Method and system for identifying high-low voltage ride-through control parameters of doubly-fed fan

Abstract

The invention relates to a double-fed fan high-low voltage ride-through control parameter identification method and system, and the method comprises the steps: analyzing a high-low voltage ride-through threshold value according to simulation data outputted by a pre-constructed double-fed fan control structured model; the high-low voltage ride-through threshold value is used as a voltage initial value of high-low voltage ride-through control, a high-low voltage ride-through control mode of the double-fed fan is identified, and a high-low voltage ride-through strategy is formulated; and on the basis of a high-low voltage ride-through strategy, identifying each parameter in the control link to obtain each control parameter value in the control link. According to the scheme, a set of complete method for identifying the fault ride-through control mode and the control parameters is formed based on the doubly-fed fan control structured model, and identification of the control parameters can be rapidly and effectively achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power control, and in particular relates to a method and system for identifying high and low voltage ride-through control parameters of a doubly-fed wind turbine. Background Art

[0002] Since the beginning of the 21st century, energy shortages and environmental pollution have become increasingly serious. The "double carbon" goals of "carbon peak and carbon neutrality" have been proposed, and the increase in the penetration rate of new energy units has become a general trend. In recent years, the installed capacity of new energy generating units such as wind power has steadily increased, which has brought challenges to the safe and reliable operation of traditional power grid structures. At present, the installed capacity of new energy in Shaanxi Power Grid has reached 17 million kilowatts, and it is expected to reach 60 million kilowatts in 2025, of which 80% of new energy is concentrated in northern Shaanxi, with wind and solar power each accounting for 50%.

[0003] For the parameter identification of the electrical and control parts of the wind turbine, the stator-rotor mutual inductance is estimated offline first, and then the parameters are further identified based on the offline data. This method is only applicable to theoretical analysis and cannot be effectively verified. The mutual inductance derived from the formula must have errors. In addition, the results of the identification by this method have large errors, and the identification errors of the transformation ratio and mutual inductance are both over 10%. In the electromechanical transient model of the doubly fed wind turbine, the generator part adopts a second-order model that ignores the stator voltage change, and the machine-side and grid-side converters use first-order inertia links instead of power electronic devices. The simulation results of the established model have small deviations and the simulation speed has been greatly improved. However, for the analysis of the grid voltage change or the response of the wind power generation system when the wind speed changes, whether the generator model can omit the stator transient process and whether accurate results can be obtained still need further research. For the wind turbine, a single mass block model is used, and the generator adopts a 5th-order model. The output power of the generator is used as the observed quantity to analyze the identifiability of each parameter of the generator. The research conclusion shows that the reactance of the stator and rotor cannot be identified at the same time, and the resistance of the stator and rotor is less sensitive than the inductance value, so parameter identification is difficult. In addition, whether multiple observed quantities can be identified requires further research. For the identification of some parameters of the wind turbine, such as the inductance, mutual inductance, active control PI parameters and reactive control PI parameters on the stator side, the generator model to be identified is a 1st-order model, the inner current loop is ignored and replaced by the inertia link, and the wind speed change is used as the excitation for identification. The wind speed change stimulates the slow dynamic mode of the system. The sensitivity of electrical parameters is generally small and difficult to identify, while the sensitivity of mechanical parameters is generally large and easy to identify. The single mass block model is used to identify the wind turbine drive system, and the fault data is used for the generator parameter identification. The generator model is a mathematical model that ignores the stator transient. The identification of the relevant parameters of the drive system under wind speed excitation is carried out, but the wind turbine is different from the synchronous machine. Whether the stator transient can be ignored needs further research.

[0004] In terms of controller parameter identification, existing literature has proposed that various controller parameters can be identified as much as possible through various types of disturbances such as wind speed changes, grid-side faults, and step signals applied to the reference voltage of the controller. However, only the parameters of the machine-side converter are identified. For controllers with cascades, disturbances are first applied to the measurement signal of the next-stage controller, while shielding the changes in the measurement signal of the previous-stage controller. Each controller parameter is individually identified without modifying the controller structure, and the parameter identification accuracy is improved through multiple rounds of iterations. However, when identifying parameters based on the detailed model of the original controller, due to the cascade of the proportional-integral links of the controller and the intersection of input quantities, more experiments are required and the operation is cumbersome. Summary of the invention

[0005] In order to remedy the above defects, the present invention proposes a method and system for identifying high and low voltage ride-through control parameters of a doubly-fed wind turbine, which can identify the fault ride-through control mode and control parameters through simple and fast operations, thereby effectively realizing the identification of control parameters.

[0006] In order to achieve the above-mentioned object of the invention, the present invention adopts the following technical scheme: A method for identifying high and low voltage ride-through control parameters of a doubly-fed wind turbine, the method comprising: Analyze high and low voltage ride-through thresholds based on simulation data output by the pre-built double-fed wind turbine control structured model; The high and low voltage ride-through thresholds are used as the voltage initial values ​​for high and low voltage ride-through control, the high and low voltage ride-through control modes of the double-fed wind turbine are identified, and high and low voltage ride-through strategies are formulated; Based on the high and low voltage ride-through strategies, various parameters in the control link are identified to obtain various control parameter values ​​in the control link.

[0007] Preferably, the pre-built doubly-fed wind turbine control structured model includes: an MPPT control module, a converter control module, a Chopper protection control module, a Crowbar protection control module and a pitch control module; wherein the converter control module includes a machine-side converter control unit and a grid-side converter control unit; the machine-side converter control unit is used for power outer loop control and current inner loop control; the grid-side converter control unit is used for DC voltage outer loop control, reactive power outer loop control and current inner loop control.

[0008] Preferably, the analyzing the high and low voltage ride-through thresholds includes: By identifying the voltage values ​​at which the doubly fed wind turbine enters and exits the high and low voltage ride through when the fault occurs, the high and low voltage ride through entry threshold and high and low voltage ride through exit threshold of the doubly fed wind turbine are determined.

[0009] Furthermore, the high-low penetration entry threshold and the high-low penetration exit threshold of the doubly fed wind turbine are judged to include: Measure the steady-state reactive current value. When the steady-state reactive current value satisfies the condition that the dynamic reactive current increment of the wind farm should respond to the voltage change of the grid connection point, calculate the difference between the real-time reactive current and the reactive current before the fault, that is, the reactive current increment; when the reactive current increment exceeds the preset threshold of the reactive current for the doubly fed wind turbine to enter the fault ride-through control, the current moment is considered to be the moment of fault occurrence; according to NB / T 31053-2021 "Verification Procedure for Electrical Simulation Model of Wind Turbine Generators", measure the reactive current at each moment within 140ms after the fault occurs I q and terminal voltage U t , when the formula ( I q - I q0 ) / (0.9- U t )≥1.5 or ( I q0 - I q ) / ( U t -1.1)≥1.5, then the terminal voltage at the current moment U t It is defined as the low penetration threshold or high penetration threshold of the doubly fed wind turbine; I qt is the reactive current per unit value at the current moment, I q0 is the per unit value of the steady-state reactive current before the fault, U t It is the unit value of the terminal voltage.

[0010] Furthermore, the dynamic reactive current increment of the doubly-fed wind turbine generator set should respond to the voltage change condition at the grid connection point: (1.16) (1.17) Where Δ I t is the reactive current increment at the grid connection point. is the reactive current increment at the grid connection point. K 1 is the low-pass reactive current increment coefficient, I N is the rated current of the wind turbine, K 2 is the high-voltage reactive current increment coefficient, U t is the steady-state value of the fault voltage.

[0011] Furthermore, the difference between the real-time reactive current and the steady-state value is calculated by the following formula: (1.18) (1.19) In the formula I qt is the reactive current per unit value at the current moment, I q0 is the reactive current per unit value in steady-state operation before the fault, I q_lvrt For expression , I q_hvrt For expression .

[0012] Preferably, the high and low voltage ride-through thresholds are used as the initial voltage values ​​of the high and low voltage ride-through control to identify the high and low voltage ride-through control methods of the doubly fed wind turbine, including: selecting a time point after the initial voltage values ​​of the high and low voltage ride-through control, intercepting the steady-state segment during the fault of the doubly fed wind turbine, obtaining the average values ​​of active power / current and reactive power / current, and recording the voltage value corresponding to the steady-state segment, and then recording the active power / current, reactive power / current, and voltage under various fault depths, using the least squares method to perform function fitting on multiple groups of active and reactive results under different voltages according to the high and low voltage ride-through strategy, and identifying each parameter in the active power / current and reactive power / current expressions.

[0013] Preferably, the high and low voltage ride-through strategy includes identification of active current reference value and reactive current reference value; wherein, identification of active current reference value includes: selecting reference value basis, when specifying power as reference value, performing an identification once by the following formula: (1.20) In the formula, P lvrt is the low wear active power reference value, K P_lvrt is the low wear active power slope coefficient, P 0 is the per-unit value of active power during steady state, P set_lvrt A basic value is set for the low-break active power, and the low-break active power is always above this value.

[0014] When the current is specified as the reference value, an identification is performed using the following formula: (1.21) In the formula I plvrt is the low wear active current reference value, K1 is the low wear active current coefficient 1, K 2 is the low wear active current coefficient 2, U t is the steady-state value of the fault voltage, I p0 is the steady-state value of active current before the fault, I set_lvrt Set a basic value for the low-throughput active current, and the low-throughput active current is always above this value; The secondary identification is performed by the following formula: (1.22) In the formula A 1, A 2, A 3 are all secondary low-through active current coefficients; The identification of the reactive current reference value is performed once using the following predefined reactive current identification model expression: (1.23) In the formula K 3 and K 4 are the two coefficients of low-through reactive current, U lim is the fault ride-through voltage threshold, U t is the steady-state value of the fault voltage, I q0 is the steady-state reactive current value before the fault, I qset_lvrt Sets the basic value for the reactive current.

[0015] Furthermore, the identification of various parameters in the control link based on the high and low voltage ride-through strategy to obtain various control parameter values ​​in the control link includes: Difference equation for a single PI link: (1.24) In the formula x For the control link input, y For the control link output, K p1 , K i1 They are the proportional coefficient and integral coefficient of the PI link respectively, x ref To input reference value; Using the least squares method, b0 and b1 are calculated by the following formula, and the PI control parameter K is obtained by solving. p , K i (1.25) (1.26) In the formula y i is the ith output, a 1. b 0. b 1 is the three coefficients, Δt is the sampling step length; The difference equation of the double PI link: (1) In the formula x It is the input of the outer loop control link. y For the control link output, K p1 , K i1 The proportional coefficient and integral coefficient of the outer loop PI link respectively, K p2 , K i2 They are the proportional coefficient and integral coefficient of the inner loop PI link, x ref To input the reference value, z It is the input of inner loop control link; The parameter identification is performed offline, and the discretization and term shifting are obtained: (1.27) (1.28) In the formula a1 , a2 , b1 , b2 , b3 , b4 , b5 , b6 are coefficients, which can be expressed by control link parameters and constants. Δt is the sampling step length; Use the least squares method to solve the coefficients of the following matrix: (1.29) In the formula, y i represents the i-th output, u i Represents the i-th input.

[0016] A high and low voltage ride through control parameter identification system for a doubly fed wind turbine based on a structured model, the system comprising: An analysis module, used to analyze high and low voltage ride-through thresholds based on simulation data output by a pre-built double-fed wind turbine control structured model; An identification module is used to use the high and low voltage ride-through thresholds as voltage initial values ​​for high and low voltage ride-through control, identify high and low voltage ride-through control modes for the doubly-fed wind turbine, and formulate high and low voltage ride-through strategies; The acquisition module is used to identify various parameters in the control link based on the high and low voltage ride-through strategy and obtain various control parameter values ​​in the control link.

[0017] Compared with the closest prior art, the beneficial effects of the present invention are: Aiming at the high and low voltage ride-through control parameters of doubly fed wind turbine, a high and low voltage ride-through control parameter identification method and system of doubly fed wind turbine based on structured model is proposed. Combined with the structured electromagnetic transient model of doubly fed wind turbine, the control structure expression and fault ride-through control scheme are summarized.

[0018] A method for identifying high and low voltage ride-through control parameters of a doubly-fed wind turbine based on a structured model, comprising analyzing high and low voltage ride-through thresholds according to simulation data output by a pre-built doubly-fed wind turbine control structured model; using the high and low voltage ride-through thresholds as voltage initial values ​​for high and low voltage ride-through control, identifying high and low voltage ride-through control methods of the doubly-fed wind turbine, and formulating high and low voltage ride-through strategies; based on the high and low voltage ride-through strategies, identifying various parameters in the control link, and obtaining various control parameter values ​​in the control link. The above scheme is based on a structured doubly-fed wind turbine control model, forming a complete set of methods for identifying fault ride-through control methods and control parameters, which can quickly and effectively realize the identification of control parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A flow chart of a method for identifying high and low voltage ride through control parameters of a doubly-fed wind turbine based on a structured model provided in an embodiment of the present invention; Figure 2 A schematic diagram of the structure of a doubly-fed wind turbine generator provided in an embodiment of the present invention; Figure 3 It is a structural diagram of the overall structure of the doubly-fed wind turbine provided in an embodiment of the present invention; Figure 4 A schematic diagram of active power outer loop control provided in an embodiment of the present invention; Figure 5 A single PI control structure diagram of a machine-side converter provided in an embodiment of the present invention; Figure 6 A schematic diagram of the reactive power outer loop control of the dual PI link provided in an embodiment of the present invention; Figure 7 A flow chart of the high and low penetration threshold judgment provided in an embodiment of the present invention; Figure 8 A reactive current diagram at the end of high and low wear provided in an embodiment of the present invention; Fig. 9 This is a flow chart for identifying an active reference value provided in an embodiment of the present invention; Fig.10 A flowchart of identifying a reactive reference value provided in an embodiment of the present invention; Fig.11 This is a schematic diagram of a single-input, single-output, single-PI link provided in an embodiment of the present invention; Fig.12 A schematic diagram of a single-input single-output dual PI link provided in an embodiment of the present invention; Fig.13 The schematic diagram of low-breakdown parameter identification of the 2500 model provided in the embodiment of the present invention (a) is a schematic diagram of active power identification; (b) is a schematic diagram of active current identification; (c) is a schematic diagram of reactive power identification; (d) is a schematic diagram of reactive current identification; Fig.14 Schematic diagram of high-throughput parameter identification of the 2500 model provided in the embodiment of the present invention: (a) is a schematic diagram of active power identification; (b) is a schematic diagram of active current identification; (c) is a schematic diagram of reactive power identification; (d) is a schematic diagram of reactive current identification; Fig.15 The schematic diagram of low-breakdown parameter identification of the 3600 model provided in the embodiment of the present invention is as follows: (a) is a schematic diagram of active power identification; (b) is a schematic diagram of active current identification; (c) is a schematic diagram of reactive power identification; (d) is a schematic diagram of reactive current identification; Fig.16 Schematic diagram of high-throughput parameter identification of the 3600 model provided in the embodiment of the present invention: (a) is a schematic diagram of active power identification; (b) is a schematic diagram of active current identification; (c) is a schematic diagram of reactive power identification; (d) is a schematic diagram of reactive current identification; Fig.17 K provided in the embodiment of the present invention i Schematic diagram of parameter identification results; Fig.18 K provided in the embodiment of the present invention p Schematic diagram of parameter identification results; Fig.19 P provided in the embodiment of the present invention I Schematic diagram of link input and output; Fig. 20 K provided in the embodiment of the present invention i Schematic diagram of parameter identification results; Fig.21 K provided in the embodiment of the present invention p Schematic diagram of parameter identification results; Fig. 22 P provided in the embodiment of the present invention I Schematic diagram of link input and output. DETAILED DESCRIPTION

[0020] The specific implementation modes of the present invention are further described in detail below in conjunction with the accompanying drawings.

[0021] The present invention proposes a method for identifying high and low voltage ride-through control parameters of a doubly fed wind turbine based on a structured model. First, according to the output response of the doubly fed wind turbine during fault ride-through, the basic control structure and fault ride-through scheme of the doubly fed wind turbine are studied to obtain a doubly fed wind turbine control structure model; then, the commonalities of different doubly fed wind turbine control systems are summarized to form a structured control logic and a fault ride-through scheme; then, mathematical expressions of control objects under different fault ride-through schemes and structured control logic are constructed to construct a mathematical expression of the doubly fed wind turbine control system; finally, after multiple simulations and data statistics, software is used to identify various coefficients in the expression of the fault ride-through scheme, and by applying small disturbances, the input and output of the wind turbine control are extracted, and the various coefficients of the mathematical expression of the control system are identified by software, thereby obtaining the control parameter values ​​in the control link. According to the above identification method and system, a systematic doubly fed wind turbine fault ride-through control parameter identification scheme is formed, and based on the structured doubly fed wind turbine control model, a complete set of methods for identifying fault ride-through schemes and control parameters is formed.

[0022] like Figure 1 As shown, a method and system for identifying control parameters of a doubly-fed wind turbine high and low voltage ride-through based on a structured model is proposed, comprising the following steps: S1 analyzes the high and low voltage ride-through thresholds based on the simulation data output by the pre-built double-fed wind turbine control structured model; S2 uses the high and low voltage ride-through thresholds as the voltage initial values ​​for high and low voltage ride-through control, identifies the high and low voltage ride-through control methods of the double-fed wind turbine, and formulates high and low voltage ride-through strategies; Based on the high and low voltage ride-through strategies, S3 identifies various parameters in the control link and obtains various control parameter values ​​in the control link.

[0023] The above method steps are specifically implemented through the following scheme: In step S1, the pre-built doubly-fed wind turbine control structured model includes: an MPPT control module, a converter control module, a Chopper protection control module, a Crowbar protection control module and a pitch control module; wherein the converter control module includes a machine-side converter control unit and a grid-side converter control unit; the machine-side converter control unit is used for power outer loop control and current inner loop control; the grid-side converter control unit is used for DC voltage outer loop control, reactive power outer loop control and current inner loop control.

[0024] In step S1, analyzing the high and low voltage ride-through thresholds includes: By identifying the voltage values ​​at which the doubly fed wind turbine enters and exits the high and low voltage ride through when the fault occurs, the high and low voltage ride through entry threshold and high and low voltage ride through exit threshold of the doubly fed wind turbine are determined.

[0025] Among them, the high and low penetration entry thresholds and high and low penetration exit thresholds for judging the doubly fed wind turbine include: Measure the steady-state reactive current value. When the steady-state reactive current value satisfies the condition that the dynamic reactive current increment of the wind farm should respond to the voltage change of the grid connection point, calculate the difference between the real-time reactive current and the reactive current before the fault, that is, the reactive current increment; when the reactive current increment exceeds the preset threshold of the reactive current for the doubly fed wind turbine to enter the fault ride-through control, the current moment is considered to be the moment of fault occurrence; according to NB / T 31053-2021 "Verification Procedure for Electrical Simulation Model of Wind Turbine Generators", measure the reactive current at each moment within 140ms after the fault occurs I q and terminal voltage U t , when the formula ( I q - I q0 ) / (0.9- U t )≥1.5 or ( I q0 - I q ) / ( U t -1.1)≥1.5, then the terminal voltage at the current moment U t It is defined as the low penetration threshold or high penetration threshold of the doubly fed wind turbine; I qt is the reactive current per unit value at the current moment, I q0 is the per unit value of the steady-state reactive current before the fault, U t It is the unit value of the terminal voltage.

[0026] The wind farm dynamic reactive current increment should respond to the grid connection point voltage change condition as follows: (1.30) (1.31) Where Δ I t is the reactive current increment at the grid connection point. is the reactive current increment at the grid connection point. K 1 is the low-pass reactive current increment coefficient, which is usually between 1.5 and 3. I Nis the rated current per unit value, K 2 is the high-throughput reactive current increment coefficient, which is usually between 1.5 and 3. U t is the steady-state value of the fault voltage.

[0027] The difference between the real-time reactive current and the steady-state value is calculated by the following formula: (1.32) (1.33) In the formula I qt is the reactive current per unit value at the current moment, I q0 is the reactive current per unit value in steady-state operation, I q_lvrt The expression in formula 1.30 is , I q_hvrt The expression in formula 1.31 is .

[0028] In step S2, the high and low voltage ride-through thresholds are used as the initial voltage values ​​of the high and low voltage ride-through control, and the high and low voltage ride-through control methods of the doubly fed wind turbine are identified, including: selecting the time point after the initial voltage values ​​of the high and low voltage ride-through control, intercepting the steady-state segment during the fault of the doubly fed wind turbine, obtaining the average values ​​of active power / current and reactive power / current, and recording the voltage value corresponding to the steady-state segment, and then recording the active power / current, reactive power / current, and voltage under various fault depths, using the least squares method to perform function fitting on multiple groups of active and reactive results under different voltages according to the high and low voltage ride-through strategy, and identifying each parameter in the active power / current and reactive power / current expressions.

[0029] In step S2, the high and low voltage ride-through strategy includes identification of active current reference value and reactive current reference value; wherein, identification of active current reference value includes: selecting reference value basis, when specifying power as reference value, performing identification once by the following formula: (1.34) In the formula, P lvrt is the low wear active power reference value, K P_lvrt is the low wear active power slope coefficient, P 0 is the per-unit value of active power during steady state, P set_lvrt A basic value is set for the low-penetration active power. This value can be set according to actual conditions, and the low-penetration active power is always above this value.

[0030] When the current is specified as the reference value, an identification is performed using the following formula: (1.35) In the formula I plvrt is the low wear active current reference value, K 1 is the low-pass active current increment coefficient 1, K 2 is the low wear active current coefficient 2, K 1. K 2 can be assigned as needed, U t is the steady-state value of the fault voltage, I p0 is the steady-state value of active current before the fault, I set_lvrt A basic value is set for the low-throughput active current. This value can be set according to actual conditions, and the low-throughput active current is always above this value.

[0031] The secondary identification is performed by the following formula: (1.36) In the formula A 1, A 2, A 3 are all secondary low-power active current coefficients.

[0032] The identification of the reactive current reference value is performed once using the following predefined reactive current identification model expression: (1.37) In the formula K 3 and K 4 are the two coefficients of low-power reactive current, U lim is the fault ride-through voltage threshold, U t is the steady-state value of the fault voltage, I q0 is the steady-state reactive current value before the fault, I qset_lvrt Sets the basic value for the reactive current.

[0033] In step S3, based on the high and low voltage ride-through strategy, various parameters in the control link are identified, and the values ​​of various control parameters in the control link are obtained, including: Difference equation for a single PI link: (1.38) In the formula x For the control link input, y For the control link output, K p1 , K i1They are the proportional coefficient and integral coefficient of the PI link respectively, x ref To input reference value.

[0034] Using the least squares method, b0 and b1 are calculated by the following formula, and the PI control parameter K is obtained by solving. p , K i (1.39) (1.40) In the formula y i is the ith output, a 1. b 0. b 1 is the three coefficients, Δt is the sampling step length.

[0035] The difference equation of the double PI link: (1) The parameter identification is performed offline, and the discretization and term shifting are obtained: (1.41) (1.42) In the formula a1 , a2 , b1 , b2 , b3 , b4 , b5 , b6 are coefficients, which can be expressed by control link parameters and constants. Δt is the sampling step length.

[0036] Use the least squares method to solve the coefficients of the following matrix: (1.43) In the formula, y i represents the i-th output, u i Represents the i-th input.

[0037] Based on the same inventive technical concept, the present invention also provides a method and system for identifying high and low voltage ride-through control parameters of a doubly-fed wind turbine based on a structured model, comprising: An analysis module, used to analyze high and low voltage ride-through thresholds based on simulation data output by a pre-built double-fed wind turbine control structured model; An identification module is used to use the high and low voltage ride-through thresholds as voltage initial values ​​for high and low voltage ride-through control, identify high and low voltage ride-through control modes for the doubly-fed wind turbine, and formulate high and low voltage ride-through strategies; The acquisition module is used to identify various parameters in the control link based on the high and low voltage ride-through strategy and obtain various control parameter values ​​in the control link.

[0038] Embodiment 1: Step 1: Structured model of fault ride-through control for DFIG The structure of a doubly-fed wind turbine generator is as follows: Figure 2 As shown, it mainly includes a doubly-fed wind turbine set and its control system. The wind turbine outlet is connected to the power grid through transformers ZT1, ZT2 and transmission lines ZL1, ZL2.

[0039] The overall control link of the doubly fed wind turbine is mainly divided into five parts, namely: MPPT control, converter control, Chopper protection control, Crowbar protection control, and pitch control. The converter control part is further divided into machine-side converter control and grid-side converter control, both of which can be further subdivided into outer loop control and current inner loop control. The outer loop control can input different variables according to the different control targets of the converter.

[0040] like Figure 3 As shown in the overall structure tree diagram of the doubly fed wind turbine, the doubly fed wind turbine converter control can be divided into machine-side converter control and grid-side converter control according to the control position; it can be divided into positive-sequence current control and negative-sequence current control according to the control sequence component. The machine-side converter of the doubly-fed wind turbine mainly performs power control. Through decoupling, active power and reactive power can be controlled separately, and most of the high and low voltage ride-through controls are set on the machine-side converter. By switching the control mode and control structure, the overall external output characteristics of the doubly-fed wind turbine are changed to achieve different fault ride-through goals. It is worth noting that due to the requirements of NB / T 31053-2021 "Verification Procedure for Electrical Simulation Models of Wind Turbines", the deviation value calculation is for positive-sequence electrical quantities, so the power control of the machine-side converter is based on the control of the positive-sequence component, that is, the rotor current, active power and reactive power involved are all obtained from the positive-sequence component, which belongs to positive-sequence current control. The negative-sequence current control part currently lacks rigid regulatory constraints, and some wind turbine manufacturers do not set it. Therefore, the relevant content of negative-sequence current control is analyzed and explained at the end, and the following description of control is based on positive-sequence control.

[0041] According to the control method, the machine-side converter can be divided into stator voltage-oriented control and stator flux-oriented control. The difference between the two control methods mainly lies in the different control objects of the dq-axis control and the different decoupling compensation items. The actual control effects are not much different.

[0042] The phase-locked loop monitors the three-phase stator voltage and outputs the grid voltage phase angle, which is then subtracted by 90° to obtain the stator flux angle. The d-axis is aligned with the stator flux, and the doubly-fed wind turbine rotor current is positioned in this coordinate system. The d-axis control object is the doubly-fed wind turbine reactive power, and the q-axis control object is the doubly-fed wind turbine active power.

[0043] When the doubly fed wind turbine is working in steady state, it operates in the mode of active power following active power reference value. When receiving the low voltage ride-through signal, it switches to the low voltage ride-through strategy: the positive sequence voltage is fitted with a coefficient through a specific function and multiplied by the active power signal as the active power reference value; the active power outer loop control is used to generate the d-axis current reference value for the machine-side converter control. Figure 4 The active outer loop control on the machine side is formed by the PI link. The reference values ​​of active and reactive power can be obtained by using the curve fitting element of ADPSS according to the reactive and active power emitted by the package model under different faults under different working conditions, or by a specific power-voltage relationship. The advantage of this outer loop control is that the reference value setting is intuitive and simple, and there is no need for complex substitution calculations from stator electrical quantities to rotor electrical quantities. The disadvantage is that the oscillation and overshoot of active and reactive power during the transient period are large, which is specifically reflected in the deviation comparison with the package model. The active power and active current deviation values ​​are large, and due to the long oscillation time, the active current enters the steady state during the fault period later than the package model, resulting in a large maximum deviation value in the steady state interval.

[0044] The control parameters under this control mode are shown in Table 1.1 Table 1.1 Dual PI link outer loop active control parameters

[0045] Under the active outer loop control structure of the dual PI link, there are four control link parameters that need to be set by the user. A PI control link of the active outer loop and the PI link of the current inner loop constitute a dual PI control structure. This control structure has high requirements for the setting of the PI link control parameters. There are many ways to set the fault active reference value. According to the expected active waveform output, the active power function can be constructed as the reference value input; or the multi-segment element in the ADPSS can be directly used to form the active function.

[0046] Based on the above analysis, other outer loop control methods can be used. Among the currently more commonly used high and low wear control methods, there is also an outer loop control method that directly gives the rotor current reference value. Compared with the previous control method, directly giving the rotor current reference value can reduce the PI link and simplify the control logic. The specific control structure is as follows: Figure 5 As shown. Figure 5It can be seen that although the PI link is reduced, its calculation process is relatively complex and not intuitive. The reference value of the rotor current can be compared with the package model and fitted using the ADPSS function fitting element, or it can be represented by the identification expression of constant power control or constant current control, and then obtained by mathematical conversion of the stator and rotor currents. The forms of the two expressions will be described in detail in the subsequent parameter identification link. Due to the introduction of the expressions of constant power control and constant current control, the control amount that needs to be set increases, and for doubly fed wind turbines with different circuit parameters, the expressions need to be modified before use.

[0047] The control parameters that need to be configured in this control mode are shown in Table 1.2 below: Table 1.2 Single PI link outer loop active control parameter table

[0048] The reactive power outer loop is equipped with low voltage ride-through control and high voltage ride-through control, and can adjust the control strategy after receiving low voltage ride-through and high voltage ride-through signals. The low voltage ride-through and high voltage ride-through of the reactive power outer loop are both the positive sequence voltage fitting a reactive reference value through a certain function, so that the reactive power generated by the doubly fed wind turbine follows this reference value, and provides the corresponding q-axis current reference value for the machine-side converter control.

[0049] like Figure 6 The setting of high and low reactive reference values ​​under the reactive power outer loop control of the dual PI link is very intuitive, and the desired reactive output can be achieved directly through the function. The control parameters under this control mode are shown in Table 1.3 below: Table 1.3 Dual PI link outer loop reactive power control parameter table

[0050] Under the reactive outer loop control structure of dual PI links, there are four control link parameters that need to be set by the user. A PI control link of the reactive outer loop and a PI link of the current inner loop constitute a dual PI control structure. This control structure has high requirements for the setting of PI link control parameters.

[0051] Step 2: High and low voltage ride-through threshold analysis and identification For the identification of high and low voltage ride-through parameters of doubly fed wind turbines, the high and low voltage ride-through control thresholds can determine when the doubly fed wind turbine switches to high and low voltage ride-through control. The subsequent division of the various parts of the output characteristic curve of the doubly fed wind turbine also requires a determined threshold to provide a basis for division. Therefore, the first step in the identification of the high and low voltage ride-through parameters of the doubly fed wind turbine is to identify the voltage values ​​at which the doubly fed wind turbine enters and exits the high and low voltage ride-through.

[0052] Based on the simulation data of Sungrow's double-fed wind turbine, the following criteria are summarized to determine whether high and low voltage ride-through has occurred: 1. When entering high and low voltage ride-through, the terminal voltage of the doubly-fed wind turbine will change.

[0053] 2. When entering high and low voltage ride through, the reactive power and reactive current output by the doubly fed wind turbine will change.

[0054] 3. When entering high and low voltage ride through, the active power and active current output by the doubly fed wind turbine will change.

[0055] 4. When entering high and low voltage ride-through, the DC bus voltage on the rotor side of the doubly-fed wind turbine will change.

[0056] 5. When entering high and low voltage ride through, the stator output current of the doubly fed wind turbine changes, which may generate a large fault current.

[0057] Integrating the above phenomena, each basis is analyzed one by one. For basis 1, for different types of doubly fed wind turbines, when the fault location, fault type, external circuit and fault ride-through strategy adopted are different, the terminal voltage of the doubly fed wind turbine will be quite different, which is difficult to be used as a direct criterion for judging whether it has entered high or low ride-through. However, the voltage entering high or low ride-through can be used to locate the timing of the doubly fed wind turbine to enter high or low ride-through control, as the high or low ride-through threshold of the doubly fed wind turbine; for basis 3, the active power of the doubly fed wind turbine will change during high or low ride-through, and will not exceed its rated value. However, according to the model simulation results of Sungrow Power Supply, the active power control strategy of the doubly fed wind turbine during fault ride-through may be different under different operating conditions. For example, the active power control measurement of the doubly fed wind turbine under low power conditions is the output value before the fault, while under high power conditions, the control target value changes with the voltage change, and it is difficult to have a unified basis; for basis 4, the DC voltage will fluctuate during high and low voltage ride-through, but it is impossible to distinguish whether it is the low ride-through threshold or the high ride-through threshold.

[0058] For basis 5, the stator current will indeed change significantly when high and low voltages are crossed, but it is difficult to determine a standard value as the threshold of high and low voltages. In contrast, according to basis 2, after high and low voltages are crossed, the reactive power and reactive current will change significantly, and the high and low voltages can be distinguished based on the positive and negative values ​​of the output reactive power. At the same time, since the current national standards have made clear requirements on the reactive support of wind turbines to the power grid, the high and low voltage crossing thresholds of the wind turbine can be easily determined based on the change in reactive power.

[0059] If the threshold for entering high and low penetration is determined based on basis 2, the basis for judgment needs to be determined. The reactive power of the doubly fed wind turbine is related to the terminal voltage. Under normal circumstances, the reactive power output of the doubly fed wind turbine is kept at 0 in steady state. Since the wind turbine is controlled by AVC, its steady-state reactive power is not necessarily 0. In order to maintain the accuracy of the threshold judgment, a detailed relationship between reactive power and voltage is required to determine the high and low penetration thresholds. According to GB / T 19963.1-2021 "Technical Regulations for Wind Farm Access to Power System Part 1: Onshore Wind Power", the dynamic reactive current increment of the wind farm should respond to the voltage change at the grid connection point and meet the following requirements: (2.1) (2.2) The formula specifies the relationship between the reactive current increase and the voltage. All wind turbines connected to the grid must satisfy the above expression during a fault. Therefore, the doubly-fed wind turbines used for parameter identification by default can satisfy the above formula. At the same time, it is specified that K1 and K2 must be greater than or equal to 1.5.

[0060] It is possible to calculate whether the reactive current and voltage at each point after the fault occurs meet the requirements of GB / T 19963.1-2021. If they do, it is considered that the doubly fed wind turbine enters fault ride-through control. The extraction of the fault point can also be done by subtracting the real-time reactive current from the steady-state value. If the deviation exceeds a certain value, it is considered that this moment is the moment of fault occurrence, that is, (2.3) (2.4) In the formula I qt is the reactive current per unit value at the current moment, I q0 is the reactive current per unit value in steady-state operation before the fault, I q_lvrt The expression in claim 5 , I q_hvrt The expression in Article 5 When the reactive current difference satisfies the above two equations, it is considered that the voltage at the terminal of the doubly fed wind turbine drops or rises, and the threshold judgment process of high and low crossing is entered as follows: Figure 7 .

[0061] For the high and low penetration exit thresholds of the doubly fed wind turbine, the high and low penetration entry thresholds cannot be simply used as the exit thresholds. The high and low penetration exit thresholds of the doubly fed wind turbine usually adopt a hysteresis link, and exit after the voltage meets a certain value, but this value is not necessarily the entry value. However, this value can also be obtained by a method similar to that of obtaining the entry threshold.

[0062] When the high and low penetration control is exited, the reactive current of the doubly fed wind turbine has a relatively obvious cut-off point. The high and low penetration exit threshold judgment process of the doubly fed wind turbine is as follows Figure 8 .

[0063] By analyzing and identifying the high and low voltage ride-through thresholds in step 2, the initial voltage values ​​of the high and low voltage ride-through control can be determined, which serve as the initial values ​​of the voltage variables in the subsequent identification of the ride-through control strategy.

[0064] Step 3: Identify high and low voltage ride-through strategy parameters a. Active power control method: (1) No additional control (2) Specified power (3.1) In the formula, P lvrt is the low wear active power reference value, K P_lvrt is the low wear active power slope coefficient, P 0 is the per-unit value of active power during the steady state before the fault, P set_lvrt A basic value is set for the low-penetration active power. This value can be set according to actual conditions, and the low-penetration active power is always above this value.

[0065] (3) Specified current (3.2) In the formula I plvrt is the low wear active current reference value, K 1 and K 2 are the two coefficients of low-pass active current, which can be assigned as needed. U t is the steady-state value of the fault voltage, I p0 is the steady-state value of active current before the fault, I set_lvrt A basic value is set for the low-throughput active current. This value can be set according to actual conditions, and the low-throughput active current is always above this value.

[0066] ( The output active power is divided by the terminal voltage. The response under multiple faults is used to identify K1, K2, and I set_lvrt , which is equivalent to solving a multivariate linear equation, or establishing a horizontal axis with voltage and a vertical axis with active power, measuring a scatter plot, and fitting with the least squares method) (4) Current before entering the crossing (5) Low voltage limit active current b. Reactive power control method: (1) No additional control (2) Specify the power (according to the national standard, this method cannot be used) (3) Specified current (3.3) In the formula K 3 and K 4 are the two coefficients of low-through reactive current, U lim is the fault ride-through voltage threshold, U t is the steady-state value of the fault voltage, I q0 is the steady-state reactive current value before the fault, I qset_lvrt Sets the basic value for the reactive current.

[0067] Control strategy 3.3 is mainly based on national standards: (3.4) In the formula K is the fault ride-through coefficient, I n is the rated current of the wind turbine. Control strategy 3.3 is the most commonly used reactive power control strategy for fault ride-through. After simulation of various models, it is found that even if the operation mode, fault type and fault depth are different, the reactive power control strategy can be expressed by strategy 3.3.

[0068] The identification process of active reference value can be used as follows Fig. 9 As shown. Secondly, regarding the identification of reactive reference values, my country has currently made detailed provisions for the reactive current of onshore wind power during fault ride-through. Compared with the various identification situations of active current, the identification of reactive reference values ​​is more single. According to GB / T 19963.1-2021, the reactive current emitted by my country's onshore wind power system during fault ride-through should satisfy equations (3.3) and (3.4), and the value of K1 should be greater than 1.5 and should not be greater than 3, and the value of K2 should be greater than 1.5. Therefore, the determination of reactive reference values ​​can be unified as the identification of reactive current reference values, and the expression of reactive current reference values ​​is assumed to be 1.

[0069] The identification of reactive current reference value is similar to that of active part, and the least square method can also be used for identification. The identification mathematical model of reactive reference value is very clear, and it is 1-time. The fault type can be changed and multiple sets of corresponding relationships can be measured. Normally, under one type of fault, when the doubly fed wind turbine reaches the steady state after the fault, the high and low voltage ride-through control of the doubly fed wind turbine will no longer continue to change and enter the temporary steady-state control. The voltage-reactive current relationship at this time can be extracted as the input-output relationship. During simulation, a longer fault period can be used, and then a longer time period from the low-voltage ride-through input to the end of the fault can be intercepted. The voltage and active power during the period are averaged as the input-output relationship of a fault, and then multiple sets of data are measured in the same way. The identification process of reactive reference value is as follows: Fig.10 shown.

[0070] According to the parameter identification ideas described above, the parameter identification program is designed. The first step is to identify the high and low voltage ride-through thresholds. The high and low voltage ride-through thresholds are determined by reactive characteristics. After the fault occurs, the doubly fed wind turbine has a transient process, during which the reactive power and reactive current do not completely conform to the reactive current control formula. The control has a response process. The identification is not accurate if it is completely based on the reactive current increase formula required by the national standard. The characteristics of the doubly fed wind turbine model that the high and low voltage ride-through control can be disabled can be used to measure the reactive values ​​when the high and low voltage ride-through is started and the reactive values ​​when the high and low voltage ride-through is not started, and compare them in the program. When a fault occurs and the voltage exceeds the threshold, the high and low ride-through control is put into operation. At this time, the difference in reactive values ​​between the two gradually increases. A reactive judgment difference PD can be set. When the difference between the two is greater than PD, it is judged that the doubly fed wind turbine enters the high and low ride-through control. There are certain differences in the reactive response characteristics under different fault depths. Multiple fault depths can be selected, and the reactive values ​​when high and low wear-through are started and not started can be measured multiple times. After obtaining multiple groups of thresholds, the average value can be obtained to reduce the error.

[0071] After obtaining the high and low voltage ride-through thresholds, the next step is to identify the high and low voltage ride-through control methods of the doubly fed wind turbine. The steady-state segment during the fault of the doubly fed wind turbine can be intercepted to obtain the average values ​​of active power / current and reactive power / current, and record the voltage value corresponding to the steady-state segment. Then, under multiple fault depths, record the active power / current, reactive power / current, and voltage, and use the least squares method to fit the active and reactive results under multiple groups of different voltages according to the high and low voltage ride-through strategies described above, and identify the various parameters in the active power / current and reactive / current expressions.

[0072] Step 4: Identification of high and low voltage ride-through control parameters a. Difference equation of single input, single output and single PI link In order to facilitate the identification of the subsequent double-fed wind turbine control parameters, the overall double-fed wind turbine control structure can be split into multiple basic units for analysis. First, starting from the simplest PI link, the control parameters and differential equations of the PI link are analyzed to facilitate the parameter identification and differential equation writing of the subsequent complex control structure.

[0073] The input quantity of the PI link in the doubly fed wind turbine is actually the reference value of the physical quantity x and the physical quantity x. If offline parameter identification is used, the reference value of x and the value of x at each moment are known, and the difference between the two is input into the PI control as the input quantity, thus forming a single-input and single-output PI link, such as Fig.11 .

[0074] Next, we derive the differential equation of a single PI link with a single input and a single output. First, we write the frequency domain expression of the control according to the control structure: (4.1) In the formula x For the control link input, y For the control link output, K p1 , K i1 They are the proportional coefficient and integral coefficient of the PI link respectively, x ref is the input reference value. Write equation 4.1 in the form of a standard difference equation (4.2) (4.3) In the formula y i is the ith output, a 1. b 0. b 1 is the three coefficients, is the sampling step. The PI link is derived into the standard difference equation form of formula (4.2), and the least squares parameter identification process can be used to calculate b0 and b1, and then solve the equation group to obtain the PI control parameter K p , K i .

[0075] b. Difference equation of double PI link Based on the derivation of the control structure of a single PI link, an input z is added between the two PI controls to form the most common dual-loop control in the double-fed wind turbine control. The reference value of x and the value of x at each moment are known, and the difference between the two is input into the PI control as an input quantity. z is also an input quantity, so this control structure is a multi-input control system. Fig.12 shown.

[0076] Next, we derive the differential equation of the dual-loop PI control with multiple inputs and a single output. First, we write the frequency domain expression of the control according to the control structure: (4.4) In the formula x It is the input of the outer loop control link. y For the control link output, K p1 , K i1 The proportional coefficient and integral coefficient of the outer loop PI link respectively, K p2 , K i2 They are the proportional coefficient and integral coefficient of the inner loop PI link, x ref To input the reference value, z It is the input of the inner loop control link.

[0077] The parameter identification is performed offline, and the discretization and term transfer can be obtained. (4.5) (4.6) In the formula a1 , a2 , b1 , b2 , b3 , b4 , b5 , b6 are coefficients, which can be expressed by control link parameters and constants. Δt is the sampling step length.

[0078] To solve for each coefficient, write it in matrix form: (4.7) In the formula, y i represents the i-th output, u i Represents the i-th input.

[0079] According to the above derivation, a set of parameter estimates can be obtained by using the least squares method. Due to the large amount of data, a new set of estimates can be continuously obtained.

[0080] Embodiment 2: According to the specific implementation method and the parameter identification ideas described in Example 1, a parameter identification program is designed. The first thing to do is to identify the high and low voltage crossing thresholds. The high and low voltage crossing thresholds are determined by reactive characteristics. Since there is a transient process in the doubly fed wind turbine after the fault occurs, the reactive power and reactive current do not completely conform to the reactive current control formula during the period. The control has a response process. The identification is completely in accordance with the reactive current increase formula required by the national standard, and the identification result is not accurate. In the doubly fed wind turbine model, the characteristics of not enabling high and low voltage crossing control can be used to measure the reactive value when the high and low voltage crossing is started and the reactive value when the high and low crossing is not started, and compare them in the program. When a fault occurs and the voltage exceeds the threshold, the high and low crossing control is put into operation. At this time, the difference in reactive values ​​between the two gradually increases. A reactive judgment difference PD can be set. When the difference between the two is greater than PD, it is judged that the doubly fed wind turbine enters high and low crossing control. There are certain differences in the reactive response characteristics under different fault depths. Multiple fault depths can be selected, and the reactive values ​​when high and low wear-through are started and not started can be measured multiple times. After obtaining multiple groups of thresholds, the average value can be obtained to reduce the error.

[0081] After obtaining the high and low voltage ride-through thresholds, the next step is to identify the high and low voltage ride-through control methods of the doubly fed wind turbine. The steady-state segment during the fault of the doubly fed wind turbine can be intercepted to obtain the average values ​​of active power / current and reactive power / current, and record the voltage value corresponding to the steady-state segment. Then, under various fault depths, the active power / current, reactive power / current, and voltage are recorded. The least squares method is used to fit the active and reactive power under multiple groups of different voltages according to the high and low voltage ride-through strategies described above, and the various parameters in the active power / current and reactive power / current expressions are identified.

[0082] According to the above principle, a doubly-fed asynchronous wind turbine model is built in ADPSS. The parameter identification process designed above is used to identify the high and low voltage ride-through control parameters of the electromagnetic transient packaging models of two models of Sungrow Power Supply, namely: WG2500KDF and WG3600KDF.

[0083] First, the high and low voltage ride-through strategy parameters of the WG2500KDF package model are identified. The identification results are as follows: Fig.13 shown.

[0084] Active power identification, such as Fig.13 (a) KP_LVRT=0.066995Pset_LV=0.074206; Active current identification, such as Fig.13 (b) K1_IP_LV=-0.114550K2_IP_LV=0.159070IPset_LV=0.173036; Reactive power identification, such as Fig.13 (c) KQ_LV=0.480987Qset_LV=0.337212; Reactive current identification, such as Fig.13 (d) K1_IQ_LV=2.620926K2_IQ_LV=0.0IQset_LV=0.136088; Recovery slope fit K=0.189535; The upper limit of the parameter is lower than that: the maximum active power value is not detected; Qmax=0.127315; IPmax=0.261346; IQmax=1.250659.

[0085] like Fig.14 , 2500 model high wear parameter identification: Active power identification, such as Fig.14 (a) KP_LVRT=0.000000Pset_LV=0.896361; Active current identification, such as Fig.14 (b) K1_IP_LV=-0.778378K2_IP_LV=0.833976IPset_LV=0.907194; Reactive power identification, such as Fig.14 (c) KQ_LV=1.616853Qset_LV=-0.278378; Reactive current identification, such as Fig.14 (d) K1_IQ_LV=1.166660K2_IQ_LV=0.0IQset_LV=-0.263170.

[0086] Then the high and low voltage ride-through strategy parameters of the WG3600KDF package model are identified. The identification results are as follows: Fig.15 shown.

[0087] Active power identification, such as Fig.15 (a) KP_LVRT=0.082682Pset_LV=0.091353; Active current identification, such as Fig.15 (b) K1_IP_LV=-0.114335K2_IP_LV=0.168896IPset_LV=0.182681; Reactive power identification, such as Fig.15 (c) KQ_LV=0.339989Qset_LV=0.472057; Reactive current identification, such as Fig.15 (d) K1_IQ_LV=2.570136K2_IQ_LV=0.0IQset_LV=0.325411; Recovery slope fit K=0.712469; The upper limit of the low-pass parameter is: Pmax=0.175896; the maximum reactive power value is not detected; IPmax=0.294000; IQmax=1.378345.

[0088] like Fig.16 , 3600 model high wear parameter identification: Active power identification, such as Fig.16 (a) KP_LVRT=0.000000Pset_LV=0.914192; Active current identification, such as Fig.16 (b) K1_IP_LV=-0.671870K2_IP_LV=0.781900IPset_LV=0.845718; Reactive power identification, such as Fig.16 (c) KQ_LV=3.010329Qset_LV=-0.131622; Reactive current identification, such as Fig.16 (d) K1_IQ_LV=2.621456K2_IQ_LV=0.0IQset_LV=-0.103514.

[0089] Next, the control parameters are identified. Based on the above derivation of the PI link parameter identification, the above parameter identification process can be realized through the MATLAB program. By applying a given step signal to the control link and analyzing the relationship between the input and output, the control parameters to be identified can be obtained. However, for the electromagnetic transient packaging model of the doubly fed wind turbine, the data that can be observed externally is limited, and the control input and output and control structure that can be adjusted by the user are also limited. Therefore, it is difficult to obtain the ideal input and output signals during parameter identification. It may be necessary to obtain the corresponding signals indirectly, including converting and calculating other signals, replacing step signals at other positions, etc. However, when there are many signals that can be adjusted by the user, the most ideal parameter identification method can be achieved by appropriately adjusting the input method of the step signal, that is, the PI input signal is a step signal, and the output signal can be obtained. Therefore, the signal selection and parameter identification algorithm of parameter identification are equally important.

[0090] The method for identifying the parameters of a doubly-fed wind turbine is to apply a step signal to the input port of the control structure to be identified, obtain its output signal, and then calculate the approximate value of the control parameter through the difference equation and the least squares method. At present, all control links of mainstream doubly-fed wind turbines adopt PI control, and the control link parameters that need to be identified are the control parameters K of the PI link. p , K i Therefore, a step signal can be applied before the PI link, and then the output can be obtained for parameter identification.

[0091] This identification method can accurately identify the integral link parameter K in the PI link i , there is almost no error in the identification process. First, write the parameter identification program of the PI link. Based on the least squares identification method and the identification structure of applying a step signal to the PI link, the corresponding program can be written. The results of each round of the least squares method are calculated by 16 sets of input and output.

[0092] Normally, the PI link can use the built-in PI link components of ADPSS, but there are also cases where PI control is built by oneself. To ensure that both PI links can be accurately identified in ADPSS, parameter identification tests are performed on the two PI links below.

[0093] First, test the PI link components in the ADPSS. Test the reactive outer loop control of the grid-side converter in the control link. The setting value of the PI link in this link is: K p =1, T i =0.01(Ki=100). The step signal occurs at 6s, from 0 to 0.5. K obtained by program identification i Parameters such as Fig.17 shown.

[0094] By observation Fig.17 It can be found that K i The error of parameter identification result is small, the final result fluctuates around 100, and its fluctuation error can be ignored. i The identification of parameters is relatively accurate.

[0095] By deriving the differential equation of the PI link in ADPSS, since the equation contains K p (x i- x i-1 ), the input at each moment of the steady state after the step occurs is equal, so it contains K p The coefficient of the term is 0, and the parameter K p Unable to identify. There are two solutions.

[0096] Method 1: Use the data before part of the step signal.

[0097] This method can directly obtain K p However, since 16 inputs and outputs are intercepted for least squares calculation, the parameter values ​​will change irregularly before and after the step signal. p The identification results are as follows Fig.18 As shown, observe Fig.18 It can be found that after 6s, that is, in the period after the step occurs, K p The identification result is always 0. Before 6 seconds, K p The identification result is about 0.997. During the steady-state operation of the fan, K p The identification results fluctuate around 1.

[0098] Method 2: Calculate based on the instantaneous input and output of the step signal.

[0099] Since the PI link consists of a proportional link and an integral link, only the proportional link is effective, so at the moment the step signal occurs: (5.1) Fig.19 To capture the input and output of the PI link, by analyzing Fig.19 It can be found that before the step, the input and output of the PI link are both 0. Immediately after the step occurs, the input becomes 0.5 and the output becomes 0.5037. Substituting into formula (1), we get K p About 1.0074.

[0100] Then, the PI link built in ADPSS is tested. The control link used in the test is the d-axis current inner loop control of the machine-side converter of the white box model. The setting value of the PI link of this link is: K p =0.5, T i=10(K i =0.1). This control link is as follows Fig.19 As shown, the step signal occurs in 6s, from 0.11 to 0.5.

[0101] K obtained by program identification i Parameters such as Fig. 20 shown.

[0102] By observation Fig. 20 It can be found that K i The error of parameter identification result is small, the final result fluctuates around 0.1, and its fluctuation error can be ignored. i The identification of parameters is relatively accurate.

[0103] Through the derivation of the differential equation of the PI link in ADPSS in the above method, it can be concluded that since the formula contains K p (x i -x i-1 ), the input at each moment of the steady state after the step occurs is equal, so it contains K p The coefficient of the term is 0, and the parameter K p Unable to identify. The above two processing methods can still be used.

[0104] Method 1: Use the data before part of the step signal.

[0105] This method can directly obtain K p However, since 16 inputs and outputs are intercepted for least squares calculation, the parameter values ​​will change irregularly before and after the step signal. p The identification results are as follows Fig.21 As shown, observe Fig.21 It can be found that after 6s, that is, in the period after the step occurs, K p The identification result is always 0. Before 6 seconds, K p The identification result is approximately 0.499.

[0106] Method 2: Calculate based on the instantaneous input and output of the step signal.

[0107] Since the PI link is composed of a proportional link and an integral link, and the function of the integral link is to find the integral within a certain time period, it takes time to respond, and the proportional link is only a simple multiplication calculation. The output can respond at the moment the input changes. At the moment the step occurs, it can be considered that the integral link is ineffective and only the proportional link is effective.

[0108] Fig. 22 To capture the input and output of the PI link, by analyzing Fig. 22It can be found that before the step, the input of the PI link is 0.0658 and the output is 0.11. Immediately after the step occurs, the input becomes 0.5 and the output becomes 0.3272. Substituting into formula (1), we get K p About 0.5.

[0109] Through the above analysis, it is found that for the two PI link forms in ADPSS, the identification method can have good accuracy. For the discrete algorithm of the PI controller in ADPSS, the parameter K p The algorithm needs to be considered separately, and the parameter identification procedure can be adjusted appropriately. The two methods described above can effectively identify the parameter K. p .

[0110] In summary, adding a step function to the PI link input has a good identification effect and high identification accuracy. However, for the package model, not all PI link inputs and outputs are directly visible and adjustable, so this method also has certain limitations, that is, for the package model, it is necessary to consider whether the PI link input and output are adjustable and observable.

[0111] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

Claims

1. A method for identifying high and low voltage ride-through control parameters of a doubly-fed wind turbine, characterized in that: The method comprises: Analyze high and low voltage ride-through thresholds based on simulation data output by the pre-built double-fed wind turbine control structured model; The high and low voltage ride-through thresholds are used as the voltage initial values ​​for high and low voltage ride-through control, the high and low voltage ride-through control modes of the doubly-fed wind turbine are identified, and high and low voltage ride-through strategies are formulated; Based on the high and low voltage ride-through strategies, various parameters in the control link are identified to obtain various control parameter values ​​in the control link.

2. The method according to claim 1, characterized in that The pre-built doubly-fed wind turbine control structured model includes: an MPPT control module, a converter control module, a Chopper protection control module, a Crowbar protection control module and a pitch control module; wherein the converter control module includes a machine-side converter control unit and a grid-side converter control unit; the machine-side converter control unit is used for power outer loop control and current inner loop control; the grid-side converter control unit is used for DC voltage outer loop control, reactive power outer loop control and current inner loop control.

3. The method according to claim 1, characterized in that The analysis of high and low voltage ride-through thresholds includes: By identifying the voltage values ​​at which the doubly fed wind turbine enters and exits the high and low voltage ride through when a fault occurs, the high and low voltage ride through entry threshold and exit threshold of the doubly fed wind turbine are determined.

4. The method according to claim 3, characterized in that The method for determining the high and low penetration entry threshold and exit threshold of the doubly fed wind turbine includes: Measure the steady-state reactive current value. When the steady-state reactive current value satisfies the condition that the dynamic reactive current increment of the wind farm should respond to the voltage change of the grid connection point, calculate the difference between the real-time reactive current and the reactive current before the fault, that is, the reactive current increment; when the reactive current increment exceeds the preset threshold of the reactive current for the doubly fed wind turbine to enter the fault ride-through control, the current moment is considered to be the moment of fault occurrence; according to NB / T 31053-2021 "Verification Procedure for Electrical Simulation Model of Wind Turbine Generators", measure the reactive current at each moment within 140ms after the fault occurs I q and terminal voltage U t , when the formula ( I q - I q0 ) / (0.9- U t )≥1.5 or ( I q0 - I q ) / ( U t -1.1)≥1.5, then the terminal voltage at the current moment U t It is defined as the low penetration threshold or high penetration threshold of the doubly fed wind turbine; I qt is the reactive current per unit value at the current moment, I q0 is the per unit value of the steady-state reactive current before the fault, U t It is the unit value of the terminal voltage.

5. The method according to claim 4, characterized in that The dynamic reactive current increment of the doubly-fed wind turbine generator set should respond to the voltage change condition at the grid connection point: (1.1) (1.2) Where Δ I t is the reactive current increment at the grid connection point. is the reactive current increment at the grid connection point. K 1 is the low-pass reactive current increment coefficient, I N is the rated current per unit value, K 2 is the high-voltage reactive current increment coefficient, U t is the steady-state value of the fault voltage.

6. The method according to claim 5, characterized in that The difference between the real-time reactive current and the steady-state value is calculated by the following formula: (1.3) (1.4) In the formula I qt is the reactive current per unit value at the current moment, I q0 is the reactive current per unit value in steady-state operation before the fault, I q_lvrt The expression in claim 5 , I q_hvrt The expression in claim 5 .

7. The method according to claim 1, characterized in that The method of using the high and low voltage ride-through thresholds as the voltage initial values ​​of the high and low voltage ride-through control to identify the high and low voltage ride-through control methods of the doubly fed wind turbine includes: selecting a time point after the voltage initial values ​​of the high and low voltage ride-through control, intercepting the steady-state segment during the fault of the doubly fed wind turbine, obtaining the average values ​​of the active power / current and the reactive power / current, and recording the voltage value corresponding to the steady-state segment, and then recording the active power / current, reactive power / current, and voltage under various fault depths, using the least squares method to perform function fitting on multiple groups of active and reactive results under different voltages according to the high and low voltage ride-through strategies, and identifying each parameter in the active power / current and reactive power / current expressions.

8. The method according to claim 1, characterized in that The high and low voltage ride-through strategy includes identification of active current reference value and reactive current reference value; wherein, identification of active current reference value includes: selecting reference value basis, when specifying power as reference value, performing an identification once by the following formula: (1.5) In the formula, P lvrt is the low wear active power reference value, K P_lvrt is the low wear active power slope coefficient, P 0 is the per-unit value of active power during steady state, P set_lvrt Set a basic value for the low-pass active power, and the low-pass active power is always above this value; When the current is specified as the reference value, an identification is performed using the following formula: (1.6) In the formula I plvrt is the low wear active current reference value, K 1 is the low wear active current coefficient 1, K 2 is the low wear active current coefficient 2, U t is the steady-state value of the fault voltage, I p0 is the steady-state value of active current before the fault, I set_lvrt Set a basic value for the low-throughput active current, and the low-throughput active current is always above this value; The secondary identification is performed by the following formula: (1.7) In the formula A 1, A 2, A 3 are all secondary low-through active current coefficients; The identification of the reactive current reference value is performed once using the following predefined reactive current identification model expression: (1.8) In the formula K 3 and K 4 are the two coefficients of low-through reactive current, U lim is the fault ride-through voltage threshold, U t is the steady-state value of the fault voltage, I q0 is the steady-state reactive current value before the fault, I qset_lvrt Sets the basic value for the reactive current.

9. The method according to claim 5, characterized in that Based on the high and low voltage ride-through strategy, the various parameters in the control link are identified to obtain the values ​​of various control parameters in the control link, including: Difference equation for a single PI link: (1.9) In the formula x For the control link input, y For the control link output, K p1 , K i1 They are the proportional coefficient and integral coefficient of the PI link respectively. x ref To input reference value; Using the least squares method, b0 and b1 are calculated by the following formula, and the PI control parameter K is obtained by solving. p1 , K i1 (1.10) (1.11) In the formula y i is the ith output, a 1. b 0. b 1 is the three coefficients, Δt is the sampling step length; The difference equation of the double PI link: (1.12) In the formula x It is the input of the outer loop control link. y For the control link output, K p1 , K i1 The proportional coefficient and integral coefficient of the outer loop PI link respectively, K p2 , K i2 They are the proportional coefficient and integral coefficient of the inner loop PI link, x ref To input the reference value, z is the input of the inner loop control link; the parameter identification is performed offline, and the discretization and term shifting can be obtained: (1.13) (1.14) In the formula a1 , a2 , b1 , b2 , b3 , b4 , b5 , b6 are coefficients, which can be expressed by control link parameters and constants. Δt is the sampling step length; Use the least squares method to solve the coefficients of the following matrix: (1.15) In the formula, y i represents the i-th output, u i Represents the i-th input.

10. A high and low voltage ride through control parameter identification system for a doubly fed wind turbine based on a structured model, characterized in that: The system comprises: An analysis module, used to analyze high and low voltage ride-through thresholds based on simulation data output by a pre-built double-fed wind turbine control structured model; An identification module is used to use the high and low voltage ride-through thresholds as voltage initial values ​​for high and low voltage ride-through control, identify high and low voltage ride-through control modes for the doubly-fed wind turbine, and formulate high and low voltage ride-through strategies; The acquisition module is used to identify various parameters in the control link based on the high and low voltage ride-through strategy and obtain various control parameter values ​​in the control link.

Images