Energy storage system control parameter step-by-step identification method based on CMOPSO algorithm

Abstract

The invention discloses an energy storage system control parameter identification method based on a CMOPSO algorithm, and the method comprises the steps: building an energy storage system main circuit model on a hardware-in-the-loop platform, carrying out the interaction with an actual energy storage system controller, and collecting an energy storage system output response data set under various fault ride-through working conditions. An energy storage system identification model is established, typical values are adopted for non-key parameters such as inner ring PI parameters, a universal energy storage system fault ride-through control formula is established, and parameters to be identified are determined. The fault ride-through control parameters of the energy storage system are identified step by step, a fault ride-through voltage threshold value is identified firstly, then a reactive current amplitude limiting value is identified, and finally the fault ride-through control parameters of the energy storage system are identified by combining test data and adopting a CMOPSO algorithm. And finally, verifying a fault ride-through identification result of the energy storage system, calculating an error, and evaluating the accuracy of the identification result. Compared with a traditional algorithm, the method is higher in identification precision, and the parameter identification result is higher in consistency.

Description

Technical Field

[0001] The present invention belongs to the technical field of energy storage system fault analysis, and particularly relates to a method for step-by-step identification of control parameters of an energy storage system based on the CMOPSO algorithm. Background Art

[0002] In recent years, with the rapid growth of the grid-connected capacity of new energy, the proportion of new energy power generation in the power system has increased significantly. However, new energy units are mainly connected to the grid through power electronic inverters, and their non-linear characteristics lead to serious wide-frequency oscillation phenomena and large harmonic components, making the grid operation show strong uncertainty, low inertia, weak disturbance resistance, and significant non-linear dynamic characteristics. At the same time, due to the influence of weather conditions and day-night changes on new energy power generation, the power output shows volatility and intermittency. These complex characteristics interact with each other, changing the operation characteristics of the grid.

[0003] As an important device for regulating the volatility of new energy, the energy storage system plays a crucial role in the modern power system. Through functions such as peak shaving and valley filling, frequency modulation and voltage regulation, and emergency response, the energy storage system can effectively alleviate the impact of the volatility and intermittency of new energy power generation on the grid. At the same time, when the output of new energy power generation is insufficient or a sudden failure occurs, the energy storage system can provide a stable power output to support the normal operation of the grid.

[0004] In a new energy grid with a very high proportion, the stability and reliability of the energy storage system are crucial, and one of its key indicators is the fault ride-through ability. The fault ride-through ability of the energy storage system refers to the ability of the energy storage device to maintain the grid-connected state and continuously provide voltage support and power output to the grid when a short-term grid fault (such as voltage dip, voltage rise, or frequency deviation) occurs. The fault ride-through performance of the energy storage system is directly related to the stability and recovery ability of the grid under fault conditions. To comprehensively evaluate the fault ride-through ability of the energy storage system, it is necessary to accurately identify its dynamic response characteristics and key control parameters.

[0005] When using the existing particle swarm algorithm to identify the control parameters for the fault ride-through of the energy storage system, there are problems of low efficiency and inaccurate results. The main manifestations are as follows: When the traditional particle swarm algorithm is applied to the identification problem of the energy storage system model, the algorithm is prone to falling into local optimum, which will lead to incorrect identification results. Moreover, in the existing particle swarm optimization algorithm for identification, the coupled influence of each control parameter is not considered, which is likely to lead to an extension of the optimization time and a decrease in the identification accuracy during the parameter optimization process. In addition, the existing particle swarm optimization algorithm is a single-objective optimization, and it is difficult to balance the requirements of multiple objective functions during the identification process, resulting in a reduction in the identification efficiency. In summary, the existing particle swarm optimization algorithm is not suitable for identifying the control parameters for the fault ride-through of the energy storage system. Summary of the Invention

[0006] To solve the above technical problems, the present invention provides a step-by-step identification method for the control parameters of an energy storage system based on the CMOPSO algorithm, which can identify the voltage crossing threshold, the reactive current limit value, and the fault crossing control parameters of the energy storage system, effectively improving the efficiency and accuracy of parameter identification.

[0007] The technical solution adopted by the present invention is as follows:

[0008] A step-by-step identification method for the control parameters of an energy storage system based on the CMOPSO algorithm includes the following steps:

[0009] Step 1: Based on a hardware-in-the-loop test platform, build a main circuit model of the energy storage system, interact with the actual energy storage system controller, and collect the output response data set of the energy storage system under various fault crossing conditions;

[0010] Step 2: Build an energy storage system identification model, adopt typical values for non-critical parameters such as double closed-loop PI parameters, establish a general fault crossing control formula for the energy storage system, and determine the parameters to be identified;

[0011] Step 3: Adopt step-by-step identification for the fault crossing control parameters of the energy storage system: first identify the fault crossing voltage threshold, then identify the reactive current limit value, and finally identify the fault crossing control parameters of the energy storage system by using the CMOPSO algorithm in combination with the test data.

[0012] It also includes Step 4: Verify the fault crossing identification results of the energy storage system, calculate the error, and evaluate the accuracy of the identification results.

[0013] In the above Step 1, as Figure 1 shown, build a main circuit model of the energy storage system in the hardware-in-the-loop, and complete the interaction of digital and analog signals between the main circuit and the energy storage controller through DB37 wiring, realize the grid connection of the energy storage system, and conduct tests under various conditions to collect the output response data set corresponding to the conditions.

[0014] To accurately obtain the parameter identification data set, it is necessary to extract the key points required for identification according to the output response characteristics of each fault crossing of the energy storage system. As Figure 2 shown, A, B, and C respectively correspond to the three stages before the fault, during the fault, and after the fault. Among them, section B can be divided into a transient interval B 1 and a steady-state interval B 2 , section C can be divided into a transient interval C 1 and a steady-state interval C 2 . Calculate the average values of voltage, active current, and reactive current in each fault condition interval A and interval B 2 respectively, and use them as the output response data set of the energy storage system.

[0015] Calculate the average values of voltage, active current, and reactive current in each fault condition interval A and interval B 2The average values of voltage, active current, and reactive current are as follows:

[0016] Among them, taking voltage as an example, for fault interval B 2 The calculation formula is:

[0017]

[0018] In formula (1): b 1 , b 2 are the start and end times of the fault interval respectively; u(t) is the collected voltage value; u b is the average value of this interval segment.

[0019] In step 2, build an energy storage system identification model, interact with the actual energy storage system controller, and adopt typical values for non-critical parameters such as the inner-loop PI parameters;

[0020] As Figure 3 shown, establish a power outer-loop controller model, and the expression of the power outer-loop controller model is:

[0021]

[0022] In formula (2): i dref , i qref are the reference values of active current and reactive current output by the power outer-loop controller respectively; K p , K i are the proportional coefficient and integral coefficient of the power outer-loop controller respectively; P, P ref are the actual value and reference value of active power respectively; Q, Q ref are the actual value and reference value of reactive power respectively.

[0023] Establish a current inner-loop controller model, and the expression of the current inner-loop controller model is:

[0024]

[0025] In formula (3), u d , u q are the d-axis and q-axis voltage control signals respectively; K pi , K ii are the proportional coefficient and integral coefficient of the inner-loop current controller respectively; U sd , U sq are the d-axis and q-axis components of the grid-connected voltage respectively, L is the filter inductor, and ω is the synchronous angular velocity.

[0026] The proportional coefficient K p and integral coefficient K i of the power outer-loop controller in the established double-loop controller model of the energy storage system, and the proportional coefficient K of the current inner-loop controllerpi And the integral coefficient K ii Is set to a typical value.

[0027] The calculation formula for the reactive current during the energy storage system failure is:

[0028]

[0029] In Equation (4): I qb Is the calculated value of the reactive current during the failure, U t Is the per-unit value of the grid connection point voltage during the failure, U t_HV Is the entry into the high voltage ride-through threshold, U t_LV The entry into the low voltage ride-through threshold, I q0 Is the initial reactive current, I N Is the rated current.

[0030] K 1_Iq_HVRT Is the reactive current support coefficient during high ride-through, K 2_Iq_HVRT Is the reactive current calculation coefficient during high voltage ride-through, I qHV Is the set value of the reactive current during high voltage ride-through. K 1_Iq_LVRT Is the reactive current support coefficient during low voltage ride-through, K 2_Iq_LVRT Is the reactive current calculation coefficient during low voltage ride-through, I qLV Is the set value of the reactive current during low voltage ride-through;

[0031] To limit the reactive current during the failure and avoid over-limit of the reactive current, the reactive current is limited, and the calculation formula is:

[0032] I qref = min(I qb , I q,max )(5);

[0033] In Equation (5): I qref Is the command value of the reactive current during the failure, I q,max Is the limit value of the reactive current during the failure.

[0034] The calculation formula for the active current during the energy storage system failure is:

[0035]

[0036] In Equation (6): I dref Is the command value of the active current during the failure, K 1_Id_HVRT Is the active current support coefficient during high voltage ride-through, K 2_Id_HVRT Is the active current calculation coefficient during high voltage ride-through, I dHV Is the set value of the active current during high voltage ride-through. Among them, K 1_Id_LVRTis the active current support coefficient during low-voltage ride-through, K 2_Id_LVRT is the active current calculation coefficient during low-voltage ride-through, I dLV is the set value of active current during low-voltage ride-through; I d0 represents the initial active current. According to the energy storage system fault ride-through control calculation formulas (4)(5)(6), determine the remaining control parameters to be identified during the fault ride-through process: U t_HV 、U t_LV 、I q,max 、K 1_Iq_HVRT 、K 2_Iq_HVRT 、I qHV 、K 1_Iq_LVRT 、K 2_Iq_LVRT 、I qLV 、K 1_Id_HVRT 、K 2_Id_HVRT 、I dHV 、K 1_Id_LVRT 、K 2_Id_LVRT 、I dLV 。

[0037] Step 3 includes the following steps:

[0038] Step 3.1: The energy storage system decides whether to switch to fault ride-through control by detecting the magnitude of the grid connection point voltage. In the case of discharging, when the grid connection point voltage is within [U t_LV ,U t_HV , the energy storage system outputs constant active and reactive power. When the grid connection point voltage is outside the range of [U t_LV ,U t_HV , both the active power and the reactive power will change. Recording the voltage values when the power changes under different fault ride-through conditions can obtain U t_LV and U t_HV ;

[0039] Step 3.2: After identifying the voltage ride-through threshold, select the test data under the 20% voltage fault dip condition to obtain the reactive current value during the fault. At this time, the reactive current reaches the maximum value, which is the reactive current limit value I q,max ;

[0040] Step 3.3: Substitute the obtained voltage ride-through threshold and reactive current limit value into equations (4), (5), and (6). The parameter identification problem is converted into a multi-objective optimization problem. Take the errors between the calculated values and the measured values of the active current and the reactive current as the objective functions f 1 and f 2 , and the expressions are as follows:

[0041]

[0042] In equation (7): I qtestis the measured value of the reactive current, I dtest is the measured value of the active current, I qsim is the calculated value of the identified reactive current, I dsim is the calculated value of the identified active current.

[0043] The CMOPSO algorithm is used to obtain the optimal solutions of the objective functions f 1 and f 2 to obtain the fault ride-through control parameters of the energy storage system corresponding to the optimal solution set: K 1_Iq_HVRT 、K 2_Iq_HVRT 、I qHV 、K 1_Iq_LVRT 、K 2_Iq_LVRT 、I qLV 、K 1_Id_HVRT ,K 2_Id_HVRT 、I dHV 、K 1_Id_LVRT 、K 2_Id_LVRT 、I dLV 。

[0044] The CMOPSO algorithm is used to identify the fault ride-through control parameters of the energy storage system, including the following steps:

[0045] Step1. Initialize the particle population and the external archive, set the number of particle populations as N, the particle dimension as D, the external archive as A, the size of the external archive as M, and set the maximum number of iterations as T max ,and the critical fitness value ε;

[0046] Step2. Calculate the fitness value of each particle according to the objective function formula (7), and judge whether each particle is a non-dominated solution according to the Pareto dominance relationship;

[0047] Among them, the calculation formula of the Pareto dominance relationship is:

[0048]

[0049] In formula (8): m, n are the numbers of the objective functions; x 1 ,x 2 are the positions of two particles in the population. It represents that the condition for x 1 to dominate x 2 is that x 1 is not inferior to x 2 in all objectives and is superior to x 2 in at least one objective.

[0050] If so, put it into the external archive. For the particles in the external archive, the crowding distance and contribution degree need to be calculated, and the particles with low contribution degree and large crowding distance are deleted according to the size limit of the external archive. The calculation formula of the crowding distance is as follows:

[0051]

[0052] In Equation (9): i is the particle number in the population; j is the objective function number, and d i is the crowding distance of the particle; f j (i + 1) and f j (i - 1) are the fitness values of adjacent particles respectively; f j,max and f j,min are the maximum and minimum values of the objective function f j respectively.

[0053] In addition, the calculation formula for the contribution degree of particles in the external archive is:

[0054]

[0055] In Equation (10): C(x) represents the particle contribution degree, x represents the particle position; I(x, x i ) represents whether x dominates x i , if it dominates, then I(x, x i ) = 1; if it does not dominate, then I(x, x i ) = 0, and w is the sparsity weight.

[0056] Step 3: The process of particle optimization is affected by the individual optimal value and the population optimal value. The velocity and position update formulas of the particles are shown as follows:

[0057] v i = ω × v i + c 1 × rand × (p best - x i ) + c 2 × rand × (g best - x i ) (11)

[0058] x i = v i + x i (12)

[0059] In the formula: v i represents the velocity of the i-th particle; x i represents the position of the i-th particle; rand represents a random number between 0 and 1; p best is the individual optimal value, g best is the global optimal value selected through the competition mechanism, c 1 , c 2 are learning factors, and ω is the adaptive inertia weight;

[0060] Step 4. Update the individual optimal value p of the particle based on the crowding distance in the external archive best , and update the global optimal value g according to the contribution degree of the particle best ;

[0061] Step 5. If g best < ε or the iteration number T is reached max , stop the iteration, output the optimal solution as the identification result, otherwise repeat Step 2 to Step 4 until g best < ε or the iteration number T is reached max , then stop the iteration.

[0062] In the said Step 4, the result verification includes the following steps:

[0063] Step 4.1. Set the voltage dip from 10% to 80% with a step size of 5%;

[0064] Step 4.2. Substitute the identification result into the energy storage system identification model, run the energy storage system identification model, and obtain the active and reactive current waveforms;

[0065] Step 4.3. Calculate the error between the simulation result and the measured result, and the formula is as follows:

[0066]

[0067] In the formula: δ d and δ q are the relative errors of the active and reactive currents respectively; i d_act and i q_act are the measured active and reactive current values respectively; i d_sim and i q_sim are the active and reactive current values simulated by the identification model respectively.

[0068] A method for identifying the fault ride-through control parameters of an energy storage system based on the CMOPSO algorithm of the present invention has the following beneficial effects:

[0069] 1) By stepwise identifying the control parameters of the energy storage system, the present invention stepwise identifies the voltage crossing threshold, the reactive current limit value, and the fault ride-through control parameters, reduces the interference between the identified parameters, and improves the efficiency of parameter identification.

[0070] 2) Based on the CMOPSO algorithm, the present invention transforms the parameter identification problem into a multi-objective optimization problem, sets the errors between the calculated values and the measured values of the active current and the reactive current as the objective functions respectively, and iteratively updates the population position until the fault ride-through control parameters of the energy storage system that satisfy the global optimal solution are found and output. Compared with the traditional method, the parameter identification result has higher accuracy.

[0071] 3) The present invention introduces a competition mechanism in the parameter identification process, selects the optimal solution according to the contribution degree and crowding distance of particles, and further improves the convergence and diversity of the population.

[0072] 4) The present invention identifies the fault ride-through control parameters of the energy storage system model through the CMOPSO algorithm, which can effectively improve the efficiency and accuracy of parameter identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] The present invention will be further described below in conjunction with the drawings and examples;

[0074] Figure 1 It is a block diagram of the hardware-in-the-loop test of the energy storage system.

[0075] Figure 2 It is a sectional view of the fault ride-through process.

[0076] Figure 3 It is a double closed-loop control block diagram of the energy storage system.

[0077] Figure 4 It is a schematic diagram of the fault ride-through control parameter identification method for the energy storage system of the present invention.

[0078] Figure 5 It is a flow chart of the CMOPSO fault ride-through control parameter identification algorithm of the present invention.

[0079] Figure 6 It is a schematic diagram of the comparison of the output results of voltage, active power, and reactive power under the 20% voltage ride-through condition of the present invention.

[0080] Figure 7 It is a schematic diagram of the comparison of the output results of voltage, active power, and reactive power under the 125% voltage ride-through condition of the present invention.

[0081] Figure 8 It is an error bar chart of the output data under the 20% voltage ride-through condition of the present invention.

[0082] Figure 9 It is an error bar chart of the output data under the 125% voltage ride-through condition of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0083] The energy storage system control parameter identification method based on the CMOPSO algorithm includes the following steps:

[0084] Step 1: Based on the hardware-in-the-loop test platform, build the main circuit model of the energy storage system, interact with the actual energy storage system controller, and collect the output response data set of the energy storage system under various fault ride-through conditions.

[0085] Step 2: Build an energy storage system identification model, adopt typical values for non-critical parameters such as the inner-loop PI parameters, establish a general fault ride-through control formula for the energy storage system, and determine the parameters to be identified.

[0086] Step 3: Identify the fault ride-through control parameters of the energy storage system step by step. First, identify the fault ride-through voltage threshold, then identify the reactive current limit value, and finally, use the CMOPSO algorithm to identify the fault ride-through control parameters of the energy storage system in combination with the test data.

[0087] Step 4: Verify the fault ride-through identification results of the energy storage system, calculate the error, and evaluate the accuracy of the identification results.

[0088] Specifically, in Step 1, to accurately obtain the parameter identification data set, it is necessary to extract the key points required for identification according to the output response characteristics of each fault ride-through of the energy storage system. As Figure 2 shown, A, B, and C correspond to the three stages before the fault, during the fault, and after the fault respectively. Among them, section B can be divided into the transient interval B 1 and the steady-state interval B 2 , and section C can be divided into the transient interval C 1 and the steady-state interval C 2 . Calculate the average values of voltage, active current, and reactive current in each fault condition interval A and interval B 2 respectively, and use them as the output response data set of the energy storage system.

[0089] Specifically, in Step 2, the formula for calculating the reactive current during the fault of the energy storage system is:

[0090]

[0091] In the formula: I qb is the calculated value of the reactive current during the fault, U t is the per-unit value of the grid connection point voltage during the fault, U t_HV is the threshold for entering high-voltage ride-through, U t_LV is the threshold for entering low-voltage ride-through, I q0 is the initial reactive current, I N is the rated current. K 1_Iq_HVRT is the reactive current support coefficient during high-voltage ride-through, K 2_Iq_HVRT is the calculation coefficient of the reactive current during high-voltage ride-through, I qHV is the set value of the reactive current during high-voltage ride-through. K 1_Iq_LVRT is the reactive current support coefficient during low-voltage ride-through, K 2_Iq_LVRT is the calculation coefficient of the reactive current during low-voltage ride-through, I qLV is the set value of the reactive current during low-voltage ride-through.

[0092] To limit the reactive current during a fault and avoid over-limit of the reactive current, the reactive current is amplitude-limited, and the calculation formula is:

[0093] I qref =min(I qb ,I q,max ) (5)

[0094] In the formula: I qref is the reactive current command value during the fault, and I q,max is the reactive current limit value during the fault.

[0095] The calculation formula for the active current of the energy storage system during the fault is:

[0096]

[0097] In the formula: I dref is the active current command value during the fault, K 1_Id_HVRT is the active current support coefficient during high voltage ride-through, K 2_Id_HVRT is the active current calculation coefficient during high voltage ride-through, I dHV is the active current setting value during high voltage ride-through. Among them, K 1_Id_LVRT is the active current support coefficient during low voltage ride-through, K 2_Id_LVRT is the active current calculation coefficient during low voltage ride-through, I dLV is the active current setting value during low voltage ride-through. I d0 represents the initial active current.

[0098] According to the energy storage system fault ride-through control calculation formula, determine the remaining control parameters to be identified during the fault ride-through process, U t_HV 、U t_LV 、I q,max 、K 1_Iq_HVRT 、K 2_Iq_HVRT 、I qHV 、K 1_Iq_LVRT 、K 2_Iq_LVRT 、I qLV 、K 1_Id_HVRT 、K 2_Id_HVRT 、I dHV 、K 1_Id_LVRT 、K 2_Id_LVRT 、I dLV 。

[0099] Specifically, the step-by-step identification in step three is mainly divided into three steps: voltage ride-through threshold identification, reactive current limit value identification, and fault ride-through control parameter identification.

[0100] Step A1. The energy storage system determines whether to switch to fault ride-through control by detecting the magnitude of the grid connection point voltage. In the case of discharging, when the grid connection point voltage is within [Ut_LV , U t_HV , the active and reactive power output by the energy storage system remains unchanged. When the grid connection point voltage is within [U t_LV , U t_HV , the active power output decreases and the reactive power output increases. By calculating the voltage value when the power changes under different fault ride-through conditions, U t_LV and U t_HV can be obtained.

[0101] Step A2: After identifying the voltage ride-through threshold, select the test data under the 20% voltage fault dip condition and calculate the reactive current value at this time, which is I q,max .

[0102] Step A3: After substituting the obtained voltage ride-through threshold and reactive current limit value into equations (1) and (2), the parameter identification problem is converted into a multi-objective optimization problem. The errors between the calculated values and the measured values of the active current and reactive current are used as the objective functions f 1 and f 2 , respectively. The expressions are as follows:

[0103]

[0104] Where: I qtest is the measured value of the reactive current, I dtest is the measured value of the active current, I qsim is the identified calculated value of the reactive current, I dsim is the identified calculated value of the active current.

[0105] Use the CMOPSO algorithm to find the optimal solutions of the objective functions f 1 and f 2 , and obtain the fault ride-through control parameters K 1_Iq_HVRT , K 2_Iq_HVRT , I qHV , K 1_Iq_LVRT , K 2_Iq_LVRT , I qLV , K 1_Id_HVRT , K 2_Id_HVRT , I dHV , K 1_Id_LVRT , K 2_Id_LVRT , I dLV of the energy storage system corresponding to the optimal solution set.

[0106] Specifically, the steps of using the CMOPSO algorithm to identify the fault ride-through control parameters of the energy storage system in step three include the following steps:

[0107] Step B1: Initialize the particle swarm and the external archive. Set the number of particle swarms as N, the particle dimension as D, the external archive as A, the size of the external archive as M, and set the maximum number of iterations as T max, and the convergence threshold ε.

[0108] Step B2: Calculate the fitness value of each particle according to the objective function formula (4), and judge whether each particle is a non-dominated solution according to the Pareto dominance relationship. If so, put it into the external archive. The particles in the external archive need to calculate the crowding distance and contribution degree, and delete the particles with low contribution degree and large crowding distance according to the size limit of the external archive. The calculation formula of the crowding distance is as follows:

[0109]

[0110] In the formula: i is the particle number in the population, j is the objective function number, d i is the crowding distance of the particle; f j (i + 1) and f j (i - 1) are the fitness values of adjacent particles respectively; f j,max and f j,min are the maximum and minimum values of the objective function f j respectively.

[0111] In addition, the calculation formula of the contribution degree of the particles in the external archive is:

[0112]

[0113] In the formula: C(x) represents the particle contribution degree, x represents the particle position, I(x, x i ) represents whether x dominates x i , if it dominates, then I(x, x i ) = 1, if it does not dominate, then I(x, x i ) = 0, and w is the sparsity weight.

[0114] Step B3: The process of particle optimization is affected by the individual optimal value and the population optimal value. The velocity and position update formulas of the particles are shown as follows:

[0115] v i = ω × v i + c 1 × rand × (p best - x i ) + c 2 × rand × (g best - x i ) (11)

[0116] x i = v i + x i (12)

[0117] In the formula: v represents the particle update velocity, p best is the individual optimal value, gbest is the global optimal value selected through the competition mechanism, c 1 , c 2 are learning factors, and ω is the adaptive inertia weight;

[0118] Step B4: Update the individual optimal value p of the particle based on the crowding distance of the particle in the external archive best , and update the global optimal value g according to the contribution degree of the particle best , and finally calculate the optimal fitness B best .

[0119] Step B5: If B best < ε or the iteration number T is reached max , then stop the iteration, and output the optimal solution as the identification result; otherwise, repeat steps B2 - B4 until B best < ε or the iteration number is reached, then stop the iteration.

[0120] Specifically, the result verification of the above step four includes the following steps:

[0121] Step C1: Set the voltage dip from 10% to 80%, with a step size of 5%;

[0122] Step C2: Substitute the identification result into the energy storage system identification model, run the simulation model, and obtain the active and reactive current waveforms;

[0123] Step C3: Calculate the error between the simulation result and the measured result, and the formula is as follows:

[0124]

[0125] In the formula: δ d and δ q are the relative errors of the active and reactive currents respectively, i d_act and i q_act are the measured active and reactive current values respectively, i d_sim and i q_sim are the active and reactive current values simulated by the identification model respectively.

[0126] Figure 6 is a schematic diagram of the comparison of the voltage, active power, and reactive power output results of the present invention under the voltage 20% crossing condition. Figure 6 It can be seen that in the low voltage ride - through test, the CMOPSO algorithm can accurately identify the low voltage ride - through control parameters of the energy storage system.

[0127] Figure 7 is a schematic diagram of the comparison of the voltage, active power, and reactive power output results of the present invention under the voltage 125% crossing condition.

[0128] Figure 7 It can be seen that in the high-voltage ride-through test, the CMOPSO algorithm can accurately identify the high-voltage ride-through control parameters of the energy storage system.

[0129] Figure 8 This is the error bar chart of the output data of the present invention under the voltage 20% ride-through condition. Figure 8 It can be seen that in the low-voltage ride-through test, the use of the CMOPSO algorithm for identification can effectively reduce the identification error and improve the identification accuracy.

[0130] Figure 9 This is the error bar chart of the output data of the present invention under the voltage 125% ride-through condition. Figure 9 It can be seen that in the high-voltage ride-through test, the use of the CMOPSO algorithm for identification can effectively reduce the identification error and improve the identification accuracy.

[0131] The present invention performs step-by-step identification of the control parameters of the energy storage system, step-by-step identification of the voltage ride-through threshold, the reactive current limit value, and the fault ride-through control parameters, reduces the interference between the identified parameters, and improves the efficiency of parameter identification. Based on the CMOPSO algorithm, the present invention transforms the parameter identification problem into a multi-objective optimization problem, sets the error between the calculated values and the measured values of the active current and the reactive current as the objective functions respectively, iteratively updates the population position until the fault ride-through control parameters of the energy storage system that satisfy the global optimal solution are found and output. Compared with the traditional method, the parameter identification result has higher accuracy. The present invention introduces a competition mechanism in the parameter identification process, selects the optimal solution according to the contribution degree and crowding distance of the particles, and further improves the convergence and diversity of the population. The present invention uses the CMOPSO algorithm to identify the fault ride-through control parameters of the energy storage system model, which can effectively improve the efficiency and accuracy of parameter identification.

Claims

1. A step-by-step identification method for energy storage system control parameters based on CMOPSO algorithm, characterized by The following steps are involved: Step 1: Build the main circuit model of the energy storage system, interact with the actual energy storage system controller, and collect the energy storage system output response data set under various fault ride-through conditions; Step 2: Build an energy storage system identification model, use typical values ​​for the dual closed-loop PI parameters, establish the energy storage system fault ride-through control formula, and determine the parameters to be identified; Step 3: The fault ride-through control parameters of the energy storage system are identified step by step: first, the fault ride-through voltage threshold is identified, then the reactive current limit value is identified, and finally, the CMOPSO algorithm is used to identify the fault ride-through control parameters of the energy storage system in combination with the test data.

2. The step-by-step identification method of energy storage system control parameters based on CMOPSO algorithm according to claim 1 is characterized in that: The method also includes step 4: verifying the fault ride-through identification result of the energy storage system, calculating the error, and evaluating the accuracy of the identification result.

3. The step-by-step identification method of energy storage system control parameters based on CMOPSO algorithm according to claim 1 is characterized in that: In step 1, a main circuit model of the energy storage system is built in the hardware-in-the-loop, and the interaction of digital and analog signals between the main circuit and the energy storage controller is completed to realize the grid connection of the energy storage system, and tests of various working conditions are performed to collect output response data sets under corresponding working conditions; In order to accurately obtain the parameter identification data set, the key points required for identification are extracted according to the output response characteristics of each fault ride-through of the energy storage system; A, B and C correspond to the three stages before, during and after the fault, respectively. The B section is divided into the transient interval B1 and the steady-state interval B2, and the C section is divided into the transient interval C1 and the steady-state interval C2. The mean values ​​of voltage, active current and reactive current in each fault condition interval A and interval B2 are calculated respectively and used as the output response data set of the energy storage system.

4. The step-by-step identification method of energy storage system control parameters based on CMOPSO algorithm according to claim 3 is characterized by: When calculating the voltage mean, the calculation formula for the fault interval B2 is as follows: In formula (1), b1 and b2 are the start and end time of the fault interval respectively; u(t) is the collected voltage value; u b is the mean value of this interval.

5. The step-by-step identification method of energy storage system control parameters based on CMOPSO algorithm according to claim 1 is characterized in that: In step 2, an energy storage system identification model is built, interacting with the actual energy storage system controller, and typical values ​​are used for the inner loop PI parameters; A power outer loop controller model is established. The expression of the power outer loop controller model is: In formula (2): i dref 、i qref They are respectively the active current and reactive current reference values ​​output by the power outer loop controller; K p , K i are the proportional coefficient and integral coefficient of the power outer loop controller respectively; P, P ref are the actual value and reference value of active power respectively; Q, Q ref They are the actual value and reference value of reactive power respectively; A current inner loop controller model is established, and the expression of the current inner loop controller model is: In formula (3), u d 、u q are d-axis and q-axis voltage control signals respectively; K pi , K ii are the proportional coefficient and integral coefficient of the inner loop current controller respectively; U sd , U sq are the d-axis and q-axis components of the grid-connected voltage, L is the filter inductance, and ω is the synchronous angular velocity; The proportional coefficient K of the power outer loop controller in the established energy storage system dual closed-loop controller model p and the integration coefficient K i , the proportional coefficient K of the current inner loop controller pi and the integration coefficient K ii Set to typical values; The reactive current calculation formula during the energy storage system fault is: In formula (4): I qb is the calculated value of reactive current during fault period, U t is the per unit value of the grid connection point voltage during the fault period, U t_HV To enter the high voltage ride-through threshold, U t_LV Entering the low voltage ride-through threshold, I q0 is the initial reactive current, I N is the rated current; K 1_Iq_HVRT is the reactive current support factor during high-voltage run-through, K 2_Iq_HVRT is the reactive current calculation coefficient during high voltage ride-through, I qHV K is the reactive current setting value during high voltage ride-through; 1_Iq_LVRT is the reactive current support factor during low voltage ride-through, K 2_Iq_LVRT is the reactive current calculation coefficient during the low voltage ride-through period, I qLV Reactive current setting value during low voltage ride-through; The reactive current is limited, and the calculation formula is: IN qref =min(I qb ,IN q,max )(5); In formula (5): I qref is the reactive current command value during the fault period, I q,max is the reactive current limit value during the fault period; The calculation formula of active current during energy storage system fault is: In formula (6): I dref is the active current command value during the fault period, K 1_Id_HVRT is the active current support factor during high voltage ride through, K 2_Id_HVRT is the active current calculation coefficient during high voltage ride through, I dHV is the active current setting value during high voltage ride-through; where K 1_Id_LVRT is the active current support factor during low voltage ride-through, K 2_Id_LVRT is the active current calculation coefficient during low voltage ride-through, I dLV It is the active current setting value during the low voltage ride-through period; I d0 represents the initial active current; according to the energy storage system fault ride-through control calculation formula (4)(5)(6), determine the remaining control parameters to be identified during the fault ride-through process: U t_HV , U t_LV ,I q,max , K 1_Iq_HVRT , K 2_Iq_HVRT ,I qHV , K 1_Iq_LVRT , K 2_Iq_LVRT ,I qLV , K 1_Id_HVRT , K 2_Id_HVRT ,I dHV , K 1_Id_LVRT , K 2_Id_LVRT ,I dLV .

6. The step-by-step identification method of energy storage system control parameters based on CMOPSO algorithm according to claim 5 is characterized in that: The step 3 comprises the following steps: Step 3.1: The energy storage system determines whether to switch to fault ride-through control by detecting the grid voltage. In the case of discharge, when the grid voltage is [U t_LV ,U t_HV ], the output active and reactive power of the energy storage system remains unchanged. When the grid connection point voltage is [U t_LV ,U t_HV ] range, both active power and reactive power will change. By recording the voltage value when the power changes under different fault ride-through conditions, U t_LV and U t_HV ; Step 3.2: After the voltage crossing threshold is identified, select the test data under the 20% voltage fault drop condition to obtain the reactive current value during the fault period. At this time, the reactive current reaches the maximum value, which is the reactive current limit value I q,max ; Step 3.3: Substitute the obtained voltage crossing threshold and reactive current limit value into equations (4), (5), and (6), and transform the parameter identification problem into a multi-objective optimization problem. The errors between the calculated and measured values ​​of active current and reactive current are used as objective functions f1 and f2, respectively. The expressions are as follows: In formula (7): I qtest is the measured value of reactive current, I dtest is the measured value of active current, I qsim Reactive current identification calculation value, I dsim Active current identification calculation value; The CMOPSO algorithm is used to find the optimal solution of the objective functions f1 and f2, and the fault ride-through control parameters of the energy storage system corresponding to the optimal solution set are obtained: K 1_Iq_HVRT , K 2_Iq_HVRT ,I qHV , K 1_Iq_LVRT , K 2_Iq_LVRT ,I qLV , K 1_Id_HVRT , K 2_Id_HVRT ,I dHV , K 1_Id_LVRT , K 2_Id_LVRT ,I dLV .

7. The step-by-step identification method of energy storage system control parameters based on CMOPSO algorithm according to claim 7 is characterized in that: The CMOPSO algorithm is used to identify the fault ride-through control parameters of the energy storage system, including the following steps: Step 1. Initialize the particle population and external archive, set the particle population to N, particle dimension to D, external archive to A, external archive size to M, and set the maximum number of iterations to T max , and critical fitness value ε; Step 2: Calculate the fitness value of each particle according to the objective function formula (7), and determine whether each particle is a non-dominated solution based on the Pareto dominance relationship; Among them, the Pareto dominance relationship calculation formula is: In formula (8), m and n are the numbers of the objective functions; x1 and x2 are the positions of two particles in the population; the condition for x1 to dominate x2 is that x1 is not inferior to x2 in all objectives and is superior to x2 in at least one objective; If yes, it is put into the external archive, and the crowding distance and contribution of the particles in the external archive are calculated. Particles with low contribution and large crowding distance are deleted according to the size limit of the external archive; the calculation formula of crowding distance is as follows: In formula (9), i is the particle number in the population; j is the objective function number, d i is the crowding distance of particles; f j (i+1) and f j (i-1) are the fitness values ​​of the adjacent particles; f j,max and f j,min The objective function f j The maximum and minimum values ​​of ; In addition, the calculation formula for the particle contribution in the external archive is: In formula (10), C(x) represents the particle contribution, x represents the particle position; I(x,x i ) indicates whether x dominates x i , if it dominates, then I(x,x i )=1; if it does not dominate, then I(x,x i )=0, w is the sparsity weight; Step 3: The particle optimization process is affected by the individual optimal value and the population optimal value. The particle speed and position update formula are as follows: v i =ω×v i +c1×rand×(p best -x i )+c2×rand×(g best -x i ) (11) x i =v i +x i (12) Where: v i represents the velocity of the ith particle; x i represents the position of the i-th particle; rand represents a random number between 0 and 1; p best is the individual optimal value, g best is the global optimal value selected by the competition mechanism, c1 and c2 are learning factors, and ω is the adaptive inertia weight; Step 4: Update the optimal value p of the individual particle based on the crowding distance of the particle in the external archive best , update the global optimal value g according to the particle's contribution best ; Step 5, if g is satisfied best <ε or reaches the number of iterations T max , then stop the iteration and output the optimal solution as the identification result. Otherwise, repeat Step 2 to Step 4 until g is satisfied. best <ε or reaches the number of iterations T max , the iteration stops.

8. The step-by-step identification method of energy storage system control parameters based on CMOPSO algorithm according to claim 2 is characterized by: In step 4, the result verification includes the following steps: Step 4.1, set the voltage drop from 10% to 80% with a step size of 5%; Step 4.2, substitute the identification results into the energy storage system identification model, run the energy storage system identification model, and obtain the active and reactive current waveforms; Step 4.3: Calculate the error between the simulation results and the measured results. The formula is as follows: Where: d and δ q are the relative errors of active and reactive current respectively; i d_act and i q_act are the measured active and reactive current values ​​respectively; i d_sim and i q_sim are the active and reactive current values ​​of the identification model simulation respectively.

Images