potential impact of adverse factors such as islanding effect on the power grid.

[0142] Inspired by the ideal embodiments of the present invention described above, through the above description, relevant staff can make various changes and modifications without departing from the scope of the present invention. The technical scope of this invention is not limited to the content in the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. A method for identifying the low-voltage ride-through characteristics of a hybrid photovoltaic and energy storage inverter, characterized in that: It includes the following steps: S1: Establish the LVRT mathematical model of the photovoltaic-storage hybrid inverter; S2: Determine the parameters to be identified in the LVRT mathematical model and conduct trajectory sensitivity analysis; S3: Collect the measured data for the parameters to be identified; S4: Use the improved African vulture optimization algorithm to identify the parameters to be identified; S5: Calculate the error of the identification result and output the optimal identification result; S6: Update the parameters to be identified according to the output optimal identification result.

2. The low-voltage ride-through characteristic identification method of the photovoltaic and energy storage hybrid inverter according to claim 1, wherein: The specific process of establishing the LVRT mathematical model of the photovoltaic-storage hybrid inverter in step S1 is as follows: Implement low-voltage transient control on the photovoltaic-storage hybrid inverter, convert the controlled quantity of the reactive power link into the grid-side voltage, and use the reactive power priority control strategy for the grid-side current limiting link. During the fault duration, the reference values of the active current and reactive current are as follows: Among them, U term is the per-unit value of the three-phase voltage phasor on the AC side of the hybrid photovoltaic and energy storage inverter; I d_ref is the reference value of the active current; I q_ref is the reference value of the reactive current; I d_cmd is the active current output by the hybrid photovoltaic and energy storage inverter; I q_cmd is the reactive current output by the hybrid photovoltaic and energy storage inverter; K q is the reactive current support coefficient of the LVRT mathematical model; K q(zero) is the zero-voltage crossing reactive current support coefficient; I N is the rated current; In the power recovery stage after the fault, establish the mathematical model of the current reference value in the dq coordinate system: I d_ref = k id (t - t1)+ I d0 , t1 ≤ t < t2 (3) I q_ref = k iq (t - t1)+I q0 , t1 ≤ t < t2 (4) Among them, k id is the active power recovery rate; k iq is the reactive power recovery rate; I d0 and I q0 are respectively the initial active current value and the initial reactive current value of the hybrid PV and energy storage inverter at the end of the fault; t1 is the fault clearing time; t2 is the time when the power recovery is completed; The above formulas (1), (2), (3) and (4) together constitute the LVRT mathematical model of the photovoltaic-storage hybrid inverter.

3. The low-voltage ride-through characteristic identification method of the photovoltaic-storage hybrid inverter according to claim 2, characterized in that: In step S2, determine the parameters to be identified in the LVRT mathematical model and conduct trajectory sensitivity analysis. The specific process is as follows: Determine K according to the LVRT mathematical model determined in step S1 q , k id , k iq , I d0 , I q0 are the parameters to be identified, which jointly determine the dynamic characteristics of the output power after the low-voltage ride-through; and calculate the trajectory sensitivity of the parameters to be identified, and analyze the influence degree of the parameters to be identified on the output characteristics of the photovoltaic-storage hybrid inverter. The trajectory sensitivity formula is as follows: Among them, is the trajectory sensitivity of parameter θ i ; Δθ i is the perturbation of the i-th parameter; θ i0 is the given value of parameter θ i ; y is one of the observed quantities; m is the number of parameters to be identified, t is the time parameter, and i = 1, 2... m.

4. The method for identifying the low voltage ride-through characteristics of the photovoltaic-storage hybrid inverter according to claim 3, characterized in that: The specific process of collecting the measured data for the parameters to be identified in step S3 is as follows: Adopt direct power control for the photovoltaic-storage hybrid inverter, set different voltage disturbances by changing the active power command value, and obtain the electrical data under multiple working conditions as the measured data; The electrical data collected under multiple working conditions are the fundamental wave component U of the grid-side voltage and the active power P.

5. The method for identifying the low-voltage ride-through characteristics of the photovoltaic-storage hybrid inverter according to claim 4, characterized in that: In step S4, use the improved African vulture optimization algorithm to identify the parameters to be identified. The specific process is as follows: Take the parameters K to be identified q , k id , k iq , I d0 , I q0 as the position vector of the African vulture population, and the weighted absolute value of the power difference during different working condition faults and the steady state period is used as the fitness function for identifying the parameters of the LVRT mathematical model. The specific steps are as follows: S4.1: Initialize the population through Tent-Logistic-Cosine chaotic mapping and hybrid opposition-based learning strategy: Set the African vulture population B to be large enough, and the maximum number of iterations to be J. Substitute the parameters K to be identified q , k id , k iq , I d0 , I q0 into the Tent-Logistic-Cosine chaotic mapping formula in turn. The formula is as follows: where r is a control parameter with a value of 0.35; X j and X j+1 are the position vectors of the j-th and (j + 1)-th iterations of the African vulture individual, where j is the current iteration number with a value range of [1, J]; the position of the African vulture individual after chaotic mapping is subjected to hybrid opposition-based learning, and the hybrid opposition-based learning formula is as follows: Among them, MOX j is the mixed opposition number of X j , OX j is the traditional opposition number of X j , lb and ub respectively represent the lower and upper bounds of the search space taking 0 and 1, rand1 represents a random number between [0, 1], and the distance parameter k changes with the number of iterations. The calculation formula is as follows: where k max and k min are the maximum and minimum values of the distance parameter, and J is the maximum number of iterations; S4.2: Calculate the fitness value f of each African vulture in the population j , and the expression of the fitness value function is as follows: Among them, P max and P min are respectively the maximum and minimum active power values of the measured data; A i is the absolute value of the difference between the active power and the reactive power during the fault period and the steady state period; C is the number of different working conditions, i = 1, 2... C, and a1 is the fitness coefficient; S4.3: Group the initialized African vulture population Divide it into 3 groups according to the fitness value of the African vultures. The African vulture corresponding to the optimal solution is assigned to the first group, the sub-optimal solution is placed in the second group, and the remaining African vultures are in the third group; The individuals in the third group within the population approach the two optimal individuals in the first two groups, and the behavior is shown in the following formula: Among them, BestMOX 1j and BestMOX 2j respectively represent the position vectors of the optimal vulture and the sub-optimal vulture at the j-th iteration; R j is one of the best vultures at the j-th iteration; L1 takes a value of 0.5; f1 and f2 represent the fitness values of the first group and the second group of vultures; n represents the total number of vultures in the first group and the second group; p is the grouping coefficient; S4.4: The African vulture group enters the global exploration stage ① Establish the African vulture starvation rate model Starving African vultures become particularly aggressive, so they will approach vultures with food. Based on the above behavior, the following mathematical model can be constructed: Among them, F represents the starvation rate; g represents the starvation weight; L2 is a random parameter in the interval (0, 1); z represents the starvation coefficient, indicating starvation when the value of z is less than 0 and being more inclined to satiety when greater than 0; h is a random parameter in the interval (-2, 2); ω represents the exploration rate, with a value of j / J; By controlling the dynamic decrease of the starvation rate F, the balance ratio between global exploration and local exploitation in the algorithm iteration process is regulated; ② Global exploration stage Introduce the adaptive trust coefficient δ j , and at the same time increase the population diversity in the early stage of the algorithm. When |F|≥1, the algorithm enters the exploration stage, and the vultures randomly search different regions through the following formula: D j = |L4·R j -MOX j | (16) Among them, D j represents the exploration coefficient; L3 and L4 are parameters between the interval (0, 2); P j represents the position of the vulture after update; e is the natural constant; the values of rand2 and rand3 are 0.2 and 0.3 respectively; the adaptive trust coefficient δ j will start from close to 0 and grow non-linearly as the number of iterations increases until it reaches 1 when the algorithm enters the global exploration stage; S4.5: Local exploration stage When the value of |F| is between 0.5 and 1, the IAVOA enters the first stage of the local exploration stage; In the first stage, two different rotational flight and siege strategies are executed, and the strategy selection is based on the vulture position P1 after the first iteration. The specific expression is: where rand4 and rand5 are both numbers between (0, 1); P j+1 and P j represent the vulture positions after the (j + 1)-th and j-th iterations respectively; S1 and S2 represent the spiral flight mode of the vulture; When the value of |F| is less than 0.5, this stage of the algorithm is executed. The actions of the two vultures gather several types of vultures on the food source and launch a siege and an aggressive struggle for food. Different strategies are selected according to the vulture position P2 after the second iteration. The specific process is as follows: Among them, A1 is the optimal aggression coefficient, and A2 is the sub-optimal aggression coefficient; the Levyflight parameter is used to improve the effectiveness of the algorithm. This parameter is calculated by the Levy flight mode. μ and v are random numbers in the interval (0, 1), β is the fixed numerical parameter 1.5, and σ is the flight coefficient.

6. The method for identifying the low-voltage ride-through characteristics of the photovoltaic-storage hybrid inverter according to claim 5, characterized in that: In step S5, the error of the identification result is calculated, and the optimal identification result is output. The specific process is as follows: The position parameter after the vulture exploration is used as the output optimal identification result, and the voltage disturbance process is divided into three intervals A, B, and C. The error calculation formula for each interval is as follows: Among them, E z is the error of each interval, and Z is the three intervals of A, B, and C. is the per-unit value of active power, and K end , K start are respectively the first and last serial numbers of the measured data within the calculated error interval. The upper limit of the weighted average error of the voltage in the three error intervals is 0.05, and the upper limit of the error of the reactive power is 0.

15. The optimal identification result and the measured data are substituted into formula (24) for error calculation. If it is within the allowable error range, the optimal identification result is directly output. Otherwise, the vulture individual position is updated, and it is judged whether the current iteration number is greater than the maximum iteration number. If not, return to step S4.

2. If so, output the optimal identification result.

7. The low-voltage ride-through characteristic identification method of the photovoltaic-storage hybrid inverter according to claim 1, characterized in that: In step S6, according to the output optimal identification result, the parameter to be identified is updated. The specific process is as follows: The optimal identification parameter value is stored in the DSP control module. After completing the parameter update operation, the system restart is triggered, which greatly enhances the grid connection stability of the photovoltaic energy storage hybrid inverter during the grid fault and effectively reduces the potential impact of adverse factors such as the islanding effect on the grid.

8. The low-voltage ride-through characteristic identification method of the photovoltaic-storage hybrid inverter according to claim 1, wherein: The photovoltaic energy storage hybrid inverter includes an MPPT controller, a bidirectional rectifier-inverter circuit, a bidirectional charge-discharge circuit, a DSP control module, a monitoring module, and a touch screen. The photovoltaic energy storage hybrid inverter can perform maximum power tracking on the direct current generated by the photovoltaic module through the MPPT controller, and then invert it into alternating current through the bidirectional rectifier-inverter circuit to supply power to the load and the grid; or perform maximum power tracking on the direct current generated by the photovoltaic module through the MPPT controller, and then transfer it into the energy storage battery through the bidirectional charge-discharge circuit for storage. When needed, the electric energy in the energy storage battery is supplied to the load and the grid through the bidirectional charge-discharge circuit and the bidirectional rectifier-inverter circuit. At the same time, the electric energy of the grid can also be rectified through the bidirectional rectifier-inverter circuit and then charge the energy storage battery through the bidirectional charge-discharge circuit; during the whole process, the data of the entire system can be communicated and displayed on the touch screen through the DSP control module and the monitoring module.

9. The low-voltage ride-through characteristic identification method of the photovoltaic-storage hybrid inverter according to claim 8, characterized in that: The photovoltaic module adopts a DC power photovoltaic module. The bidirectional rectifier-inverter circuit includes a three-phase full-bridge converter, a filter inductor L f , a capacitor C f and a static transfer switch STS. The load is a local load R Load . The power grid includes a grid-side inductor L g , a grid-side resistor R g and the power grid Grid. The bidirectional charge-discharge circuit adopts a bidirectional DC-DC converter. When the static transfer switch STS is closed, the photovoltaic-storage hybrid inverter is connected to the main power grid, that is, it is connected to the power grid Grid through the grid-side inductor L g , the grid-side resistor R g . The inverter system operates in the grid-connected mode. When STS is disconnected, the system operates in the island mode.

Images