Voltage sag assessment method based on low-voltage ride-through control parameter identification

Abstract

Disclosed in the present invention is a voltage sag assessment method based on low-voltage ride-through control parameter identification. The method comprises the following steps: S1, establishing a mathematical model of output currents of an inverter-interfaced distributed generator, and determining key control parameters to be identified; S2, converting a voltage and a current in monitoring data into a positive-sequence component and a negative-sequence component in a dq coordinate system; S3, on the basis of a regression tree algorithm and a least square method, performing identification on the key control parameters; S4, establishing a fault equivalent model of the inverter-interfaced distributed generator during low-voltage ride through; S5, calculating a voltage at each node, so as to obtain an output short-circuit current; and S6, on the basis of a Monte Carlo method and the output short-circuit current of the inverter-interfaced distributed generator, performing fault-induced voltage sag assessment. The present invention provides a control parameter identification method for an inverter-interfaced distributed generator during low-voltage ride through, so as to improve the accuracy of voltage sag assessment.

Description

Voltage Sag Assessment Method Based on Low Voltage Ride-Through Control Parameter Identification Technical Field

[0001] This invention relates to the field of voltage sag assessment technology, and in particular to a voltage sag assessment method based on low voltage ride-through control parameter identification. Background Technology

[0002] With the increasing grid-connected capacity of inverter-type distributed generation (DG), large-scale grid disconnection of DG during faults will further exacerbate the voltage sag problem. During grid faults, the output current characteristics of DG are related to its low-voltage ride-through (LVRT) strategy. However, the control strategy parameters of DG are not directly disclosed due to factors such as inconsistent equipment characteristics and trade secrets, and there is currently no direct method to obtain LVRT control parameters. With the introduction of technical specifications such as photovoltaic grid connection regulations and photovoltaic power plant power quality testing standards, online monitoring data of power quality at the DG grid connection point can now be obtained. Considering the differences in voltage and current parameter trajectories of DG under different control parameters, identifying the key control parameters of the LVRT strategy through online monitoring data is of great significance for voltage sag assessment.

[0003] Distributed power sources (DPGs) need to fully consider their output characteristics under fault conditions during low-voltage ride-through (LVRT). However, the output characteristics under fault conditions are related to the control strategy employed. Different manufacturers use different control strategies, and due to trade secrets, they typically do not disclose LVRT control parameters. Therefore, these parameters are difficult to obtain, making it impossible to establish an accurate equivalent model for the DPG and thus difficult to accurately assess voltage sags. Existing methods face the following challenges: there is currently no method to directly obtain LVRT control parameters from online monitoring data; and when an asymmetrical fault occurs, the total short-circuit current output by multiple inverter-type DPGs cannot be accurately calculated. Summary of the Invention

[0004] To address the problem that control parameters are difficult to obtain during low-voltage ride-through in inverter-type distributed power sources, resulting in an inability to accurately assess voltage sags, this invention proposes a voltage sag assessment method based on low-voltage ride-through control parameter identification, thus resolving the aforementioned issue.

[0005] This application discloses a voltage sag assessment method based on low voltage ride-through control parameter identification, including the following steps:

[0006] S1. Based on the negative sequence current under asymmetrical fault, establish a mathematical model of the output current of the inverter-type distributed power source and determine the key control parameters to be identified.

[0007] S2. Acquire online monitoring data and convert the voltage and current in the monitoring data into positive and negative sequence components in the dq coordinate system;

[0008] S3. Identify key control parameters based on regression tree algorithm and least squares method;

[0009] S4. Based on the identification results of key control parameters, establish an equivalent fault model for inverter-type distributed power sources during low-voltage ride-through.

[0010] S5. Calculate the voltage of each node of the inverter-type distributed power source after the fault based on the fault equivalent model, and then obtain the output short-circuit current of the inverter-type distributed power source.

[0011] S6. Fault voltage sag assessment based on Monte Carlo method and output short-circuit current of inverter-type distributed power source.

[0012] Preferably, the mathematical model of the output current of the inverter-type distributed power source includes a mathematical model of reactive power compensation current, a mathematical model of active power output current, and a mathematical model of negative sequence control current.

[0013] Preferably, the mathematical model expression for the reactive power compensation current is as follows:

[0014] in, For the positive sequence component of reactive power compensation current, U pcc I is the grid connection point voltage. N K1 and K2 are the rated current, K1 and K2 are the reactive power compensation current control parameters, α is the reactive power compensation current segmentation point, and the magnitudes of K1 and K2 satisfy the requirement that they are continuous at the segmentation point.

[0015] The mathematical model expression for the active power output current is as follows:

[0016] in, K represents the positive-sequence component of the active output current. MAX These are the current limiting control parameters for the inverter;

[0017] The mathematical model expression for the negative sequence control current is as follows:

[0018] in, This is the negative sequence control current on the d-axis. K is the negative sequence control current on the q-axis. N For negative sequence control parameters, The negative sequence component of the grid connection point voltage along the d-axis. This represents the negative q-axis component of the grid connection point voltage. This represents the positive sequence component of the voltage at the grid connection point along the d-axis.

[0019] Preferably, the key control parameters to be identified include reactive power compensation current control parameters K1 and K2, reactive power compensation current segmentation point α, and inverter current limiting control parameter K. MAX Negative sequence control parameter K N .

[0020] Preferably, step S2 includes the following steps:

[0021] Using rotation factor Symmetrical component transformation is performed on the three-phase voltage and three-phase current to separate the positive and negative sequences of the three-phase voltage and current:

[0022] in, This represents the positive sequence component of the phase a voltage or current. p is the negative sequence component of phase a voltage or current. a For phase a voltage or current, p b For phase b voltage or current, p c This refers to the voltage or current of phase c.

[0023] The positive and negative sequence components of the three-phase voltage and current are transformed into positive and negative sequence components of voltage and current in the dq coordinate system through positive sequence component Park transformation and negative sequence component Park transformation.

[0024] The positive-order component Park transform is shown below:

[0025] The negative-order component Park transform is shown below:

[0026] Where, p + d p represents the positive sequence component of voltage or current along the d-axis. + q p represents the positive-sequence component of the voltage or current along the q-axis. + b p represents the positive sequence component of phase b voltage or current. + c p is the negative sequence component of the c-phase voltage or current. - d p represents the negative sequence component of the voltage or current along the d-axis. - q p is the negative sequence component of the voltage or current along the q-axis. - b p is the negative sequence component of phase b voltage or current. - c Let θ be the negative sequence component of phase c voltage or current, and let θ be the rotation angle of the abc coordinate system relative to the dq coordinate system.

[0027] Preferably, the reactive power compensation current segmentation point α is identified using a regression tree algorithm, including the following steps:

[0028] Suppose an input set of data is (X,Y)={(x1,y1),(x2,y2),(x3,y3),...,(x n ,y n For all initial data points, calculate the squared error. in Represents the mean of Y

[0029] For each possible segmentation point s, the data is divided into a left subset X. L Y L With right subset X R Y R Calculate the mean and squared error for each subset separately. The overall weighted squared error for the segment point s is:

[0030] Among them, MSE L The squared error of the left subset, MSE R The right subset squared error;

[0031] Among all possible segmentation points, select the segmentation point s that minimizes the weighted squared error. * =minMSE(s) is the optimal segmentation point, then the optimal segmentation point s * This is the reactive power compensation current segmentation point α.

[0032] Preferably, the reactive power compensation current control parameters K1 and K2, and the inverter current limiting control parameter K... MAX Negative sequence control parameter K N Identification is performed using least squares fitting, including the following steps:

[0033] At segment point s * At the point s, ensure that the fitting function at both ends is at the piecewise point s. * The function values ​​are the same, that is:

[0034] Where a1 and b1 are the parameters of the linear fitting function before the segmentation point, and b2 is the parameter of the linear fitting function after the segmentation point;

[0035] Within each segment, let i be the starting point and j be the ending point. For each segment [i,j], the least squares method is used for linear fitting, and the fitting equation is: y = ax + b, where the slope a and intercept b are obtained by minimizing the following sum of squared errors:

[0036] Where, x k y is the independent variable of the input data. k The dependent variable is the input data.

[0037] Find the values ​​of a and b that minimize E(a,b):

[0038] Among them, the reactive power compensation current control parameter fitting satisfies: K1=-a1, K2=b2, and the negative sequence control parameter K N =a;

[0039] The mathematical model of active power output current is transformed as follows:

[0040] when To determine if the inverter has entered current-limiting mode, assume there are n sets of data under current-limiting conditions. Then we can obtain n Ks. MAXi Value, for all K MAXi Take the average:

[0041] All K MAXi The average value is the inverter current limiting control parameter K. MAX .

[0042] Preferably, step S4 includes the following steps:

[0043] The identification results are obtained by using key control parameters, and the inverter output short-circuit current I is obtained. DG With grid connection point U PCC Relationship I DG =f(U PCC Therefore, an equivalent fault model of inverter-type distributed power sources during low-voltage ride-through is established.

[0044] Preferably, step S5 includes the following steps:

[0045] S51. For a normal component network, the inverter-type distributed generation is considered as a PQ node, and its current is taken as the output current during normal operation. The node voltage equation is then:

[0046] Among them, U 0 I is a column vector consisting of the normal components of the voltages at each node. 0 The column vector formed by the normal components of the injected current at each node. Let be the normal voltage component at node i. Let f be the injected current at node i, where i = 1, 2, ..., f, ..., n, Z be the impedance matrix of the power grid nodes, f be the faulty node, and n be the total number of network nodes.

[0047] S52. For the faulty component network, the synchronous power supply and inverter-type distributed power supply are set to zero, and a reverse current source is added at the fault point. The node voltage equation is:

[0048] Where ΔU is the column vector composed of the fault components of the voltage at each node, I f The column vector formed by the fault components of the injected current at each node. Let be the voltage fault component at node i. Inject current into the fault point;

[0049] The injected current at the fault point is:

[0050] Among them, Z ff Let z be the self-impedance of node f. f The fault resistor;

[0051] S53. Superimpose the normal voltage components and fault voltage components of each node to obtain the voltage of each node after the fault.

[0052] Based on the negative sequence current under asymmetrical faults, the formula for calculating the output short-circuit current of the i-th distributed power source is as follows:

[0053] The short-circuit current output by N distributed power sources is:

[0054] S54. Repeat S51 to S53 until the grid connection point voltage of the inverter-type distributed power source meets the following convergence condition in the two consecutive calculations:

[0055] in, ε is the voltage at the grid connection point of the distributed power source after the kth iteration, and ε is the convergence accuracy.

[0056] S55. Calculate the fault current of each branch in the distribution network based on the voltage of each node.

[0057] Preferably, step S6 includes the following steps:

[0058] S61. Obtain the parameters of each component in the power grid and set the number of Monte Carlo simulations;

[0059] S62. Based on the number of Monte Carlo simulations, multiple random faults are generated.

[0060] S63. Calculate the generated short-circuit current under each fault using the output short-circuit current calculation method in S5 in parallel.

[0061] S64. Based on the short-circuit current under each fault, obtain the voltage sag of the bus of interest under different faults.

[0062] S65. Based on the voltage sag amplitude of the bus of interest under multiple faults, the expected voltage sag amplitude of the bus of interest is calculated using the following formula:

[0063] Where N is the number of Monte Carlo simulations, and V m The voltage sag of the bus of interest is calculated for the m-th fault.

[0064] The beneficial effects of this invention are:

[0065] (1) This invention combines regression tree algorithm, least squares method and mathematical model derivation to propose a method for identifying control parameters of inverter distributed power source during low voltage ride-through. The acquisition of control parameters provides a more reliable basis for the calculation of short-circuit current, the evaluation of voltage sag, and the setting of relay protection current.

[0066] (2) This invention utilizes the identification results of key control parameters for low voltage ride-through of distributed power sources to establish a short-circuit current calculation model for multiple inverter-type distributed power sources connected to the system, corrects the calculation formula for short-circuit current of inverter-type distributed power sources, and improves the accuracy of voltage sag assessment. Attached Figure Description

[0067] Figure 1 is a flowchart of the voltage sag assessment method based on low voltage ride-through control parameter identification according to an embodiment of the present invention;

[0068] Figure 2 is a schematic diagram of the relationship between reactive power compensation current control parameters in an embodiment of the present invention;

[0069] Figure 3 is a schematic diagram of the relationship between the active power output current control parameters in an embodiment of the present invention;

[0070] Figure 4 is a schematic diagram of the relationship between the negative sequence control current control parameters in an embodiment of the present invention;

[0071] Figure 5 is a schematic diagram of the fault equivalent model of the inverter-type distributed power source according to an embodiment of the present invention;

[0072] Figure 6 is an equivalent circuit diagram of a three-sequence network under asymmetric fault according to an embodiment of the present invention;

[0073] Figure 7 is a schematic diagram of the normal component network and fault component network of the distributed power supply access in an embodiment of the present invention. Detailed Implementation

[0074] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided with reference to the accompanying drawings and embodiments.

[0075] The voltage sag assessment method based on low voltage ride-through control parameter identification is shown in Figure 1.

[0076] First, the key control parameters for low voltage ride-through are identified based on the regression tree algorithm and the least squares method.

[0077] S1. Based on the negative sequence current under asymmetrical fault, establish a mathematical model of the output current of the inverter-type distributed power source and determine the key control parameters to be identified.

[0078] The fault current characteristics of inverter-type distributed generation (DGG) are related to its low-voltage ride-through (LVRT) strategy. Based on this control strategy, a corresponding equivalent short-circuit calculation model for the DGG is established. Generally, the short-circuit calculation model for an inverter-type DGG is equivalent to a controlled current source. Considering the LVRT control strategy for negative-sequence current under asymmetrical faults, its output current calculation model is as follows:

[0079] The mathematical model expression for reactive power compensation current is as follows:

[0080] in, For the positive sequence component of reactive power compensation current, U pcc I is the grid connection point voltage. N Let K1 and K2 be the rated current, K1 and K2 be the reactive power compensation current control parameters, and α be the reactive power compensation current segmentation point. The magnitudes of K1 and K2 must be continuous at the segmentation point. The parameters to be identified in the reactive power compensation current mathematical model include the reactive power compensation current segmentation point α and the reactive power compensation current control parameters K1 and K2.

[0081] The mathematical model expression for active power output current is as follows:

[0082] in, K represents the positive-sequence component of the active output current. MAX These are the inverter current-limiting control parameters. The parameter to be identified in the active power output current mathematical model is the inverter current-limiting control parameter K. MAX .

[0083] The mathematical model expression for negative sequence control current is as follows:

[0084] in, This is the negative sequence control current on the d-axis. K is the negative sequence control current on the q-axis. N For negative sequence control parameters, The negative sequence component of the grid connection point voltage along the d-axis. This represents the negative q-axis component of the grid connection point voltage. This represents the positive-sequence d-axis component of the grid-connected point voltage. The parameter to be identified in the negative-sequence control current mathematical model is the negative-sequence control parameter K. NIn this implementation, the value range is [-1, 1]. When an asymmetrical fault occurs in the power grid, a negative sequence component will be generated in the system. The negative sequence control current can suppress the second harmonic fluctuations of active and reactive power in the power grid.

[0085] S2. Acquire online monitoring data and convert the voltage and current in the monitoring data into positive and negative sequence components in the dq coordinate system.

[0086] Three-phase voltage and current can be obtained through online monitoring data, and the rotation factor can be used. Symmetrical component transformation is performed on the three-phase voltage and three-phase current to separate the positive and negative sequences of the three-phase voltage and current:

[0087] in, This represents the positive sequence component of the phase a voltage or current. p is the negative sequence component of phase a voltage or current. a For phase a voltage or current, p b For phase b voltage or current, p c This refers to the voltage or current of phase c.

[0088] The positive and negative sequence of phase a voltage and current is separated by equation (4). The positive and negative sequence components of phases b and c satisfy the three-phase symmetry with the positive and negative sequence components of phase a. After separating the positive and negative sequence of the three-phase voltage and current, the voltage and current in the three-phase synchronous rotating abc coordinate system can be transformed into the voltage and current in the two-phase synchronous rotating dq coordinate system by equation (5) Park transformation. The positive and negative sequence components of the monitored three-phase voltage and current can be transformed into the positive and negative sequence components of the voltage and current in the dq coordinate system by equations (5) and (6) Park transformation. The dq coordinate system can convert the three-phase AC input or output signal into DC quantity on the dq axis, which is convenient for controlling the inverter.

[0089] The positive-order component Park transform is shown below:

[0090] The negative-order component Park transform is shown below:

[0091] Where, p + d p represents the positive sequence component of voltage or current along the d-axis. + q p represents the positive-sequence component of the voltage or current along the q-axis. + b p represents the positive sequence component of phase b voltage or current. + c p is the negative sequence component of the c-phase voltage or current. - d p represents the negative sequence component of the voltage or current along the d-axis. -q p is the negative sequence component of the voltage or current along the q-axis. - b p is the negative sequence component of phase b voltage or current. - c Let θ be the negative sequence component of phase c voltage or current, and let θ be the rotation angle of the abc coordinate system relative to the dq coordinate system.

[0092] S3. Key control parameters are identified based on regression tree algorithm and least squares method.

[0093] Figure 2 shows the relationship between the grid connection point voltage and reactive power compensation current for identifying key control parameters for low voltage ride-through. When the grid connection point voltage U pcc The maximum reactive power compensation current that the inverter can provide when the current is less than point α is defined as the segment point α. As shown in the graph, the change in K1 affects the slope of the straight line. Fitting α... pcc The slope of a straight line with a value less than 0.9 can be used to identify the control parameter K1. The control parameter K1 ranges from 1.5 to 3. Since the two straight lines are continuous at the segmentation point α, that is, the three satisfy K2 = K1(0.9-α), the control parameter K2 can be obtained when the segmentation point α and the control parameter K1 are identified.

[0094] To ensure sufficient reactive current support during low-voltage ride-through, the active power output current must meet the following requirements. Figure 3 shows the current limiting control parameter K. MAX It is a circular arc curve radius, through and The relationship can be used to identify K MAX .

[0095] To facilitate the identification of negative-order control parameters, the independent variable N is defined. d N q To describe the relationship in equation (3), where Therefore, equation (3) can be simplified to:

[0096] Through the The slope fitting of the straight line can identify the negative-order control parameter K. N .

[0097] The reactive power compensation current segmentation point α is identified using a regression tree algorithm. This algorithm finds the optimal segmentation point based on the input dataset, thus identifying the parameter α. A tree structure is constructed by recursively dividing the dataset into several subsets. The goal of the partitioning is to make the output values ​​of each subset (child node) as similar as possible, achieved by minimizing the variance within each subset. ​

[0098] Suppose an input set of data is (X,Y)={(x1,y1),(x2,y2),(x3,y3),…,(x n ,y n For all initial data points, calculate the squared error. in Represents the mean of Y

[0099] For each possible segmentation point s, the data is divided into a left subset X. L Y L With right subset X R Y R Calculate the mean and squared error of each subset, and the mean of the left subset. Left subset squared error right subset mean Right subset squared error The overall weighted squared error of segment point s is:

[0100] Among all possible segmentation points, select the segmentation point s that minimizes the weighted squared error. * =minMSE(s) is the optimal segmentation point, then the optimal segmentation point s * This is the reactive power compensation current segmentation point α.

[0101] Reactive power compensation current control parameters K1 and K2, inverter current limiting control parameter K MAX Negative sequence control parameter K N Identification is performed using least squares fitting.

[0102] In piecewise linear fitting of a parametric model, the fitted curve should remain continuous at the piecewise points. Therefore, at the piecewise point s*, the fitted function at both ends should be ensured to remain continuous at the piecewise point s*. * The function values ​​are the same, that is:

[0103] Where a1 and b1 are the parameters of the linear fitting function before the segmentation point, and b2 is the parameter of the linear fitting function after the segmentation point. This continuity constraint ensures the continuity of the fitted curve.

[0104] Within each segment, let i be the starting point and j be the ending point. For each segment [i,j], the least squares method is used for linear fitting, and the fitting equation is: y = ax + b, where the slope a and intercept b are obtained by minimizing the following sum of squared errors:

[0105] Where, xk y is the independent variable of the input data. k The dependent variable is the input data.

[0106] To find the values ​​of a and b that minimize E(a,b), take the partial derivatives of each parameter with respect to 0:

[0107] Solving the equation yields:

[0108] Among them, the reactive power compensation current control parameter fitting satisfies: K1=-a1, K2=b2, and the negative sequence control parameter K N =a;

[0109] The mathematical model of active power output current is transformed as follows:

[0110] when To determine if the inverter has entered current-limiting mode, a set of data under current-limiting mode is provided. satisfy K MAX We can find a K. MAX1 Assume there are n sets of data under current limiting conditions. Then we can obtain n Ks. MAXi Value, for all K MAXi Take the average:

[0111] All K MAXi The average value is the inverter current limiting control parameter K. MAX .

[0112] After identifying the key control parameters for low voltage ride-through using the parameter identification module, the voltage sag is evaluated by considering the connection of multiple inverter-type distributed power sources.

[0113] S4. Based on the identification results of key control parameters, establish an equivalent fault model for inverter-type distributed power sources during low-voltage ride-through.

[0114] The identification results are obtained by using key control parameters, and the inverter output short-circuit current I is obtained. DG With grid connection point U PCC Relationship I DG =f(U PCC Therefore, a fault equivalent model of the inverter-type distributed power source during the low voltage ride-through process, as shown in Figure 5, was established.

[0115] In traditional short-circuit current calculations for distributed generation systems with inverters, the formula for calculating the short-circuit current of the distributed generation system is as follows: The calculation of the short-circuit current in the formula does not consider the negative-sequence component of the inverter output short-circuit current under asymmetrical fault conditions, as shown in Figure 6. In Figure 6(a), the positive-sequence network is represented; in Figure 6(b), the negative-sequence network is represented; and in Figure 6(c), the zero-sequence network is represented. This application's embodiment employs an iterative calculation method considering negative-sequence current injection. The equivalent current source of the inverter-type distributed power supply will appear in the negative-sequence component network, correcting the short-circuit current output by the inverter-type distributed power supply to...

[0116] S5. Calculate the voltage of each node in the inverter-type distributed power source after the fault based on the fault equivalent model, and then obtain the output short-circuit current of the inverter-type distributed power source. When considering multiple distributed power sources connected as shown in Figure 7, the short-circuit currents output by each distributed power source are superimposed. The specific steps are as follows:

[0117] S51. For a normal component network, the inverter-type distributed generation is considered as a PQ node, and its current is taken as the output current during normal operation. The node voltage equation is then:

[0118] Among them, U 0 I is a column vector consisting of the normal components of the voltages at each node. 0 The column vector formed by the normal components of the injected current at each node. Let be the normal voltage component at node i. Let be the injected current at node i, i = 1, 2, ..., f, ..., n, Z be the impedance matrix of the power grid nodes, f be the faulty node, and n be the total number of network nodes.

[0119] S52. For the faulty component network, the synchronous power supply and inverter-type distributed power supply are set to zero, and a reverse current source is added at the fault point. The node voltage equation is:

[0120] Where ΔU is the column vector composed of the fault components of the voltage at each node, I f The column vector formed by the fault components of the injected current at each node. Let be the voltage fault component at node i. Inject current into the fault point.

[0121] The voltage fault component at the fault point is:

[0122] Among them, Z ff Let f be the self-impedance of node f.

[0123] The voltage at the fault point is:

[0124] The injected current at the fault point is then obtained as:

[0125] Substituting equation (18) into equation (15) yields the voltage fault components at each node.

[0126] S53. Superimpose the normal voltage components and fault voltage components of each node to obtain the voltage of each node after the fault, and correct the output short-circuit current of the inverter-type distributed power source according to equations (1), (2), and (3). The magnitude of the short-circuit current output by the i-th distributed power source is... N distributed power sources output short-circuit current I DGs for:

[0127] S54. Repeat S51 to S53 until the grid connection point voltage of the inverter-type distributed power source meets the following convergence condition in the two consecutive calculations:

[0128] in, ε represents the voltage at the grid connection point of the distributed power source after the k-th iteration, and ε is the convergence accuracy.

[0129] S55. Calculate the fault current of each branch in the distribution network using the voltage of each node in equation (16).

[0130] S6. Fault voltage sag assessment based on Monte Carlo method and output short-circuit current of inverter-type distributed power source.

[0131] S61. Obtain the parameters of each component in the power grid and set the number of Monte Carlo simulations.

[0132] S62. Based on the number of Monte Carlo simulations, multiple random faults are generated.

[0133] S63. Calculate the generated short-circuit current under each fault using the output short-circuit current calculation method in S5 in parallel.

[0134] S64. Based on the short-circuit current under each fault, obtain the voltage sag of the bus of interest under different faults using equation (17).

[0135] S65. Based on the voltage sag amplitude of the bus of interest under multiple faults, the expected voltage sag amplitude of the bus of interest is calculated using the following formula:

[0136] Where N is the number of Monte Carlo simulations, and V mThe voltage sag of the bus of interest is calculated for the m-th fault. The basic principles, main features, and advantages of the present invention have been shown and described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A method for voltage sag assessment based on low voltage ride through control parameter identification, characterized in that, Includes the following steps: S1. Based on the negative sequence current under asymmetrical fault, establish a mathematical model of the output current of the inverter-type distributed power source and determine the key control parameters to be identified. S2. Acquire online monitoring data and convert the voltage and current in the monitoring data into positive and negative sequence components in the dq coordinate system; S3. Identify key control parameters based on regression tree algorithm and least squares method; S4. Based on the identification results of key control parameters, establish an equivalent fault model for inverter-type distributed power sources during low-voltage ride-through. S5. Calculate the voltage of each node of the inverter-type distributed power source after the fault based on the fault equivalent model, and then obtain the output short-circuit current of the inverter-type distributed power source. S6. Fault voltage sag assessment based on Monte Carlo method and output short-circuit current of inverter-type distributed power source.

2. The method for voltage sag assessment based on low voltage ride through control parameter identification according to claim 1, characterized in that, The mathematical model for the output current of the inverter-type distributed power source includes a mathematical model for reactive power compensation current, a mathematical model for active power output current, and a mathematical model for negative sequence control current.

3. The voltage sag assessment method based on low voltage ride-through control parameter identification according to claim 2, characterized in that, The reactive compensation current mathematical model expression is as follows: wherein U is the positive sequence component of the reactive compensation current pcc I is the grid-connected point voltage N I is the rated current, K1, K2 are the reactive compensation current control parameters, and α is a reactive compensation current segmentation point. The sizes of K1 and K2 satisfy continuity at the segmentation point. The active output current mathematical model expression is as follows: wherein K for active output current positive sequence component MAX K for inverter current limit control parameter; The negative sequence control current mathematical model expression is as follows: wherein for negative sequence control current on d-axis, For q-axis negative sequence control current, K N For negative sequence control parameter, for the d-axis negative sequence component of the grid point voltage, for the q-axis negative sequence component of the grid point voltage, This represents the positive sequence component of the voltage at the grid connection point along the d-axis.

4. The method for voltage sag assessment based on low voltage ride through control parameter identification according to claim 3, characterized in that, The key control parameters to be identified include reactive compensation current control parameters K1 and K2, reactive compensation current segmentation point a, inverter current limiting control parameter K MAX , negative sequence control parameter K N .

5. The voltage sag assessment method based on low voltage ride-through control parameter identification according to claim 4, characterized in that, S2 includes the following steps: Using rotation factor Symmetrical component transformation is performed on the three-phase voltage and three-phase current to separate the positive and negative sequences of the three-phase voltage and current: in, This represents the positive sequence component of the phase a voltage or current. Vp is the negative sequence component of the a-phase voltage or current, p a Vp is the a-phase voltage or current, p b Vp is the b-phase voltage or current, p c Vp is the c-phase voltage or current; The positive and negative sequence components of the three-phase voltage and current are transformed into positive and negative sequence components of voltage and current in the dq coordinate system through positive sequence component Park transformation and negative sequence component Park transformation. The positive-order component Park transform is shown below: The negative-order component Park transform is shown below: where p + d is the positive sequence component of the voltage or current of the d-axis, p + q is the positive sequence component of the voltage or current of the q-axis, p + b is the positive sequence component of the voltage or current of the b-phase, p + c is the negative sequence component of the voltage or current of the c-phase, p - d is the negative sequence component of the voltage or current of the d-axis, p - q is the negative sequence component of the voltage or current of the q-axis, p - b is the negative sequence component of the voltage or current of the b-phase, p - c is the negative sequence component of the voltage or current of the c-phase, and θ is the rotation angle of the abc coordinate system with respect to the dq coordinate system.

6. The voltage sag assessment method based on low voltage ride-through control parameter identification according to claim 5, characterized in that, The reactive power compensation current segmentation point α is identified using a regression tree algorithm, including the following steps: Suppose a set of input data is (X,Y)={(x 1, y1),(x 2, y2),(x 3, y3),…,(x n, y n For all initial data points, calculate the squared error. in Represents the mean of Y For each possible segmentation point s, the data is divided into a left subset X. L Y L With right subset X R Y R Calculate the mean and squared error for each subset separately. The overall weighted squared error for the segment point s is: where MSE L is the left subset squared error, MSE R is the right subset squared error; Among all possible segmentation points, the segmentation point s that minimizes the weighted squared error is selected * = min MSE(s) is the optimal segmentation point, then the optimal segmentation point s * is the reactive compensation current segmentation point a.

7. The voltage sag assessment method based on low voltage ride-through control parameter identification according to claim 6, characterized in that, The reactive compensation current control parameters K1 and K2, inverter current limiting control parameters K MAX , negative sequence control parameters K N The identification is performed by least square fitting, including the following steps: At segment point s * At the point s, ensure that the fitting function at both ends is at the piecewise point s. * The function values ​​are the same, that is: Where a1 and b1 are the parameters of the linear fitting function before the segmentation point, and b2 is the parameter of the linear fitting function after the segmentation point; Within each segment, let i be the starting point and j be the ending point. For each segment [i,j], the least squares method is used for linear fitting, and the fitting equation is: y = ax + b, where the slope a and intercept b are obtained by minimizing the following sum of squared errors: where x k is the independent variable of the input data, y k is the dependent variable of the input data; Find the values ​​of a and b that minimize E(a,b): wherein the reactive compensation current control parameters fit to satisfy: K1 = -a1, K2 = b2, negative sequence control parameters K N = a; The mathematical model of active power output current is transformed as follows: when To determine if the inverter has entered current-limiting mode, assume there are n sets of data under current-limiting conditions. Then we can obtain n Ks. MAXi Value, for all K MAXi Take the average: All K MAXi The mean of all K MAX is the inverter current limit control parameter K .

8. The voltage sag assessment method based on low voltage ride-through control parameter identification according to claim 7, characterized in that, S4 includes the following steps: The identification result is obtained through a key control parameter, and an inverter output short-circuit current I is obtained DG The relationship between the grid-connected point U PCC and the inverter output short-circuit current I DG =f(U PCC ), thereby establishing a fault equivalent model of the inverter-type distributed power supply in the low-voltage ride-through process.

9. The voltage sag assessment method based on low voltage ride-through control parameter identification according to claim 8, characterized in that, S5 includes the following steps: S51. For a normal component network, the inverter-type distributed generation is considered as a PQ node, and its current is taken as the output current during normal operation. The node voltage equation is then: Among them, U 0 I is a column vector consisting of the normal components of the voltages at each node. 0 The column vector formed by the normal components of the injected current at each node. Let be the normal voltage component at node i. Let f be the injected current at node i, where i = 1, 2, ..., f, ..., n, Z be the impedance matrix of the power grid nodes, f be the faulty node, and n be the total number of network nodes. S52. For the faulty component network, the synchronous power supply and inverter-type distributed power supply are set to zero, and a reverse current source is added at the fault point. The node voltage equation is: Where ΔU is the column vector composed of the fault components of the voltage at each node, I f The column vector formed by the fault components of the injected current at each node. Let be the voltage fault component at node i. Inject current into the fault point; The injected current at the fault point is: wherein Z ff is the self-impedance of node f, z f is the fault resistance; S53. Superimpose the normal voltage components and fault voltage components of each node to obtain the voltage of each node after the fault. Based on the negative sequence current under asymmetrical faults, the formula for calculating the output short-circuit current of the i-th distributed power source is as follows: The short-circuit current output by N distributed power sources is: S54. Repeat S51 to S53 until the grid connection point voltage of the inverter-type distributed power source meets the following convergence condition in the two consecutive calculations: in, ε is the voltage at the grid connection point of the distributed power source after the kth iteration, and ε is the convergence accuracy. S55. Calculate the fault current of each branch in the distribution network based on the voltage of each node.

10. The voltage sag assessment method based on low voltage ride-through control parameter identification according to claim 9, characterized in that, S6 includes the following steps: S61. Obtain the parameters of each component in the power grid and set the number of Monte Carlo simulations; S62. Based on the number of Monte Carlo simulations, multiple random faults are generated. S63. Calculate the generated short-circuit current under each fault using the output short-circuit current calculation method in S5 in parallel. S64. Based on the short-circuit current under each fault, obtain the voltage sag of the bus of interest under different faults. S65. Based on the voltage sag amplitude of the bus of interest under multiple faults, the expected voltage sag amplitude of the bus of interest is calculated using the following formula: where N is the number of Monte Carlo simulations, V m is the voltage sag magnitude of the bus of interest calculated for the mth fault.

Images