Integrated identification method for high-low transition control parameters and electrical parameters of photovoltaic inverter

Abstract

Based on the integration of photovoltaic inverter high and low voltage ride through control parameters and electrical parameters identification method, including: building photovoltaic inverter semi-physical test platform, and under different conditions of high and low voltage ride through test, get the data set required for parameter identification; With the current and power measured value under high and low voltage ride through as the optimization target, the hybrid frog leap algorithm is adopted (the control parameters under corresponding control mode are identified respectively; According to the measured voltage value under high and low voltage ride through as the optimization target, the improved dichotomy search method is adopted to optimize the fault impedance, so that the simulation voltage and the measured voltage tend to be consistent; The PSASP photovoltaic mechanical model is established and the control parameters and electrical parameters identification results are brought in, and the error value of voltage, current and power under corresponding control mode is calculated. This method realizes the batch automatic identification and accurate setting of new energy mechanical and electrical simulation model impedance under voltage drop multi working condition, greatly improves the multi working condition setting precision and data acquisition efficiency of new energy mechanical and electrical simulation model.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic inverter control, and in particular to a method for integrated identification of high and low ride-through control parameters and electrical parameters of photovoltaic inverters. Background Technology

[0002] Power system flow calculations and simulation analyses rely on high-precision simulation models. Photovoltaic (PV) models are categorized into electromagnetic and electromechanical models based on simulation scale. National standards require PV models to simulate faults using fault impedance. Therefore, the simulation accuracy of PV models is affected not only by control parameters but also by the accuracy of fault impedance settings. Due to the nonlinear characteristics of PV ports, fault impedance cannot be directly calculated analytically, making electromechanical modeling of PV inverters a challenging problem where both control parameters and fault impedance (electrical parameters) are unknown. Therefore, to obtain an accurate PV electromechanical simulation model, it is necessary to perform integrated and precise identification of control parameters and fault impedance parameters. Summary of the Invention

[0003] This invention provides an integrated identification method for high and low ride-through control parameters and electrical parameters of photovoltaic inverters. This method enables batch automatic identification and precise setting of impedance in new energy electromechanical simulation models under multiple voltage dip conditions, greatly improving the accuracy of multi-condition setting and data acquisition efficiency of new energy electromechanical simulation models.

[0004] The technical solution adopted in this invention is as follows:

[0005] The method for integrating high and low ride-through control parameters and electrical parameters of photovoltaic inverters includes the following steps: Step 1: Build a semi-physical test platform for photovoltaic inverters and conduct high and low voltage ride-through tests under different operating conditions to obtain the dataset required for parameter identification; Step 2: Using the measured current and power values ​​under high and low voltage ride-through conditions as the optimization target, the hybrid frog leaping algorithm (SFLA) is used to identify the control parameters under the corresponding control modes. Step 3: Based on the measured voltage value under high and low voltage ride-through, the improved binary search method is used to optimize the fault impedance so that the simulated voltage and the measured voltage tend to be consistent. Step 4: Establish the PSASP photovoltaic electromechanical model and input the identification results of control parameters and electrical parameters. Calculate the error values ​​of voltage, current and power under the corresponding control mode and verify that the error is within the required range.

[0006] In step 1, the photovoltaic inverter semi-physical test platform includes components such as the RT-LAB real-time simulation system and host computer, the OP8665 control development board and host computer, and an oscilloscope. The photovoltaic controller HIL test interacts with the current / voltage sampling signal (Analog Output (AO), switching device trigger signal (DI), and status detection signal (DO) of the RT-LAB photovoltaic inverter through three ports: analog input (AI), digital input (DI), and digital output (DO). The controller communicates with its host computer, and the RT-LAB communicates with its host computer via optical fiber.

[0007] The new energy controller hardware-in-the-loop test platform mainly consists of two parts: a converter primary power circuit model and a real converter controller, which are connected via an I / O interface. High and low voltage ride-through tests are performed on this platform under different operating conditions to obtain the dataset needed for parameter identification.

[0008] In step 2, the active power control during low-voltage ride-through mainly includes four modes: steady-state control, specified power, specified current, and control based on the pre-ride-through current. The four main modes are described below: 1) Steady-state control: In this control mode, the current command remains unchanged, and the photovoltaic inverter's control system continues to use the active current I during normal operation. p0 With reactive current value I q0 However, the integrator in the control loop needs to be frozen; (1); In formula (1): Active power command; This is a reactive power command; This is the rated voltage.

[0009] 2) Specified power: Active power during the preset crossing period And combined with the terminal voltage The active current command can be calculated. The calculation formula is as follows: (2); In formula (2): Indicates the active power coefficient of the low-voltage system; This is the setpoint for low-voltage active power. This represents the initial active power.

[0010] 3) Specified Current: Specifies the active current command during the low-voltage ride-through period. The calculation formula is: (3); In formula (3): and All are calculation coefficients for low-voltage active current; This is the setpoint for the low-voltage active current. This is the initial active current.

[0011] 4) Based on pre-crossing current: During low-voltage ride-through, the pre-crossing current is maintained. This control method can be considered as... Take 0, Take 1, Special case when taking 0: (4); The low-voltage ride-through reactive power control strategy includes three main modes: steady-state control, specified power, and specified current. These three modes are described below: (1) Steady-state control: The current command is the reactive current command processed by the freeze integral stage during normal operation. This is intended to maintain the stability of the current command when abnormal voltage ride-through occurs, and to avoid instability caused by sudden command changes, so that the photovoltaic inverter low-voltage ride-through control system can maintain operation under a relatively stable current reference; reactive current command The expression is as follows: (5); In equation (5): t0 is the time when the fault occurs; This is the reactive current command value under normal operating conditions at the instant before the fault. This value is generated by the outer-loop PI regulator, and the integrator is frozen at the moment of the fault. Specifically: (6); The discrete form is: (7); In equations (6) and (7): Reference voltage and measured voltage amount of deviation; and These are the proportional coefficient and the integral coefficient, respectively. This is the accumulated value of the integrator at the moment of failure; Indicates the time when the fault occurred. The voltage deviation value; This indicates that during the integration process (from the moment the fault occurred) arrive The voltage deviation value within a small time period. It is an integral variable; These are the sampled values ​​at the fault time in discrete time.

[0012] (2) Specify power: First specify the reactive power during the low voltage ride-through period. Then, based on the terminal voltage Calculate reactive current command The calculation formula is: (8); In equation (8): The low-voltage reactive power coefficient; This is the setpoint for low-voltage reactive power. Specify the initial reactive current; introduce a low voltage ride-through threshold. At this time, the current command is: (9); In equation (9): and The calculation factor for low-voltage reactive current; This is the low-through reactive current setting value; This is the initial reactive current; The active and reactive power control methods for high-voltage ride-through are similar to those for low-voltage ride-through. However, in the designated current control section of the reactive power control method for high-voltage ride-through, the reactive current command is as follows: (10); In formula (10): This indicates the reactive current command value during high-voltage ride-through. and All are calculation coefficients for high-voltage reactive current. High voltage ride-through threshold; This is the set value for high-voltage reactive current. This is the initial reactive current.

[0013] Step 2 includes: 2.1: In the process of performing the Mixed Frog Leaping Algorithm (SFLA), firstly, within a given feasible region, an initial population model containing P frogs is constructed randomly; each individual in the population, i.e., the i-th frog, corresponds to a potential vector solution to the problem, mathematically represented as X. i =(x i1 ,x i2 ,…,x iD ), where: X i x represents the D-dimensional vector corresponding to the i-th frog; i1 ,x i2 ,…,x iD Each of these represents a set of control parameters to be optimized; the D parameter represents the dimension of the solution. In the specified power control mode, Xi=(K P,LVRT ,Pset,LV ), K P,LVRT P represents the low-voltage active power factor. set,LV This indicates the low-voltage active power setting value; In the specified current control mode, Xi=(K 1,Ip,LV ,K 2,Ip,LV ,I pset,LV ); K 1,Ip,LV and K 2,Ip,LV All are low active power coefficients; I pset,LV This is the setting value for low-voltage active current.

[0014] After completing the initial population model construction, the hybrid frog leaping algorithm (SFLA) needs to calculate the objective function value for all individual frogs. The evaluation calculations include: Define the mean absolute error of voltage U as : (11); In equation (11): This represents the simulated voltage value at the i-th sampling point; The voltage at the i-th sampling point is the measured value; N is the total number of sampling points. This represents the per-unit value of the level error, in pu. This is the reference voltage.

[0015] Similarly, the per-unit value of the mean absolute error of active power P can be defined. The per-unit value of the mean absolute error of reactive power Q Active current Mean absolute error per unit reactive current I q Mean absolute error per unit Furthermore, an objective function containing multiple feature quantities can be obtained. : (12); Obtaining the objective function values ​​of all individuals Then, the individuals in the population are sorted in ascending order according to the magnitude of their objective function values ​​as follows: (13); In equation (13): This represents the objective function value of the individual with the smallest objective function value among all individuals. This represents the objective function value of the second smallest individual; and so on. This represents the maximum objective function value.

[0016] The entire frog population system is then divided into m subpopulations. Let the sorted frog individuals be numbered k=1,2,…,P, where P represents the total number of individuals in the population.

[0017] Then the subgroup number to which the k-th frog belongs It is determined by the following formula: (14) In equation (14): This represents the modulo operation. Each subgroup contains a collection of frog indexes. for: (15) satisfy That is, each subgroup contains n frog individuals.

[0018] 2.2: For each subgroup, the individual with the best performance in the objective function is denoted as... The individual with the worst performance in the objective function is denoted as Meanwhile, within the entire initial population, the individual with the globally optimal solution is denoted as... During the iterative optimization of the algorithm, the focus is on the subgroups within each subgroup. The update operation employs the following strategy for its specific iterative update mechanism: (15); (16); Where j = 1, 2, ..., D; the moving step size. Satisfy constraint condition -D max ≤D j ≤D max , representing the distance moved along the j-th dimension; D max This represents the maximum step length of an individual frog. This represents a random number that is uniformly distributed in the interval [0,1].

[0019] If the new solution obtained after the update Better than the original solution Then use the new solution. Replace the corresponding individual in the original population. Otherwise, use X. g Replace X b Repeated update mechanism: (17); If a better solution is not found after multiple updates, the hybrid frog-jumping algorithm will automatically generate a new random solution and replace the current X with it. w Solve the vector; repeat this update operation until the preset maximum number of iterations is reached to terminate the process.

[0020] (18); In equation (18): This indicates the updated solution; and , respectively, are the lower and upper bound vectors of each dimension; rand is a random vector of the same dimension as the solution vector, and each vector follows a uniform distribution in [0,1]; ⊙ represents element-wise multiplication.

[0021] Then, all individuals are remixed, sorted and divided into several subpopulations according to Equations (13) to (15), and then a local depth search is performed. This process is repeated until the predetermined number of mixing times is reached.

[0022] In step 2, the hybrid frog leaping algorithm (SFLA) is used to calculate the identification error of low voltage ride-through under different control modes. The specific identification strategy is as follows: Step 2.1.1. Data Input Stage: Experimental data were collected under various operating conditions, covering different voltage drop depths and initial active power levels; Step 2.1.2. Data Processing Stage: Extract the rated capacity Sn, sampling period T, and low voltage ride-through threshold of the photovoltaic inverter's low voltage ride-through control system. Key information such as...

[0023] Step 2.1.3. Feature Extraction and Parameter Selection: Based on the steady-state operating data before and after the fault, the average steady-state active current before the fault is extracted. Average terminal voltage before the fault Average active power before the fault The average steady-state active power current during the fault period Average extreme voltage during the fault .

[0024] Preliminary identification was performed using the hybrid frog leaping algorithm (SFLA), and usable parameter sets (K) that met the error range were selected. P,LVRT ,P set,LV ) and (K 1,Ip,LV ,K 2,Ip,LV ,I pset,LV ).

[0025] Step 2.1.4. Pattern Matching and Parameter Optimization: Based on the selected active power control structure, the hybrid leapfrog hopping algorithm (SFLA) is used to control key parameters. and Optimize and identify the parameters. Calculate the theoretical power value using the obtained parameters. and compared with the measured power Compare and define error metrics Used to evaluate model accuracy; (19); In equation (19): Here are the parameters to be identified, where D=2; Let be the active power of the i-th data set before it enters the low-voltage circuit, i=1,2,…,n, where n represents the number of operating conditions; This represents the measured active power under the i-th data set; This represents the theoretical active power under the i-th set of data.

[0026] Step 2.1.5. Based on the specified current calculation formula (3), identify the current using the same method. , and Using the identified parameters, the current calculation value is obtained according to the specified current control mode. Its relationship with the measured value Difference ; (20); In equation (20): Here are the parameters to be identified, where D=3; U ti (i=1,2,…,n) represents the terminal voltage during the low-voltage period of the i-th data set; I p0i (i=1,2,…,n) represents the active current of the i-th data set before it enters the low-voltage circuit; This represents the measurement value of the i-th data set during low-voltage ride-through; This represents the calculated value of the i-th data set.

[0027] Step 2.1.6. Select error threshold as e th =0.1, if e1>e th And e2>e th It is determined that its control method is without additional control.

[0028] Step 2.1.7. If e1 is satisfied <e th And e2 <e th If the control structure with the smaller error is selected as the optimal match, its parameters are used as the identification result.

[0029] Step 2.1.8. If step 2.1.7 determines that it is a specified current control, then it is necessary to determine whether the current value meets the active current before entering the crossing. If the identified parameters meet K... 1,Ip,LV ≈0,K 2,Ip,LV≈1 and I pset,LV If the value is approximately 0, the control mode is based on the initial active current; otherwise, the control mode is based on the specified current.

[0030] The reactive power control mode and parameter identification method is similar to that of active power control, except that it does not include the identification process based on the pre-crossing current control mode.

[0031] The identification methods for active and reactive power control modes and parameters during high-voltage ride-through are the same as those for low-voltage ride-through.

[0032] In step 3, the initial value of the fault impedance is first calculated. This initial value is used as the starting point for the improved binary search algorithm. The calculation method is as follows: The calculation method for low-voltage ride-through impedance is as follows: (twenty one); In equation (21): Indicates the fault impedance during low-level downtime; Indicates the fault reactance under low voltage; Indicates the resistance to low-throughput faults; This indicates the target voltage value during the low-voltage period; Indicates the system impedance during low-pass operation; This represents the fault impedance ratio for low-throughput, where .

[0033] The calculation method for high-voltage ride-through impedance is as follows: (twenty two); In equation (22): Indicates the fault impedance during high-speed traversal; Indicates the fault reactance of high-voltage circuitry; Indicates high-throughput fault resistance; This indicates the target voltage value during high-voltage ride-through. Indicates the system impedance during high-speed penetration; This indicates the fault impedance ratio during high-voltage operation.

[0034] In step 3, the steps for optimizing the fault impedance using the improved binary search algorithm are as follows: Step 3.1: Initialize the search space, initially set the iteration count k=0, based on the target voltage U. f , n In the (pu) scenario, the search space for fault reactance and fault resistance during the initial high and low voltage ride-through is initialized using piecewise functions, i.e., X(0)f and R(0)f: (twenty three); In equation (23): The initial fault reactance value entered for the configuration file; This represents the maximum value of the initial fault reactance search space; This indicates the target voltage value that drops or rises during the crossing.

[0035] To ensure that the initial space covers a reasonable impedance range and to avoid missing the optimal solution due to an excessively small initial range, the following is ensured: (twenty four); In equation (24): This represents the minimum value of the initial fault reactance search space. This represents the maximum value of the initial fault reactance search space; Therefore: (25) In equation (25): This represents the search space for fault reactance during the initial high and low voltage ride-through period. This represents the minimum value of the initial fault reactance search space. This represents the maximum value of the initial fault reactance search space.

[0036] At the same time, in order to satisfy X f >>R f In practice, impedance ratio is commonly used to meet actual needs in engineering. =10. That is: (26); In equation (26): R(0)f,min and R(0)f,max are the minimum and maximum values ​​of the initial fault resistance search space, respectively.

[0037] Therefore: (27) In equation (27): R(0)f,min and R(0)f,max are the search space for the fault resistance during the initial high and low voltage ride-through; R(0)f,min and R(0)f,max are the minimum and maximum values ​​of the initial fault resistance search space, respectively.

[0038] Step 3.2: Uniform Discretization of Multiple Sampling Points: The reactance search space X(k)f of the k-th iteration is uniformly discretized to generate N sets of sampling points, and the sampling step size of the k-th iteration is... With sampling vector The general term formula is: (28); In equation (28): This represents the fault reactance sampling step size in the k-th iteration; This represents the fault reactance value at the i-th sampling point in the k-th iteration; This represents the maximum value of the fault reactance search space in the k-th iteration; This represents the minimum value of the fault reactance search space in the k-th iteration; Indicates the sampling point number; This indicates the total number of sampling points.

[0039] Resistance sampling vector R (k) Generated proportionally from reactance sampling vector: ensuring impedance ratio It conforms to the actual impedance characteristics of the power system.

[0040] Step 3.3: Voltage Deviation Calculation and Optimal Index Location: The fault simulation is performed automatically by calling PSASP using Python, and the voltage response vector of the k-th simulation is obtained. And calculate the voltage relative to the target voltage. deviation vector Optimal sampling point index satisfy: (29); In equation (29): This represents the voltage response vector for the k-th iteration drop or rise. Indicates the target voltage value that drops or rises during the crossing; express and The deviation vector.

[0041] Represents the simulated voltage at all sampling points i. With target voltage The minimum value.

[0042] This index locates the impedance parameter within the current search space that is closest to the target voltage, providing a directional basis for subsequent space shrinkage.

[0043] Step 3.4: Adaptive search space shrinkage: Based on the current voltage With target voltage The size relationship is used to dynamically shrink the search space for the next iteration. ; The contraction strategy is represented by a piecewise function: If the current voltage Below target voltage The impedance needs to be increased to raise the voltage.

[0044] (30); In equation (30): This represents the reactance value of the optimal sampling point in the k-th iteration; In the k-th iteration, the i-th... +1 sampling point reactance value; The optimal sampling point in the k-th iteration; This represents the maximum value of the fault reactance search space in the k-th iteration; This indicates the total number of sampling points.

[0045] If the current voltage Higher than the target voltage The impedance needs to be reduced to lower the voltage.

[0046] (31); In equation (31): In the k-th iteration, the i-th... -1 reactance value at sampling points.

[0047] Step 3.5: Iteration Termination Criterion. The iteration terminates and the optimal fault impedance is output when any of the following logical conditions are met. , ): (32); In equation (32): Indicates whether the termination condition is met in the k-th iteration, and returns True or False; This represents the voltage deviation corresponding to the optimal sampling point in the k-th iteration; Indicates the convergence threshold; Indicates the current iteration number; This represents the maximum allowed number of iterations. The iteration terminates when the condition is met. This indicates that the condition is not met, and the iteration continues.

[0048] The optimal impedance is: (33).

[0049] In equation (33): The optimal fault reactance value was finally identified. This represents the optimal fault resistance value obtained from the identification. This represents the fault reactance value corresponding to the optimal sampling point in the k-th iteration; This represents the fault resistance value corresponding to the optimal sampling point in the k-th iteration.

[0050] In step 4, the error values ​​of voltage, current, and power under the corresponding control mode are calculated as follows: Simulation and test results of the electromechanical transient simulation model for new energy power plants connecting to the power grid were used to calculate U, P, Q, and I. p and Iq The deviation was used to verify the accuracy of the modeling. First, divide the data into segments. , and These are the start and end times of the simulation, respectively. This is the time to clear the fault. The steady-state duration before the fault. The duration of the instantaneous voltage drop phase. To account for the duration of the voltage instantaneous recovery phase, the disturbance process is divided into five intervals: A, B1, B2, C1, and C2. Transient and steady-state deviations should be calculated for each interval separately. For each transient interval, only the average deviation is calculated; for the steady-state interval, both the average and maximum deviations are calculated. The weighted average total deviation between the simulation and experimental data of the electromechanical transient simulation model for new energy power plants connected to the grid is then calculated. (34); In equation (34): and These are the average deviations for the steady-state and transient intervals, respectively. This represents the maximum value of the deviation within the steady-state interval. The weighted average of the average deviations for each time period is calculated. This represents the simulated value of the nth sampling point; This represents the measured value of the nth sampling point.

[0051] and These are the sequence numbers of the beginning and end of each interval, respectively.

[0052] Indicates from K start To K end Take the absolute value of all data points. The maximum value; This represents the index number of the data point, corresponding to the nth sampling point in the time series.

[0053] This represents the steady-state deviation in interval A; This represents the steady-state deviation in interval B; This represents the steady-state deviation in interval C.

[0054] For U, , , and All must satisfy <0.05; For P, Q, I p ,I q All must meet <0.1、 <0.2、 <0.15、 <0.15.

[0055] This invention provides an integrated identification method for high and low ride-through control parameters and electrical parameters of photovoltaic inverters, with the following technical advantages: 1) This invention describes each control method in detail and takes into account both high-voltage and low-voltage ride-through scenarios; 2) This invention utilizes voltage dip impedance automatic identification technology, which not only saves time but also enables accurate identification of dip impedance. It enables batch automatic identification and accurate setting of impedance in new energy electromechanical simulation models under multiple voltage dip conditions, greatly improving the accuracy of multi-condition setting and data acquisition efficiency of new energy electromechanical simulation models. 3) The present invention calculates electrical parameters under no-load conditions as the initial values ​​for optimization in order to ensure that the algorithm can accurately simulate the behavior of the power grid under different voltage conditions; the ratio of the initial resistance to the reactance is set to 1:10 in the algorithm, which is consistent with the actual line and helps to simulate the voltage drop and current flow in the actual power grid. Attached Figure Description

[0056] The present invention will be further described below with reference to the accompanying drawings and examples; Figure 1 This is a schematic diagram of the low-voltage ride-through active control mode and parameter identification process of the present invention.

[0057] Figure 2 This is a diagram showing the identification results of the low-voltage ride-through control parameters of the present invention.

[0058] Figure 3 The above are simulation waveforms of high voltage ride-through under different control methods according to the present invention.

[0059] Figure 4 This is a diagram showing the identification results of the high-voltage ride-through control parameters of the present invention.

[0060] Figure 5 This is a simulation model of the new energy unit of the present invention.

[0061] Figure 6 This is a graph showing the relationship between the drop voltage and the drop impedance of the present invention.

[0062] Figure 7 The above are simulation diagrams of the photovoltaic inverter of the present invention under different active power conditions.

[0063] Figure 8 This is a flowchart of the photovoltaic inverter drop impedance identification process of the present invention.

[0064] Figure 9 A schematic diagram of the five intervals for error calculation.

[0065] Figure 10 The figure shows a comparison of simulation results and actual measurements of the photovoltaic inverter low-through fault impedance identification technology in this embodiment of the invention under the conditions of P=0.9Pn, U=0.5Un, and three-phase low-through fault 0.15s (condition 1). Figure 11 The figure shows a comparison of simulation results and actual measurements of the photovoltaic inverter low-through fault impedance identification technology in this embodiment of the invention under the conditions of P=0.9Pn, U=0.35Un, and three-phase low-through 0.92s (condition 2). Figure 12 The figure shows a comparison of simulation results and actual measurements of the photovoltaic inverter low-through fault impedance identification technology in this embodiment of the invention under the conditions of P=0.9Pn, U=0.5Un, and three-phase low-through 1.205s (condition 3). Figure 13 The figure shows a comparison of simulation results and actual measurements of the photovoltaic inverter low-through fault impedance identification technology in this embodiment of the invention under the conditions of P=0.9Pn, U=0.75Un, and three-phase low-through 1.705s (condition 4). Figure 14 The figure shows a comparison of simulation results and actual measurements of the photovoltaic inverter low-through fault impedance identification technology in this embodiment of the invention under the conditions of P=0.2Pn, U=0.05Un, and three-phase low-through fault 0.15s (condition 5). Figure 15 The figure shows a comparison of simulation results and actual measurements of the photovoltaic inverter low-through fault impedance identification technology in this embodiment of the invention under the conditions of P=0.2Pn, U=0.2Un, and three-phase low-through 0.625s (condition six). Figure 16 The figure shows a comparison of simulation results and actual measurements of the photovoltaic inverter low-through fault impedance identification technology in this embodiment of the invention under the conditions of P=0.2Pn, U=0.5Un, and three-phase low-through 1.205s (condition 7). Figure 17 The figure shows a comparison between the simulation results and actual measurements of the photovoltaic inverter low-through fault impedance identification technology in this embodiment of the invention under the conditions of P=0.2Pn, U=0.75Un, and three-phase low-through 1.705s (condition 8). Figure 18 The figure shows a comparison of simulation results and actual measurements of the photovoltaic inverter high-throughput fault impedance identification technology in this embodiment of the invention under the conditions of P=0.9Pn, U=1.2Un, and three-phase low-throughput 10s (condition nine). Figure 19 The figure shows a comparison of simulation results and actual measurements of the photovoltaic inverter high-throughput fault impedance identification technology in this embodiment of the invention under the conditions of P=0.9Pn, U=1.3Un, and three-phase low-throughput 0.5s (condition 10). Figure 20The figure shows a comparison of simulation results and actual measurements of the photovoltaic inverter high-throughput fault impedance identification technology in this embodiment of the invention under the conditions of P=0.2Pn, U=1.2Un, and three-phase low-throughput 10s (condition 11). Figure 21 The figure shows a comparison between the simulation results and actual measurements of the photovoltaic inverter high-through-fault impedance identification technology in this embodiment of the invention under the conditions of P=0.2Pn, U=1.3Un, and three-phase low-through-fault 0.5s (condition 12). Detailed Implementation

[0066] Based on an integrated identification method for high and low voltage ride-through control parameters and electrical parameters of photovoltaic inverters, the system includes a control parameter identification module and an electrical parameter (fault impedance) identification module. First, a semi-physical test platform for the photovoltaic controller is built using RT-LAB, and high and low voltage ride-through tests are conducted under different operating conditions to obtain the dataset required for parameter identification. Second, using the measured current and power values ​​under high and low voltage ride-through conditions as optimization targets, the frog-leaping algorithm is used to identify the control parameters for the corresponding control modes. Then, based on the measured voltage values ​​under high and low voltage ride-through conditions as optimization targets, an improved binary search method is used to optimize the fault impedance, making the simulated voltage tend to be consistent with the measured voltage. Finally, a PSASP photovoltaic electromechanical model is established and the identification results of the control and electrical parameters are input. The effectiveness of the proposed method is demonstrated by calculating the error values ​​of voltage, current, and power under the corresponding control modes, all of which are within the required error range.

[0067] Verification Example: To verify the effectiveness and adaptability of the method of the present invention in identifying control modes and control parameters, the specific steps are as follows: Figure 1 As shown.

[0068] Step 1: Determine the control method and parameters for low-voltage ride-through based on the error between power and current. The identification process for the low-voltage ride-through control method and parameters is as follows: Figure 1 As shown.

[0069]

[0070] In the PSASP platform, both active and reactive power control during low-voltage surges are implemented using a "specified current" control strategy. Eight sets of simulation results were obtained by varying the drop depth and initial active power. The control structure and parameters were then inverted using the method described in this invention, and the identification results are as follows: Figure 2 As shown, the maximum error of the parameter identification results is within 0.05%, indicating that accurate control modes and control parameters have been obtained.

[0071] The identification methods for active and reactive power control modes and parameters during high-voltage ride-through are the same as those for low-voltage ride-through. Taking the high-voltage ride-through reactive power control mode of the inverter as an example, the parameters to be identified are shown in Table 2:

[0072] Set the rated power P of the unit in PSASP n to 0.2 p.u., the sampling interval T is 0.001 s. Under the three-phase symmetrical condition, the high voltage ride-through threshold U Hin is 1.1 p.u., the boosting voltage U t is 1.2 p.u. Using the proposed identification method, in the "mode identification" link, taking reactive power control as an example, under the specified current control mode, e1 is 0.0007, and under the specified power control mode, e2 is 0.0725. Since e1 < 0.1 and e2 < 0.1, the no additional control mode is not satisfied. Also, since e1 < e2, this mode is the specified current control mode, and the simulation waveform is as Figure 3 shown.

[0073] The parameter identification results are as Figure 4 shown. The active power control mode is specified power, and the reactive power control mode is specified current. The identification errors of the high voltage ride-through active and reactive power control parameters are both below 1%.

[0074] Step 2: Establish an electromechanical model in which the new energy unit is connected to the equivalent external power grid through a single transformer and a section of impedance. The equivalent external power grid is an infinite voltage source, as Figure 5 shown.

[0075] Set the active power of the new energy unit to 0.9 p.u. and the reactive power to 0. Adjust the voltage of bus 72 on the equivalent external power grid side to control the terminal voltage to 1.0 p.u. Figure 6 Give the relationship between the voltage drop U f and the drop impedance X f when a three-phase short circuit occurs at bus S ( = 10). It can be seen from the figure that U f and X f are not strictly linear. The smaller X f is, the more severe the voltage drop is. When a short circuit fault occurs and X f = 0, the voltage drops suddenly from 1.0 pu to 0.0663 p.u., and the drop amplitude is 93.367%; when a minor fault occurs and X f = 5.45 p.u., the voltage only drops from 1.0 p.u. to 0.8994 p.u., and the drop amplitude is only 10.056%, which conforms to the circuit principle that the smaller the fault reactance, the larger the short circuit current and the lower the voltage division.

[0076] The experiment keeps X f=0.05pu, Q=0, only let P vary between 0.2 and 0.82pu; the results show that increasing or decreasing P only slightly adjusts the steady-state voltage and current amplitude, and the effect on the drop depth is negligible. Therefore, Z can be... f Treating the voltage as an independent variable, an improved binary search method is used for rapid approximation to ensure that the simulated voltage matches the test voltage of RT-LAB. Simulation graphs of the photovoltaic inverter under different active power are shown below. Figure 7 As shown.

[0077] Based on this principle, the fault impedance under no-load conditions is calculated as the initial value for the improved bisection method optimization. Through continuous iteration, the simulated voltage and the measured voltage are made to be consistent. The photovoltaic inverter identification flowchart is as follows. Figure 8 As shown.

[0078] Step 3: Establish a PSASP photovoltaic electromechanical model and input the identification results of control and electrical parameters to calculate the error values ​​of voltage, current, and power under the corresponding control mode. The steady-state and transient data segmentation for U is shown below. Figure 9 As shown.

[0079]

[0080] Due to space limitations, 12 operating conditions were selected, including low voltage (conditions 1-8), high voltage (conditions 9-12) under high power (P0≥0.9pu) and low power (0.1pu≤P0≤0.3pu) conditions, as shown in Table 4, and the deviation of the B interval of the model was calculated.

[0081]

[0082] Substituting the high-voltage and low-voltage ride-through control parameter identification results into the electromechanical transient simulation model, a three-phase symmetrical fault occurs at t1s, where t is the low-voltage ride-through time. The simulation obtains the terminal voltage U and reactive current I. q The response curves of active power P and reactive power Q. Figures 10-21 This is a comparison chart of simulation and test results.

[0083] Figures 10-21 It can be seen that under low voltage ride-through fault conditions, condition 1 has a larger error because the 5% pu voltage is close to a short-circuit fault, and the active power P=0.9pu is close to full power, resulting in a large fault current. The control parameters cannot adjust in time, leading to a large current surge during the fault initiation and recovery phases. The same explanation applies to condition 5. In comparison... Figure 14The impact amplitude is relatively small because the active power P=0.2pu in condition 5 is less than the active power P=0.9pu in condition 1. Under high voltage ride-through conditions, the power factor (PF) in condition 10 has a large impact during the fault initiation and recovery phases because the unit power control does not respond in time after the grid connection voltage U is suddenly raised, resulting in a sudden change in PF. Subsequently, the active power P recovers to its steady-state value. The active power P and active current Ip in condition 12 have large errors because the active power and active current control parameters of the actual controller are not optimized under high voltage ride-through conditions, and the controller cannot control P and Ip in time at the beginning of the high voltage ride-through fault.

[0084] The deviations of each electrical quantity in section B under 12 working conditions were calculated, and the results are shown in Tables 5 to 16. It can be seen that the deviations under each working condition are all less than the maximum allowable deviations specified in Table 3.

[0085]

[0086]

[0087]

[0088]

[0089]

[0090]

[0091]

[0092]

[0093]

[0094]

[0095]

[0096] .

Claims

1. A method for integrated identification of high and low ride-through control parameters and electrical parameters of photovoltaic inverters, characterized in that... Includes the following steps: Step 1: Build a semi-physical test platform for photovoltaic inverters and conduct high and low voltage ride-through tests under different operating conditions to obtain the dataset needed for parameter identification; Step 2: Using the measured current and power values ​​under high and low voltage ride-through as the optimization target, the hybrid frog leaping algorithm (SFLA) is used to identify the control parameters under the corresponding control modes. Step 3: Based on the measured voltage value under high and low voltage ride-through, the optimization target is to use the improved binary search method to optimize the fault impedance so that the simulated voltage and the measured voltage tend to be consistent. Step 4: Establish the PSASP photovoltaic electromechanical model and input the identification results of control parameters and electrical parameters to calculate the error values ​​of voltage, current and power under the corresponding control mode.

2. The method for integrated identification of high and low ride-through control parameters and electrical parameters of photovoltaic inverters according to claim 1, characterized in that: In step 2, the active power control during low-voltage ride-through includes four modes: steady-state control, specified power, specified current, and control based on the pre-ride-through current. 1) Steady-state control: In this control mode, the current command remains unchanged, and the photovoltaic inverter control system continues to use the active current I during normal operation. p0 With reactive current value I q0 However, the integrator in the control loop needs to be frozen; （1）； In formula (1): Active power command; This is a reactive power command; Rated voltage; 2) Specified power: Active power during the preset crossing period And combined with the terminal voltage The active current command can be calculated. The calculation formula is as follows: （2）； In formula (2): Indicates the active power coefficient of the low-voltage system; This is the setpoint for low-voltage active power. This represents the initial active power. 3) Specified Current: Specifies the active current command during the low-voltage ride-through period. The calculation formula is: （3）； In formula (3): and All are calculation coefficients for low-voltage active current; This is the setpoint for the low-voltage active current. This is the initial active current; 4) Based on pre-crossing current: During low-voltage ride-through, the pre-crossing current is maintained. This control method can be considered as... Take 0, Take 1, Special case when taking 0: （4）。 3. The method for integrated identification of high and low ride-through control parameters and electrical parameters of photovoltaic inverters according to claim 2, characterized in that: The low-voltage ride-through reactive power control strategy includes three modes: steady-state control, specified power control, and specified current control. (1) Steady-state control: The current command is the reactive current command processed by the freeze integral element during normal operation. The expression is as follows: （5）； In equation (5): t0 is the time when the fault occurs; The reactive current command value is generated by the outer loop PI regulator at the instant before the fault, and the integrator is frozen at the moment of the fault. Specifically: （6）； The discrete form is: （7）； In equations (6) and (7): Reference voltage and measured voltage amount of deviation; and These are the proportional coefficient and the integral coefficient, respectively. This is the accumulated value of the integrator at the moment of failure; Indicates the time when the fault occurred. The voltage deviation value; This indicates that during the integration process (from the moment the fault occurred) arrive The voltage deviation value within a small time period. It is an integral variable; These are sampled values ​​at the fault time in discrete time. (2) Specify power: First specify the reactive power during the low voltage ride-through period. Then, based on the terminal voltage Calculate reactive current command The calculation formula is: （8）； In equation (8): The low-voltage reactive power coefficient; This is the setpoint for low-voltage reactive power. Specify the initial reactive current; introduce a low voltage ride-through threshold. At this time, the current command is: （9）； In equation (9): and The calculation factor for low-voltage reactive current; This is the low-through reactive current setting value; This is the initial reactive current; In the designated current control section of the high-voltage ride-through reactive power control method, the reactive current command is: （10）； In formula (10): This indicates the reactive current command value during high-voltage ride-through. and All are calculation coefficients for high-voltage reactive current. High voltage ride-through threshold; This is the set value for high-voltage reactive current. This is the initial reactive current.

4. The method for integrated identification of high and low ride-through control parameters and electrical parameters of photovoltaic inverters according to claim 3, characterized in that: In step 2, during the operation of the hybrid frog leaping algorithm (SFLA), firstly, within the given feasible domain, an initial population model containing P frogs is constructed in a random manner. Each individual in the population, i.e., the i-th frog, corresponds to a potential vector solution to the problem, mathematically represented as X. i =(x i1 ,x i2 ,…,x iD ), where: X i x represents the D-dimensional vector corresponding to the i-th frog; i1 ,x i2 ,…,x iD Each of these represents a set of control parameters to be optimized; the D parameter represents the dimension of the solution. In the specified power control mode, Xi=(K P,LVRT ,P set,LV ), K P,LVRT P represents the low-voltage active power factor. set,LV This indicates the low-voltage active power setting value; In the specified current control mode, Xi=(K 1,Ip,LV ,K 2,Ip,LV ,I pset,LV ); K 1,Ip,LV and K 2,Ip,LV All are low active power coefficients; I pset,LV This is the setting value for low-voltage active current.

5. The method for integrated identification of high and low ride-through control parameters and electrical parameters of photovoltaic inverters according to claim 4, characterized in that: After completing the initial population model construction, the Mixed Frog Leaping Algorithm (SFLA) evaluates the objective function value of all frog individuals. The evaluation calculations include: Define the mean absolute error of voltage U as : （11）; In equation (11): This represents the simulated voltage value at the i-th sampling point; The voltage at the i-th sampling point is the measured value; N is the total number of sampling points. This represents the per-unit value of the level error; The reference voltage; Similarly, the per-unit value of the mean absolute error of active power P is defined. The per-unit value of the mean absolute error of reactive power Q Active current Mean absolute error per unit reactive current I q Mean absolute error per unit Furthermore, the objective function containing multiple features is obtained. : （12）； Obtaining the objective function values ​​of all individuals Then, the individuals in the population are sorted in ascending order according to the magnitude of their objective function values ​​as follows: （13）； In equation (13): This represents the objective function value of the individual with the smallest objective function value among all individuals. This represents the objective function value of the second smallest individual; and so on. This represents the maximum objective function value; The entire frog population system is then divided into m subpopulations. Let the sorted frog individuals be numbered k=1,2,…,P, where P represents the total number of individuals in the population. Then the subgroup number to which the k-th frog belongs It is determined by the following formula: （14） In equation (14): This represents the modulo operation; the collection of frog indexes contained in each subgroup. for: （15） satisfy That is, each subgroup contains n frog individuals.

6. The method for integrated identification of high and low ride-through control parameters and electrical parameters of photovoltaic inverters according to claim 5, characterized in that: For each subgroup, the individual with the best performance in the objective function is denoted as... The individual with the worst performance in the objective function is denoted as Simultaneously, within the entire initial population, the individual with the globally optimal solution is denoted as... During the iterative optimization process of the algorithm, for each subgroup... The update operation employs the following strategy for its specific iterative update mechanism: （15）； （16）； Where j = 1, 2, ..., D; the moving step size. Satisfy constraint condition -D max ≤D j ≤D max , representing the distance moved along the j-th dimension; D max This represents the maximum step length of an individual frog. This represents a random number that is uniformly distributed in the interval [0,1]. If the new solution obtained after the update Better than the original solution Then use the new solution. Replace the corresponding individual in the original population; otherwise, use X. g Replace X b Repeated update mechanism: （17）； If a better solution is not found after multiple updates, the hybrid frog-jumping algorithm will automatically generate a new random solution and replace the current X with it. w Solve the vector; repeat this update operation until the preset maximum number of iterations is reached to terminate the process. （18）； In equation (18): This indicates the updated solution; and These are the lower and upper bound vectors of the respective dimensions; rand is a random vector with the same dimension as the solution vector, and each vector follows a uniform distribution in [0,1]; ⊙ represents element-wise multiplication; Then, all individuals are remixed, sorted and divided into several subpopulations according to Equations (13) to (15), and then a local depth search is performed. This process is repeated until the predetermined number of mixing times is reached.

7. The method for integrated identification of high and low ride-through control parameters and electrical parameters of photovoltaic inverters according to claim 6, characterized in that: In step 2, the hybrid frog leaping algorithm (SFLA) is used to calculate the identification error of low voltage ride-through under different control modes. The specific identification strategy is as follows: Step 2.1.

1. Data Input Stage: Experimental data were collected under various operating conditions, covering different voltage drop depths and initial active power levels; Step 2.1.

2. Data Processing Stage: Extract the rated capacity Sn, sampling period T, and low voltage ride-through threshold of the photovoltaic inverter's low voltage ride-through control system. Key information; Step 2.1.

3. Feature Extraction and Parameter Selection: Based on the steady-state operating data before and after the fault, the average steady-state active current before the fault is extracted. Average terminal voltage before the fault Average active power before the fault The average steady-state active power current during the fault period Average extreme voltage during the fault ; Preliminary identification was performed using the hybrid frog leaping algorithm (SFLA), and usable parameter sets (K) that met the error range were selected. P,LVRT ,P set,LV ) and (K 1,Ip,LV ,K 2,Ip,LV ,I pset,LV ); Step 2.1.

4. Pattern Matching and Parameter Optimization: Based on the selected active power control structure, the hybrid leapfrog hopping algorithm (SFLA) is used to control key parameters. and Optimize identification; Calculate the theoretical power value using the obtained parameters. and compared with the measured power Compare and define error metrics Used to evaluate model accuracy; （19）； In equation (19): Here are the parameters to be identified, where D=2; Let be the active power of the i-th data set before it enters the low-voltage circuit, i=1,2,…,n, where n represents the number of operating conditions; This represents the measured active power under the i-th data set; This represents the theoretical active power under the i-th set of data; Step 2.1.

5. Based on the specified current calculation formula (3), identify the current using the same method. , and Using the identified parameters, the current calculation value is obtained according to the specified current control mode. Its relationship with the measured value Difference ; （20）； In equation (20): Here are the parameters to be identified, where D=3; U ti (i=1,2,…,n) represents the terminal voltage during the low-voltage period of the i-th data set; I p0i (i=1,2,…,n) represents the active current of the i-th data set before it enters the low-voltage circuit; This represents the measurement value of the i-th data set during low-voltage ride-through; This represents the calculated value of the i-th data set; Step 2.1.

6. Select error threshold as e th =0.1, if e1>e th And e2>e th Therefore, its control method is determined to be without additional control. Step 2.1.

7. If e1 is satisfied <e th And e2 <e th If the control structure with the smaller error is selected as the optimal match, its parameters are used as the identification result. Step 2.1.

8. If step 2.1.7 determines that it is a specified current control, then it is necessary to determine whether the current value meets the active current before entering the crossing. If the identified parameters meet K... 1,Ip,LV ≈0,K 2,Ip,LV ≈1 and I pset,LV If the value is approximately 0, the control mode is based on the initial active current; otherwise, the control mode is based on the specified current.

8. The method for integrated identification of high and low ride-through control parameters and electrical parameters of photovoltaic inverters according to claim 7, characterized in that: In step 3, the initial value of the fault impedance is first calculated. This initial value is used as the starting point for the improved binary search algorithm. The calculation method is as follows: The calculation method for low-voltage ride-through impedance is as follows: （21）； In equation (21): Indicates the fault impedance during low-level downtime; Indicates the fault reactance under low voltage; Indicates the resistance to low-throughput faults; This indicates the target voltage value during the low-voltage period; Indicates the system impedance during low-pass operation; This represents the fault impedance ratio for low-throughput, where ; The calculation method for high-voltage ride-through impedance is as follows: （22）； In equation (22): Indicates the fault impedance during high-speed traversal; Indicates the fault reactance of high-voltage circuitry; Indicates high-throughput fault resistance; This indicates the target voltage value during high-voltage ride-through. Indicates the system impedance during high-speed penetration; This indicates the fault impedance ratio during high-voltage operation.

9. The method for integrated identification of high and low ride-through control parameters and electrical parameters of photovoltaic inverters according to claim 8, characterized in that: In step 3, the steps for optimizing the fault impedance using the improved binary search algorithm are as follows: Step 3.1: Initialize the search space, initially set the iteration count k=0, based on the target voltage U. f , n In the (pu) scenario, the search space for fault reactance and fault resistance during the initial high and low voltage ride-through is initialized using a piecewise function, i.e., X(0)f and R(0)f: （23）； In equation (23): The initial fault reactance value entered for the configuration file; This represents the maximum value of the initial fault reactance search space; Indicates the target voltage value that drops or rises during the crossing; To ensure that the initial space covers a reasonable impedance range and to avoid missing the optimal solution due to an excessively small initial range, the following is ensured: （24）； In equation (24): This represents the minimum value of the initial fault reactance search space. This represents the maximum value of the initial fault reactance search space; Therefore: （25） In equation (25): This represents the search space for fault reactance during the initial high and low voltage ride-through period. This represents the minimum value of the initial fault reactance search space. This represents the maximum value of the initial fault reactance search space; At the same time, in order to satisfy X f >>R f In practice, impedance ratio is commonly used to meet actual needs in engineering. =10; that is: （26）； In equation (26): R(0)f,min and R(0)f,max are the minimum and maximum values ​​of the initial fault resistance search space, respectively; Therefore: （27） In equation (27): R(0)f,min and R(0)f,max are the search space for the fault resistance during the initial high and low voltage ride-through; R(0)f,min and R(0)f,max are the minimum and maximum values ​​of the initial fault resistance search space, respectively. Step 3.2: Uniform Discretization of Multiple Sampling Points: The reactance search space X(k)f of the k-th iteration is uniformly discretized to generate N sets of sampling points, and the sampling step size of the k-th iteration is... With sampling vector The general term formula is: （28）； In equation (28): This represents the fault reactance sampling step size in the k-th iteration; This represents the fault reactance value at the i-th sampling point in the k-th iteration; This represents the maximum value of the fault reactance search space in the k-th iteration; This represents the minimum value of the fault reactance search space in the k-th iteration; Indicates the sampling point number; Indicates the total number of sampling points; Resistance sampling vector R (k) Generated proportionally from reactance sampling vector: ensuring impedance ratio It conforms to the actual impedance characteristics of the power system; Step 3.3: Voltage Deviation Calculation and Optimal Index Location: The fault simulation is performed automatically by calling PSASP using Python, and the voltage response vector of the k-th simulation is obtained. And calculate the voltage relative to the target voltage. deviation vector Optimal sampling point index satisfy: （29）； In equation (29): This represents the voltage response vector for the k-th iteration drop or rise. Indicates the target voltage value that drops or rises during the crossing; express and The deviation vector; Represents the simulated voltage at all sampling points i. With target voltage The minimum value; This index locates the impedance parameter within the current search space that is closest to the target voltage; Step 3.4: Adaptive search space shrinkage: Based on the current voltage With target voltage The size relationship is used to dynamically shrink the search space for the next iteration. ; The contraction strategy is represented by a piecewise function: If the current voltage Below target voltage The impedance needs to be increased to raise the voltage; （30）； In equation (30): This represents the reactance value of the optimal sampling point in the k-th iteration; In the k-th iteration, the i-th... +1 sampling point reactance value; The optimal sampling point in the k-th iteration; This represents the maximum value of the fault reactance search space in the k-th iteration; Indicates the total number of sampling points; If the current voltage Higher than the target voltage Therefore, the impedance needs to be reduced to lower the voltage; （31）； In equation (31): In the k-th iteration, the i-th... -1 reactance value at sampling points; Step 3.5: Iteration Termination Criterion. The iteration terminates and the optimal fault impedance is output when any of the following logical conditions are met. , ): （32）； In equation (32): Indicates whether the termination condition is met in the k-th iteration, and returns True or False; This represents the voltage deviation corresponding to the optimal sampling point in the k-th iteration; Indicates the convergence threshold; Indicates the current iteration number; This represents the maximum allowed number of iterations. The iteration terminates when the condition is met. This indicates that the condition is not met, and the iteration continues; The optimal impedance is: （33）； In equation (33): The optimal fault reactance value was finally identified. This represents the optimal fault resistance value obtained from the identification. This represents the fault reactance value corresponding to the optimal sampling point in the k-th iteration; This represents the fault resistance value corresponding to the optimal sampling point in the k-th iteration.

10. The method for integrated identification of high and low ride-through control parameters and electrical parameters of photovoltaic inverters according to claim 9, characterized in that: In step 4, the error values ​​of voltage, current, and power under the corresponding control mode are calculated as follows: Simulation and test results of the electromechanical transient simulation model for new energy power plants connecting to the power grid were used to calculate U, P, Q, and I. p and I q The deviation was used to verify the accuracy of the modeling. First, divide the data into segments. , and These are the start and end times of the simulation, respectively. This is the time to clear the fault. The steady-state duration before the fault. The duration of the instantaneous voltage drop phase. To account for the duration of the voltage instantaneous recovery phase, the disturbance process is divided into five intervals: A, B1, B2, C1, and C2. Transient and steady-state deviations should be calculated for each interval separately. For each transient interval, only the average deviation is calculated; for the steady-state interval, both the average and maximum deviations are calculated. The weighted average total deviation between the simulation and experimental data of the electromechanical transient simulation model for new energy power plants connected to the grid is then calculated. （34）； In equation (34): and These are the average deviations for the steady-state and transient intervals, respectively. This represents the maximum value of the deviation within the steady-state interval. The weighted average of the average deviations for each time period is calculated. This represents the simulated value at the nth sampling point; This represents the measured value of the nth sampling point; and These are the sequence numbers of the beginning and end data of each interval, respectively; Indicates from K start To K end Take the absolute value of all data points. The maximum value; This represents the index number of the data point, corresponding to the nth sampling point in the time series; This represents the steady-state deviation in interval A; This represents the steady-state deviation in interval B; This represents the steady-state deviation in interval C; For U, , , and All must satisfy <0.05; For P, Q, I p ,I q All must meet <0.1、 <0.2、 <0.15、 <0.15.

Images