Current compensation type three-vector model predictive control method for off-grid doubly-fed wind turbine generator

Abstract

The invention discloses a current compensation type three-vector model prediction control method for an off-grid doubly-fed wind turbine generator, belongs to the technical field of doubly-fed wind turbine generator control, and aims to solve the problems of voltage sag and high total harmonic distortion rate of a micro-grid system caused by sudden load change during off-grid operation of the doubly-fed wind turbine generator. The control method comprises the steps that on the basis of virtual synchronous networking control, an off-grid doubly-fed wind turbine generator rotor current prediction model is established, load current, stator side converter current and stator voltage are introduced into three-vector model prediction control, and the voltage coupling effect and current change caused by load abrupt change are compensated in real time, so that the rotor current of the off-grid doubly-fed wind turbine generator is predicted. Obtaining a compensated rotor current reference; a cost function is designed by taking the minimum prediction error of the rotor current as a target, so that accurate prediction of the rotor current during sudden load change is realized, a voltage sag phenomenon caused by sudden load change is effectively inhibited, the harmonic content of output voltage is reduced, and meanwhile, quick response to system frequency and load power is realized.

Description

Technical Field

[0001] This invention relates to a current-compensated three-vector model predictive control method for off-grid doubly-fed wind turbines, belonging to the field of doubly-fed wind turbine control technology. Background Technology

[0002] With the increasing focus on new energy power generation, microgrids and distributed energy systems have developed rapidly. Off-grid wind power systems do not require the support of a large power grid and can operate independently to supply power to loads. They are suitable for microgrids and remote areas that are difficult to connect to the grid, such as mountainous areas, islands, and ships, offering high flexibility. Off-grid wind power systems have become an important component of microgrid systems, playing a significant role in alleviating power supply shortages and compensating for insufficient grid-connected power generation. Among them, doubly-fed induction generator (DFIG) wind turbines have developed rapidly due to their advantages such as low cost and high reliability, and have become the mainstream form of off-grid wind power generation at present.

[0003] Currently, off-grid control methods for doubly-fed induction generators (DFIGs) typically employ vector control techniques based on stator magnetic field forced orientation and proportional-integral (PI) control. Depending on the controlled variable, these methods can be categorized into indirect flux linkage control and direct voltage control. Indirect flux linkage control uses the excitation current as the controlled variable, thereby controlling the stator voltage. Direct voltage control establishes a mathematical relationship between the stator voltage amplitude and the excitation current, using the stator voltage amplitude as the controlled variable, thus simplifying the control process. Building upon this, some studies have decoupled the stator voltage outer loop control, establishing a dual voltage-current closed loop to improve voltage control performance. However, all of the above control methods rely on stator magnetic field forced orientation, and the accuracy of flux linkage observation is easily affected by generator parameters, impacting the dynamic response performance of the DFIG. Simultaneously, PI control is susceptible to system parameters, making inner and outer loop parameter matching difficult and resulting in poor dynamic performance. Therefore, some studies have introduced model predictive control (MPC) into DFIG control to achieve better dynamic response. However, the traditional MPC algorithm is a single-vector control, which has low control accuracy, resulting in large pulsation of the controlled variable and high harmonic content. Furthermore, it has not been studied and applied in conjunction with off-grid doubly fed wind turbine units.

[0004] Furthermore, due to the lack of effective support from a large power grid, off-grid doubly-fed induction generators (DFIGs) lack sufficient inertia, resulting in poor power generation immunity. When load changes abruptly, the output voltage, frequency, and power of off-grid wind power systems experience significant fluctuations, threatening the safe and stable operation of the microgrid. Grid-based control can enable DFIGs to possess inertial response characteristics and independent power supply capabilities. A commonly used grid-based control method for DFIGs employs Virtual Synchronous Generation (VSG) control technology in the outer control loop and a dual closed-loop configuration of stator voltage and rotor current in the inner loop, enabling the DFIG to support system frequency and voltage. However, this control method is based on the premise of zero stator current, essentially a no-load voltage-building control; simultaneously, it treats load current, stator-side converter current, and voltage coupling terms as disturbances, reducing the robustness of voltage control. Moreover, DFIG grid-based control technologies have mostly been researched and applied under grid-connected or weakly grid-connected conditions, relying on the grid for voltage support and neglecting the voltage coupling effect caused by sudden load changes when off-grid. Therefore, off-grid doubly-fed induction generators require a control algorithm that meets the control requirements.

[0005] In summary, the following drawbacks exist in using traditional control methods (such as PI control) for off-grid doubly-fed wind turbines:

[0006] 1. Severe voltage sag during load changes: lack of dynamic compensation for voltage coupling effects;

[0007] 2. High harmonic distortion rate: Insufficient accuracy of single-vector model predictive control (MPC);

[0008] 3. Slow response speed: The PI controller parameters are sensitive, and matching of inner and outer loops is difficult.

[0009] While Virtual Synchronous (VSG) control can provide inertia support, it does not solve the voltage coupling problem caused by sudden changes in load current. Therefore, a high-precision and robust off-grid control method is urgently needed. Summary of the Invention

[0010] To address the issues of voltage sag and high total harmonic distortion (THD) in microgrid systems caused by sudden load changes during off-grid operation of doubly-fed induction generator (DFIG) wind turbines, this invention provides a current-compensated three-vector model predictive control method for off-grid DFIG wind turbines.

[0011] The off-grid doubly-fed induction generator current-compensated three-vector model predictive control method of the present invention includes the following steps:

[0012] Step 1: Establish a rotor current prediction model in the inner loop of the RSC control of the rotor-side converter. Based on the DFIG state equation, discretize it using the forward Euler method to obtain the rotor current prediction equation in the dq coordinate system. The rotor current prediction model is associated with the rotor voltage and the switching function. Voltage vector synthesis is achieved through the combination of switching states.

[0013] Step 2: Design a current compensation strategy by introducing load current i L Stator-side converter current i g and stator voltage u s Calculate the current coupling compensation amount and generate the reference value of the rotor current after compensation;

[0014] Step 3: Execute three-vector model predictive control:

[0015] Based on the mapping relationship between rotor voltage and switching function, the eight switching states of the rotor-side converter RSC are converted into basic voltage vectors in the dq coordinate system;

[0016] Within each sampling period, two adjacent non-zero vectors are selected and combined with the zero vector to synthesize the target voltage vector;

[0017] The cost function is designed with the goal of minimizing the rotor current tracking error, and the optimal switching signal is output to the rotor-side converter RSC.

[0018] Preferably, the rotor current prediction equation in step 1 is:

[0019]

[0020] In the formula, σ is the leakage flux coefficient. L m For the equivalent mutual inductance between the stator and rotor windings, L s For the equivalent stator winding self-inductance, L r This is the equivalent rotor winding self-inductance;

[0021] u sd ,u sq and u rd ,u rq These are the stator voltages u s and rotor voltage u r The d-axis and q-axis components;

[0022] i sd i sq and i rd i rq These are the stator currents i s and rotor current i r The d-axis and q-axis components;

[0023] k represents the current sampling time, and k+1 represents the next sampling time;

[0024] C is the coupling coefficient matrix of the stator current.

[0025] B is the coupling coefficient matrix of the rotor current.

[0026] R s and R r ω is the equivalent resistance of the stator and rotor windings; s For synchronous rotational angular velocity, ω m The rotor's electric angular velocity;

[0027] T s The sampling period.

[0028] Preferably, the mapping relationship between rotor voltage and switching function is as follows:

[0029]

[0030] Where, θ sl Let θ be the slip angle, and θ be the slip angle. sl =ω sl t, ω sl Let ω be the slip angular velocity. sl =ω s -ω m ;U dc RSC DC side voltage; S R S is the switching function vector of the three-phase bridge arm of the RSC bridge. R =[S a S b S c ] T S a ,S b ,S c This is the switching function for the three-phase bridge arm.

[0031] Preferably, the current compensation strategy in step 2 includes:

[0032] Step 21: Stator voltage d-axis orientation:

[0033]

[0034] Among them, u sm This refers to the stator voltage amplitude.

[0035] And by controlling the rotor current i rd and i rq To achieve control of the stator voltage;

[0036] Step 22: Calculate the current coupling compensation amount i rdc i rqc :

[0037]

[0038] In the formula, i rdc and i rqc These are the compensation values ​​for the d-axis and q-axis components of the rotor current, respectively; Cg The filter capacitor connected in parallel at the common connection point of the load, i gd i gq For the stator-side converter SSC AC side current i g d-axis and q-axis components; i Ld i Lq For the load current i L The d-axis and q-axis components;

[0039] Step 23: Generate the compensated rotor current reference value i rdref i rqref :

[0040]

[0041] Among them, K usP and K usI For the proportional gain and integral gain of the stator voltage control loop; u sdref and u sqref The stator voltage d-axis and q-axis component reference commands are generated by the virtual synchronous generator (VSG).

[0042] Preferably, in step 21, the rotor current i rd and i rq The calculation process includes:

[0043] Substituting the stator voltage d-axis orientation relation into the DFIG voltage equation and flux linkage equation, let R s =0, and the differential term is zero in steady state. Therefore, the relationship between stator current and rotor current in the dq coordinate system is as follows:

[0044]

[0045] Considering that a filter capacitor C is connected in parallel at the common connection point PCC of the load. g After coordinate transformation, the following relationship exists at PCC:

[0046]

[0047] Substituting the relationship between stator current and rotor current in the dq coordinate system into the above equation, we obtain the rotor current equation at PCC when the stator voltage is oriented as follows:

[0048]

[0049] When performing stator voltage control on an off-grid doubly-fed induction generator (DFIG), it is necessary to obtain i through calculations of the current equation at the PCC. rd and i rq As the RSC inner loop control input reference, it is necessary to control the rotor current i rd and irq This is to achieve control of the stator voltage.

[0050] Preferably, step 3, converting the switching state into a basic voltage vector, includes:

[0051] RSC adopts a two-level three-phase voltage source inverter topology, with a total of 8 switching state combinations;

[0052] Based on the mapping relationship between rotor voltage and switching function, the eight switching states of the rotor-side converter RSC are converted into basic voltage vectors in the dq coordinate system. These basic voltage vectors include six non-zero basic voltage vectors and two zero voltage vectors; the amplitude of the non-zero vectors is 2 / 3U. dc They are distributed in a hexagonal pattern in the dq plane, with adjacent vectors spaced 60° apart.

[0053] Preferably, step 3 synthesizes the target voltage vector u. tn The process includes:

[0054] Three fundamental vectors are used in each sampling period, namely two adjacent non-zero vectors u. i and u j And a zero vector u0, combined to form a synthetic target voltage vector u tn To track the reference vector, the target voltage vector is synthesized using the following formula:

[0055]

[0056] In the formula, t i t j t0 and t0 are the three basic voltage vectors u i 、u j The duration of action of u0 and u0 satisfy the following relationship:

[0057] t i +t j +t0=T s

[0058] The time allocation for each basic voltage vector is as follows:

[0059]

[0060] Among them, g i g j and g0 are u i 、u j The cost function value when u0 is interacting with it.

[0061] Preferably, the cost function in step 3 is:

[0062] g = |i rdref -i rd(k+1)|+|i rqref -i rq (k+1)|

[0063] The cost function value g when the synthesized voltage vector is applied tn Expressed as:

[0064] g tn =g i t i +g j t j +g0t0

[0065] By using a rolling optimization of the cost function, the synthesized voltage vector u corresponding to the minimum cost function value is selected. tn This is the optimal voltage vector, and the output control signal is applied to the rotor-side converter RSC to achieve current-compensated three-vector model predictive control.

[0066] The off-grid doubly-fed induction generator (DFIG) wind turbine current-compensated three-vector model predictive control method described in this invention improves upon the traditional DFIG off-grid control method. Based on virtual synchronous grid control, a rotor current prediction model for the off-grid DFIG wind turbine is established, incorporating load current, stator-side converter current, and stator voltage into the three-vector model predictive control. By real-time compensation for voltage coupling effects and current changes caused by load mutations, a compensated rotor current reference is obtained. A cost function is designed with the goal of minimizing the rotor current prediction error, achieving accurate rotor current prediction during load mutations. This effectively suppresses voltage sags caused by load mutations, reduces harmonic content in the output voltage, and accelerates the response speed to system frequency and load power.

[0067] In summary, the present invention has the following beneficial effects:

[0068] 1. A current compensation strategy was designed to obtain a reference rotor current after compensation by real-time compensation for voltage coupling effect and current change caused by load change, which effectively suppressed the voltage sag phenomenon caused by load change.

[0069] 2. A current-compensated three-vector model predictive control was designed, which enabled accurate prediction of rotor current, reduced the harmonic content of output voltage, and accelerated the response speed to system frequency and load power.

[0070] 3. The voltage outer loop adopts VSG control, which enables the off-grid doubly-fed wind turbine to have the ability to actively support voltage and frequency, and enhances the inertia of the off-grid doubly-fed wind turbine. Attached Figure Description

[0071] Figure 1 This is a block diagram of an off-grid doubly-fed wind power system.

[0072] Figure 2 This is the overall control block diagram of RSC;

[0073] Figure 3 Here is a block diagram of the VSG control principle;

[0074] Figure 4 Block diagram of current-compensated three-vector model predictive control algorithm Detailed Implementation

[0075] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0076] Specific implementation method one, combined with Figures 1 to 4 The current-compensated three-vector model predictive control method for off-grid doubly-fed wind turbines described in this embodiment includes the following steps:

[0077] Step 1: Establish a rotor current prediction model in the inner loop of the RSC control of the rotor-side converter. Based on the DFIG state equation, discretize it using the forward Euler method to obtain the rotor current prediction equation in the dq coordinate system. The rotor current prediction model is associated with the rotor voltage and the switching function. Voltage vector synthesis is achieved through the combination of switching states.

[0078] Step 2: Design a current compensation strategy by introducing load current i L Stator-side converter current i g and stator voltage u s Calculate the current coupling compensation amount and generate the reference value of the rotor current after compensation;

[0079] Step 3: Execute three-vector model predictive control:

[0080] Based on the mapping relationship between rotor voltage and switching function, the eight switching states of the rotor-side converter RSC are converted into basic voltage vectors in the dq coordinate system;

[0081] Within each sampling period, two adjacent non-zero vectors are selected and combined with the zero vector to synthesize the target voltage vector;

[0082] The cost function is designed with the goal of minimizing the rotor current tracking error, and the optimal switching signal is output to the rotor-side converter RSC.

[0083] The structure of an off-grid doubly-fed wind power generation system is as follows: Figure 1As shown. Due to the decoupling effect of the intermediate capacitor C1 in the back-to-back converter, the back-to-back converter can be divided into RSC and SSC for separate control. The control objective of SSC is the same in both grid-connected and off-grid operation: maintaining the DC voltage stability between the converters and providing the necessary reactive power. Therefore, the control strategy is basically the same. However, the control objective of RSC differs between grid-connected and off-grid operation. In off-grid operation, the amplitude and frequency of the stator voltage are controlled by controlling RSC. The function of the energy storage unit is to provide excitation energy to RSC during initial system startup, establishing excitation and DC voltage. This invention studies the control strategy of RSC during off-grid operation, assuming that the converter excitation has already been established.

[0084] Following the conventions for electric motors, a mathematical model of the DFIG in the dq synchronous rotating coordinate system is given. The DFIG voltage equation is as follows:

[0085]

[0086] Where p is the differential operator d / dt; ψ sd ψ sq and ψ rd ψ rq These are the d-axis and q-axis components of the stator flux linkage and rotor flux linkage, respectively.

[0087] The DFIG flux linkage equations are as follows:

[0088]

[0089] With current as the state variable, the DFIG state equations are obtained as follows:

[0090]

[0091]

[0092] Based on the mathematical model of DFIG, a current-compensated three-vector model predictive control strategy for off-grid doubly-fed induction generator (DFIG) wind turbines is proposed to control the RSC. VSG control is used as the voltage outer loop to generate system frequency and voltage amplitude references. After stator voltage closed-loop control, the reference values ​​i of the d-axis and q-axis components of the rotor current before compensation are obtained. rd * and i rq * The input is fed into a current-compensated three-vector model predictive control algorithm (CC-TVMPC algorithm) for rolling optimization and predictive control, ultimately obtaining the optimal control output S. R * The overall control block diagram is as follows: Figure 2 As shown.

[0093] Traditional DFIG off-grid control strategies only consider the steady-state properties of output voltage and frequency. Their output power passively changes with load variations, lacking the ability to actively adjust power and exhibiting poor robustness. To enable the DFIG to actively support voltage and frequency during off-grid operation, a VSG control strategy is adopted in the stator voltage outer loop of the RSC control.

[0094] The VSG outer loop control of the DFIG consists of two parts: active power-frequency control and reactive power-voltage control. Its principle block diagram is shown below. Figure 3 As shown. The active-frequency control equations are as follows:

[0095]

[0096] Among them, P m The mechanical power of the VSG, including the active power command P. ref The speed controller output and the speed regulator output are two parts, K p P is the active-frequency droop factor. e ω represents the actual output active power. V ω is the electrical angular frequency output by the VSG; ω0 is the grid synchronization angular frequency, and in the off-grid case, it is the microgrid system angular frequency; D is the damping coefficient; J is the moment of inertia.

[0097] The VSG utilizes the reactive power and voltage droop characteristics to control the output voltage amplitude. Its reactive power-voltage control equation is as follows:

[0098] U V =U0+K q (Q ref -Q e )

[0099] Among them, U V U0 is the output voltage amplitude of the VSG; K is the no-load voltage of the VSG. q Q is the reactive power-voltage droop factor. ref For reactive power command, Q e This represents the actual reactive power output.

[0100] In summary, the electric angular frequency ω generated by the VSG V and voltage amplitude U V ω, respectively, serves as the angular frequency reference command for DFIG in off-grid control. sref and stator voltage amplitude reference command u smref This enables the off-grid doubly fed wind turbine units to achieve active frequency regulation and voltage support.

[0101] The rotor current inner-loop control strategy commonly used in off-grid doubly-fed induction generator (DFIG) wind turbines is basically the same as that in grid-connected systems, employing PI control as the regulator for vector-oriented control. However, PI control has inherent problems such as parameter design being susceptible to external interference and poor dynamic response. Furthermore, considering VSG-based stator voltage closed-loop control, the inner-loop PI control can also lead to issues such as parameter mismatch between the inner and outer loops and difficulty in adjustment in the overall RSC control design. To simplify the design of the inner and outer loop controllers, traditional off-grid control methods typically treat load current, stator-side converter current, and voltage coupling terms as disturbances, essentially a no-load voltage build-up control. Therefore, this invention aims to compensate for voltage coupling effects and current changes caused by sudden load changes in real time, enhancing the dynamic response capability of off-grid DFIG wind turbines to load variations. Based on the stator voltage amplitude and frequency reference commands generated by the virtual synchronous outer loop control, this invention analyzes the processing methods of voltage coupling and current compensation terms in the outer loop control and proposes a current-compensated three-vector model predictive control method for off-grid DFIG wind turbines. This method solves the aforementioned problems while improving the accuracy of the inner-loop control.

[0102] Based on the state equations of DFIG, a forward Euler transformation is performed with a sampling period T. s Discretization yields the following rotor current prediction equation:

[0103]

[0104] Based on the rotor current prediction equation, and considering that the RSC AC side output voltage vector is the same as the DFIG rotor voltage vector, it can be seen that the predicted rotor current value is determined by the rotor voltage vector. Define the switching function vector S of the RSC three-phase bridge arm. R as follows:

[0105] S R =[S a S b S c ] T

[0106] Among them, S a S b S c These represent the switching functions of each phase bridge arm, and there are only two states: 0 or 1. When the switching function is 1, it means that the upper bridge arm is on and the lower bridge arm is off; when it is 0, it means the opposite.

[0107] Based on the working principle of a three-phase voltage source inverter, the mathematical relationship between the rotor voltage and the switching function in the dq coordinate system can be obtained as follows:

[0108]

[0109] In summary, by obtaining the mathematical relationship between the rotor current and the RSC switching function, a model predictive control algorithm can be used to perform rolling optimization of the inner loop of the rotor current, thereby obtaining the optimal control output S. R * .

[0110] The current compensation strategy is designed as follows:

[0111] The stator voltage d-axis orientation is as follows:

[0112]

[0113] Substituting this into the stator voltage equation of DFIG, while neglecting the stator resistance and assuming that the differential term is zero in steady state, we have the following relationship:

[0114]

[0115] Substituting the above equation into the stator flux linkage equation of the DFIG, we obtain the following expression:

[0116]

[0117] In addition, considering that Figure 1 A filter capacitor C is connected in parallel at PCC. g After coordinate transformation, the following relationship exists at PCC:

[0118]

[0119] Combining the above two equations, the rotor current equation for stator voltage orientation can be obtained as follows:

[0120]

[0121] As shown in the above equation, when performing stator voltage control on an off-grid doubly-fed wind turbine, it is necessary to obtain i through calculation of the current equation at the PCC. rd and i rq This serves as the RSC inner loop control input reference. That is, it is necessary to control the rotor current i. rd and i rq This is to achieve control of the stator voltage. rd and i rq The control can be divided into two parts. The first is to dynamically regulate the stator voltage differential term, which has already been controlled in the voltage outer loop. The second is to overcome the influence of the coupled disturbance term and the current compensation term on the stator voltage in the equation. A current compensation strategy is designed, and the compensation amount of the coupled disturbance term is obtained through the above calculation. At the same time, the SSC AC current and the load current are introduced as current compensation terms, and the current compensation equation is obtained as follows:

[0122]

[0123] Further analysis is conducted on the decoupling effect of the current compensation strategy. After introducing compensation, the outer-loop voltage control equation is as follows:

[0124]

[0125] Combining the above three equations, we can obtain the following expression:

[0126]

[0127] As can be seen from the above formula, after introducing the current compensation strategy, the outer loop of the DFIG stator voltage achieves complete decoupling control of the d and q axes.

[0128] Traditional MPC algorithms are single-vector MPCs, meaning they use only one basic voltage vector in each sampling period, resulting in high harmonic content in both voltage and current. Therefore, this invention proposes a current-compensated three-vector model predictive control algorithm for off-grid DFIG. Based on a current compensation strategy, it uses three basic vectors in each sampling period: two adjacent non-zero vectors u... i and u j And a zero vector u0, combined to form a composite voltage vector u tn To track the reference vector.

[0129] The relationship between the synthesized voltage vector and the fundamental voltage vector is as follows:

[0130]

[0131] t i +t j +t0=T s

[0132] As can be seen from the above formula, the six voltage vectors synthesized by the current-compensated three-vector model predictive control algorithm can have their amplitude and direction freely adjusted, covering the entire vector action plane. Therefore, the control accuracy can be improved and the optimal control output can be obtained.

[0133] After determining the voltage vector combination, a modulation model predictive control algorithm is used to calculate the duration of each basic voltage vector. This algorithm assumes that the duration of each basic voltage vector is inversely proportional to its corresponding cost function, thus effectively reducing the computational load while achieving the same control effect. The durations of each basic voltage vector are thus obtained as follows:

[0134]

[0135] The above analysis shows that the stator voltage d-axis component u of the off-grid doubly-fed induction generator is... sd Directly derived from the d-axis component i of the rotor current rdControl, stator voltage q-axis component u sq Directly derived from the q-axis component i of the rotor current rq Control. Therefore, rotor current is chosen as the state variable. Based on the current compensation strategy, the control objective is to minimize the rotor current prediction error. The cost function is designed as follows:

[0136] g = |i rdref -i rd (k+1)|+|i rqref -i rq (k+1)|

[0137] The cost function value g when the synthesized voltage vector is applied. tn as follows:

[0138] g tn =g i t i +g j t j +g0t0

[0139] By using a rolling optimization of the cost function, the synthesized voltage vector u corresponding to the minimum cost function value is selected. tn This is the optimal voltage vector, and the output control signal is applied to the converter to achieve current-compensated three-vector model predictive control. In summary, the principle block diagram of the current-compensated three-vector model predictive control algorithm is as follows: Figure 4 As shown.

[0140] In summary, this implementation first employs VSG control in the outer loop control of the off-grid doubly-fed induction generator (DFIG). The angular frequency and voltage amplitude generated by VSG control serve as the angular frequency reference command and stator voltage amplitude reference command in the off-grid control, enabling the DFIG to achieve active frequency regulation and active voltage support during off-grid operation. Secondly, to address the issues of voltage sag and high total harmonic distortion (THD) in the microgrid system caused by sudden load changes during off-grid operation of the DFIG, this implementation proposes a current-compensated three-vector model predictive control by real-time compensation for voltage coupling effects and current changes caused by sudden load changes. The cost function is designed with the goal of minimizing rotor current prediction error, achieving accurate rotor current prediction during sudden load changes. This effectively suppresses voltage sags caused by sudden load changes, reduces the harmonic content of the output voltage, and simultaneously achieves rapid response to system frequency and load power.

[0141] Although the present invention is described herein with reference to specific embodiments, it should be understood that these embodiments are merely illustrative of the principles and applications of the invention. It should be understood that many modifications may be made to the illustrative embodiments, and that other arrangements may be devised, without departing from the spirit and scope of the invention as defined by the appended claims. It should be understood that the various dependent claims and features described herein may be combined in ways other than those described in the original claims. It should also be understood that features described in conjunction with individual embodiments may be used in conjunction with other described embodiments.

Claims

1. A current-compensated three-vector model predictive control method for off-grid doubly-fed wind turbines, characterized in that, The method includes the following steps: Step 1: Establish a rotor current prediction model in the inner loop of the RSC control of the rotor-side converter. Based on the DFIG state equation, discretize it using the forward Euler method to obtain the rotor current prediction equation in the dq coordinate system. The rotor current prediction model is associated with the rotor voltage and the switching function. Voltage vector synthesis is achieved through the combination of switching states. Step 2: Design a current compensation strategy by introducing load current i L Stator-side converter current i g and stator voltage u s Calculate the current coupling compensation amount and generate the reference value of the rotor current after compensation; Step 3: Execute three-vector model predictive control: Based on the mapping relationship between rotor voltage and switching function, the eight switching states of the rotor-side converter RSC are converted into basic voltage vectors in the dq coordinate system; Within each sampling period, two adjacent non-zero vectors are selected and combined with the zero vector to synthesize the target voltage vector; The cost function is designed with the goal of minimizing the rotor current tracking error, and the optimal switching signal is output to the rotor-side converter RSC.

2. The off-grid doubly-fed induction generator current-compensated three-vector model predictive control method according to claim 1, characterized in that, The rotor current prediction equation in step 1 is: In the formula, σ is the leakage flux coefficient. L m For the equivalent mutual inductance between the stator and rotor windings, L s For the equivalent stator winding self-inductance, L r This is the equivalent rotor winding self-inductance; u sd ,u sq and u rd ,u rq These are the stator voltages u s and rotor voltage u r The d-axis and q-axis components; i sd ,i sq and i rd ,i rq These are the stator currents i s and rotor current i r The d-axis and q-axis components; k represents the current sampling time, and k+1 represents the next sampling time; C is the coupling coefficient matrix of the stator current. B is the coupling coefficient matrix of the rotor current. R s and R r ω is the equivalent resistance of the stator and rotor windings; s For synchronous rotational angular velocity, ω m The rotor's electric angular velocity; T s The sampling period.

3. The off-grid doubly-fed induction generator current-compensated three-vector model predictive control method according to claim 2, characterized in that, The mapping relationship between rotor voltage and switching function is as follows: Where, θ sl Let θ be the slip angle, and θ be the slip angle. sl =ω sl t, ω sl Let ω be the slip angular velocity. sl =ω s -ω m ;U dc RSC DC side voltage; S R S is the switching function vector of the three-phase bridge arm of the RSC bridge. R =[S a S b S c ] T S a ,S b ,S c This is the switching function for the three-phase bridge arm.

4. The off-grid doubly-fed induction generator current-compensated three-vector model predictive control method according to claim 2, characterized in that, The current compensation strategy in step 2 includes: Step 21: Stator voltage d-axis orientation: Among them, u sm This refers to the stator voltage amplitude. And by controlling the rotor current i rd and i rq To achieve control of the stator voltage; Step 22: Calculate the current coupling compensation amount i rdc ,i rqc : In the formula, i rdc and i rqc These are the compensation values ​​for the d-axis and q-axis components of the rotor current, respectively; C g The filter capacitor connected in parallel at the common connection point of the load, i gd 、i gq For the stator-side converter SSC AC side current i g d-axis and q-axis components; i Ld 、i Lq For the load current i L The d-axis and q-axis components; Step 23: Generate the compensated rotor current reference value i rdref ,i rqref : Among them, K usP and K usI For the proportional gain and integral gain of the stator voltage control loop; u sdref and u sqref The stator voltage d-axis and q-axis component reference commands are generated by the virtual synchronous generator (VSG).

5. The off-grid doubly-fed induction generator current-compensated three-vector model predictive control method according to claim 4, characterized in that, In step 21, the rotor current i rd and i rq The calculation process includes: Substituting the stator voltage d-axis orientation relation into the DFIG voltage equation and flux linkage equation, let R s =0, and the differential term is zero in steady state. Therefore, the relationship between stator current and rotor current in the dq coordinate system is as follows: Considering that a filter capacitor C is connected in parallel at the common connection point PCC of the load. g After coordinate transformation, the following relationship exists at PCC: Substituting the relationship between stator current and rotor current in the dq coordinate system into the above equation, we obtain the rotor current equation at PCC when the stator voltage is oriented as follows: When performing stator voltage control on an off-grid doubly-fed induction generator (DFIG), it is necessary to obtain i through calculations of the current equation at the PCC. rd and i rq As the RSC inner loop control input reference, it is necessary to control the rotor current i rd and i rq This is to achieve control of the stator voltage.

6. The off-grid doubly-fed induction generator current-compensated three-vector model predictive control method according to claim 1, characterized in that, Step 3, converting the switch state into a basic voltage vector, includes: RSC adopts a two-level three-phase voltage source inverter topology, with a total of 8 switching state combinations; Based on the mapping relationship between rotor voltage and switching function, the eight switching states of the rotor-side converter RSC are converted into basic voltage vectors in the dq coordinate system. These basic voltage vectors include six non-zero basic voltage vectors and two zero voltage vectors; the amplitude of the non-zero vectors is 2 / 3U. dc They are distributed in a hexagonal pattern in the dq plane, with adjacent vectors spaced 60° apart.

7. The off-grid doubly-fed induction generator current-compensated three-vector model predictive control method according to claim 6, characterized in that, Step 3: Synthesize the target voltage vector u tn The process includes: Three fundamental vectors are used in each sampling period, namely two adjacent non-zero vectors u. i and u j And a zero vector u0, combined to form a synthetic target voltage vector u tn To track the reference vector, the target voltage vector is synthesized using the following formula: In the formula, t i t j t0 and t0 are the three basic voltage vectors u i u j The duration of action of u0 and u0 satisfy the following relationship: t i +t j +t0=T s The time allocation for each basic voltage vector is as follows: Among them, g i g j and g0 are u i u j The cost function value when u0 is interacting with it.

8. The off-grid doubly-fed induction generator current-compensated three-vector model predictive control method according to claim 7, characterized in that, The cost function for step 3 is: g=|i rdref -i rd (k+1)|+|i rqref -i rq (k+1)| The cost function value g when the synthesized voltage vector is applied tn Represented as: g tn =g i t i +g j t j +g0t0 By using a rolling optimization of the cost function, the synthesized voltage vector u corresponding to the minimum cost function value is selected. tn This is the optimal voltage vector, and the output control signal is applied to the rotor-side converter RSC to achieve current-compensated three-vector model predictive control.

Images