Improved ADRC controller-based rapid no-load grid connection method for variable-speed pumped storage unit

Abstract

The invention provides a variable-speed pumped storage unit rapid no-load grid connection method based on an improved ADRC controller, and the method comprises the steps: simplifying a variable-speed pumped storage unit into a single AC excitation asynchronous generator, and building a state space model of the AC excitation asynchronous generator under a no-load grid connection condition; an ADRC controller is constructed for the state space model, an extended Kalman state observer is introduced to improve the ADRC controller, and an EKO-ADRC controller is obtained; and the EKO-ADRC controller is used to carry out closed-loop control on the stator voltage amplitude, phase and frequency of the AC excitation asynchronous generator until the stator voltage amplitude, phase and frequency of the AC excitation asynchronous generator are consistent with the voltage amplitude, phase and frequency of a to-be-connected power grid, and zero impact current grid connection is realized. The invention provides a quicker no-load grid connection method for switching in power grid control, so that the soft switching in power grid working condition speed is effectively improved.

Description

Technical Field

[0001] This application relates to the field of doubly-fed generator control technology, and in particular to a method for rapid no-load grid connection of a variable-speed pumped-storage unit based on an improved ADRC controller. Background Technology

[0002] Pumped storage hydropower stations are currently the mainstream type of hydropower station. Pumped storage units, as one of the means of storing large-capacity hydropower resources, have functions such as peak shaving, valley filling, frequency regulation, phase regulation, and emergency backup. Doubly fed variable speed pumped storage generator units are a new type of hydropower storage-generation system, characterized by high economy, large capacity, wide power regulation range, and fast speed. In the context of large-scale renewable energy grid integration, variable speed units can undertake tasks such as peak shaving, valley filling, and frequency regulation.

[0003] The operating mode switching of variable speed pumped storage units mainly includes two parts: grid connection control and grid disconnection control. The mathematical models for variable speed pumped storage units in grid connection control mode and grid disconnection control mode are completely different, with grid connection control being more complex. Grid connection control for variable speed units is mainly divided into no-load grid connection, islanded grid connection, and loaded grid connection, with no-load grid connection being the simplest. No-load grid connection refers to the variable speed pumped storage unit having no load connected to its terminals, meaning the stator current is zero. In this case, the output voltage at the turbine terminals is controlled by the excitation current input to the rotor terminals. Therefore, no-load grid connection is currently the most important method for grid connection control research in existing technologies.

[0004] Because variable-speed generator sets have large unit capacity and very low stator leakage inductance, even a slight error in amplitude, frequency, and phase between the generator terminal voltage and the grid voltage can cause a huge inrush current during grid connection. This not only affects the operation of the variable-speed generator set but also significantly impacts the voltage stability of the power grid. The requirement for no-load grid connection of variable-speed generator sets is, under the premise of applying control algorithms, to ensure that the amplitude, frequency, and phase of the generator terminal voltage are consistent with the grid voltage, thus meeting the grid connection conditions and achieving zero-current inrush grid connection.

[0005] For grid connection control of variable-speed pumped storage units, the main focus is on the grid connection control problem of doubly-fed induction generators (DFIGs). Currently, the main control strategies employed by scholars include PID control, sliding mode control, and model predictive control. Existing technology "CN105915134A" discloses a method for no-load grid connection control of a DFIG, including: 1. Obtaining the stator flux linkage angular velocity from the grid voltage signal; 2. Obtaining the rotor current reference value; 3. Obtaining the rotor position angle as the reference angle for the coordinate rotation change of the control variable; 4. Calculating the rotor voltage q-axis reference value; 5. Obtaining the rotor current d-axis actual value idr; 6. Using the reference value, actual value, and their deviation e of the rotor current as inputs to a neural network controller to obtain the rotor voltage d-axis reference value, and then performing an inverse coordinate transformation to obtain the signal for regulating the converter device, adjusting the amplitude, frequency, and phase of the excitation current so that the stator terminal voltage of the DFIG meets the grid connection requirements to achieve no-load grid connection of the DFIG. This existing technology combines traditional PID control with BP neural network control to control the no-load grid connection process of a doubly-fed generator, which can make the stator voltage close to the grid voltage, and has the advantages of fast stator voltage response and small error. However, the existing technology still has the problem of insufficient speed of no-load grid connection. Summary of the Invention

[0006] To overcome the technical problem of insufficient speed of no-load grid connection in the prior art, the present invention provides a method for rapid no-load grid connection of variable speed pumped storage units based on an improved ADRC controller.

[0007] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows: A method for rapid no-load grid connection of a variable-speed pumped-storage unit based on an improved ADRC controller includes the following steps: S1: Simplify the variable speed pumped storage unit into a single AC excitation asynchronous generator, and establish the state-space model of the AC excitation asynchronous generator under no-load grid-connected conditions. S2: Construct an ADRC controller for the state-space model, and improve the ADRC controller by introducing an extended Kalman state observer to obtain the EKO-ADRC controller; S3: Using the EKO-ADRC controller, the stator voltage amplitude, phase and frequency of the AC excitation asynchronous generator are controlled in a closed loop until the stator voltage amplitude, phase and frequency of the AC excitation asynchronous generator are consistent with the voltage amplitude, phase and frequency of the grid to be connected, so as to achieve zero inrush current grid connection.

[0008] Preferably, in step S1, the state-space model of the AC excitation asynchronous generator under no-load grid-connected conditions is represented as follows:

[0009] in, The rotor voltage vector serves as the control vector. The stator voltage vector serves as the output vector. The rotor current vector serves as the state vector. This is the voltage across the d-axis of the generator rotor. This is the q-axis voltage of the generator rotor. This refers to the d-axis voltage of the generator stator. This refers to the q-axis voltage of the generator stator. This refers to the d-axis current of the generator rotor. Let be the q-axis current of the generator rotor; A, B, C, and D are the first to fourth state space matrices, respectively; The state space matrices A, B, C, and D are represented as follows:

[0010]

[0011] in For generator stator resistance, For generator stator inductance, For generator rotor inductance, The generator's self-inductance coefficient. For the generator rotor speed, For grid voltage angle, generator slip speed It can be represented as: .

[0012] Preferably, in step S2, the ADRC controller is a second-order ADRC controller, including a TD module, an ESO module, and an NLSEF module.

[0013] Preferably, the TD module adopts a single-input dual-output architecture, including a nonlinear differential tracker to avoid input jumps and smooth the input. The nonlinear differential tracker in the discrete state is represented as follows:

[0014] in For target tracking signals; Let be the input signal at time k; This is the input signal at time k+1; Input signal The differential signal; for The state at the next moment; Let be the error between the input signal and the tracked signal; r be the velocity factor; h be the sampling period; and fhan(·) be the fastest control synthesis function, expressed as:

[0015] in This represents the boundary of the linear region. The boundary of the switching zone is y; y is the predicted future error value. The differential form of the input signal ; For intermediate control quantities; It is a linear control variable; the Sgn(·) function is a sign function, and its return value is based on the sign of the input value. If the input value is positive, it returns 1; if the input value is 0, it returns 0; if the input value is negative, it returns -1. The ESO module adopts a dual-input, three-output port architecture, including an ESO extended state observer. The inputs to the ESO module are the output value and the input observation value of the controlled system, respectively. The outputs of the ESO module are the output estimate of the controlled system, its first derivative, and its second derivative, respectively. The output estimate and its first derivative of the controlled system are fed back to the nonlinear differential tracker, and the second derivative of the output estimate is fed back to the input of the controlled system as a compensation perturbation. The ESO extended state observer in the discrete state is represented as follows:

[0016] in This represents the output value of the controlled system at time k. The input observation value of the controlled system at time k; for The differential form; The observed value of the derivative is input to the controlled system at time k; for The differential form; Let be the observed value of the total disturbance of the controlled system at time k; for The differential form; , , These are the parameters for the first, second, and third ESO observers, respectively. Here are the control gain parameters for the ESO extended state observer; fal(·) is the saturation function, expressed as:

[0017] Where e is the error between the output value of the controlled system and the input signal; Indicates the nonlinear factor; The preset linear interval length; These are the nonlinear factors of the first and second fal(·) functions, respectively; The NLSEF module adopts a dual-input single-output architecture and includes a nonlinear state error feedback controller; in discrete states, the nonlinear state error feedback controller is represented as:

[0018] in This refers to the error between the preset intermediate process and the observed state; The error between the preset differential signal and the observed differential state; , These are the parameters for the fourth and fifth ESO observers, respectively. , These are the nonlinear factors of the third and fourth fal(·) functions, respectively; To control the gain; For preliminary control purposes; It is the final control variable of the controlled system.

[0019] Preferably, in step S2, the state-space model is linearized using a first-order Taylor expansion, expressed as:

[0020]

[0021] in Let k be the state vector at time k. Let this be the state vector at time k-1; Let k be the control vector at time k. This is the control vector at time k-1; This represents the process noise at time k-1. The measurement error noise at time k; Let T' be the output vector at time k; T' be the total sampling time. Let be the identity matrix, e be the natural logarithm; F, H, M, and G are the first to fourth linearization parameter matrices, respectively; The covariance matrix of the noise is constructed as follows:

[0022] Where Q is the process noise covariance matrix, matrix R is the measurement error noise covariance matrix; T represents matrix transpose; w and v represent process noise and measurement error noise, respectively; The observation estimation error of the state-space model at time k is:

[0023] in Let be the observed estimate of the state vector at time k. The error between the state vector at time k and the observed estimate of the state vector; Further construct the covariance matrix of the observation estimation error of the state-space model at time k, which is expressed as: .

[0024] Preferably, in step S2, the iterative process of the extended Kalman state observer includes: Calculate the prior state estimate:

[0025] in, This is the prior state estimate of the state vector at time k; This is the observed estimate of the state vector at time k-1; Calculate the prior estimate of the covariance matrix:

[0026] The prior state estimate is the covariance matrix of the observation estimation error at time k. Let be the covariance matrix of the observation estimation error at time k-1; Calculate the Kalman gain value :

[0027] Calculate the posterior state estimate:

[0028] Finally, update the covariance matrix: .

[0029] Preferably, in step S3, the EKO-ADRC controller is used to perform closed-loop control of the stator voltage amplitude of the AC excitation asynchronous generator, including: The voltage amplitude to be connected to the grid is used as the input of the TD module in the EKO-ADRC controller, and the stator voltage amplitude is used as the input of the ESO module in the EKO-ADRC controller. Inside the EKO-ADRC controller, the error of the two input amplitude signals and the error of their derivatives are fed back to the NLSEF module to form the control signal of the generator rotor excitation current, thereby controlling the amplitude of the generator stator voltage.

[0030] Preferably, in step S3, the EKO-ADRC controller is used to perform closed-loop control of the stator voltage phase of the AC excitation asynchronous generator, including: The voltage phase to be connected to the grid is used as the input of the EKO-ADRC controller TD module, and the stator voltage phase is used as the input of the EKO-ADRC controller ESO module. In the EKO-ADRC controller, the phase difference between the two phases is calculated to form a phase difference, which is fed back to the NLSEF module together with the error of the phase derivative of the two phases. Finally, the compensation angle control signal is obtained and converted into a radian compensation angle to directly compensate the phase of the excitation current, and thus indirectly compensate the phase of the generator stator voltage.

[0031] Preferably, in step S3, the EKO-ADRC controller is used to perform closed-loop control of the stator voltage frequency of the AC excitation asynchronous generator, including: The voltage frequency to be connected to the grid is used as the input of the TD module of the EKO-ADRC controller, and the stator voltage frequency is used as the input of the ESO module of the EKO-ADRC controller. Inside the EKO-ADRC controller, the error between the two input frequencies and the differential error are fed back to the NLSEF module to generate a control signal for the stator voltage frequency, thereby compensating for the frequency of the generator stator voltage.

[0032] Preferably, an S-Function simulation model based on the state-space model of the AC-excited asynchronous generator is designed in Simulink, and the effectiveness of the method is verified by the RTLAB hardware-in-the-loop semi-physical simulation platform.

[0033] Compared with the prior art, the beneficial effects of the technical solution of the present invention are: This invention provides a method for rapid no-load grid connection of a variable-speed pumped-storage unit based on an improved ADRC controller, comprising: simplifying the variable-speed pumped-storage unit into a single AC-excited asynchronous generator; establishing a state-space model of the AC-excited asynchronous generator under no-load grid connection conditions; constructing an ADRC controller for the state-space model; introducing an extended Kalman state observer to improve the ADRC controller, resulting in an EKO-ADRC controller; and using the EKO-ADRC controller to perform closed-loop control on the stator voltage amplitude, phase, and frequency of the AC-excited asynchronous generator until the stator voltage amplitude, phase, and frequency of the AC-excited asynchronous generator are consistent with the voltage amplitude, phase, and frequency of the grid to be connected, thereby achieving zero-inrush current grid connection.

[0034] This invention addresses the shortcomings of ADRC controllers in handling rotor current noise in AC-excited asynchronous generator systems. It designs a Kalman state observer to improve the ADRC controller and enhance its control performance. A closed-loop control strategy based on stator voltage amplitude, frequency, and phase is employed to control the AC-excited asynchronous generator's no-load grid connection, improving the speed of no-load grid connection. Compared to traditional control methods, this invention effectively accelerates the no-load grid connection speed of variable-speed pumped storage units, and the method has been verified in Simulink simulation experiments and RTLAB hardware-in-the-loop experimental platform. Compared to traditional control strategies, this invention improves the speed of soft grid connection. Attached Figure Description

[0035] Figure 1 This is a flowchart of a method for rapid no-load grid connection of a variable speed pumped storage unit based on an improved ADRC controller in Example 1; Figure 2 This is a simplified model diagram of a variable speed pumped storage unit based on an improved ADRC controller for rapid no-load grid connection in Example 2. Figure 3 This is a schematic diagram of stator flux orientation control for a method of rapid no-load grid connection of a variable speed pumped storage unit based on an improved ADRC controller in Example 2. Figure 4 This is a second-order ADRC control system framework diagram of a method for rapid no-load grid connection of a variable-speed pumped-storage unit based on an improved ADRC controller in Example 2. Figure 5 This is a structural diagram of the improved ADRC controller based on the extended Kalman state observer, which is used in Example 2 for a method of rapid no-load grid connection of a variable speed pumped storage unit based on an improved ADRC controller. Figure 6 This is a flowchart of the open-loop control for no-load pressure building based on EKO-ADRC, which is a method for rapid no-load grid connection of a variable speed pumped storage unit based on an improved ADRC controller in Example 2. Figure 7 This is a structural diagram of a closed-loop control method based on stator voltage for a fast no-load grid connection method of a variable speed pumped storage unit based on an improved ADRC controller, as shown in Example 2. Figure 8 This is a flowchart of a closed-loop control method based on stator voltage amplitude for a variable speed pumped storage unit with rapid no-load grid connection based on an improved ADRC controller, as described in Example 2. Figure 9 This is a flowchart of a closed-loop control method based on stator voltage phase angle for a variable speed pumped storage unit with rapid no-load grid connection based on an improved ADRC controller, as described in Example 2. Figure 10This is a flowchart of a closed-loop control method based on stator voltage frequency for a rapid no-load grid connection method of a variable speed pumped storage unit based on an improved ADRC controller, as described in Example 2. Figure 11 This is a Simulink model diagram of a generator fast no-load grid connection control system for a variable speed pumped storage unit based on an improved ADRC controller, as shown in Example 2. Figure 12 This is a simulation result diagram of the amplitude, phase and frequency output of the generator terminal voltage under open-loop control in Example 2, which is a fast no-load grid connection method for a variable speed pumped storage unit based on an improved ADRC controller. Figure 13 The figure shows the simulation results of the amplitude, phase and frequency output of the generator terminal voltage under closed-loop control in Example 2 of a method for rapid no-load grid connection of a variable speed pumped storage unit based on an improved ADRC controller. Figure 14 The figure shows the simulation results of the three-phase voltage error between the generator terminal and the grid segment under the classic PID, classic ADRC, and ADRC controllers based on the improved Kalman state observer in Example 2, which is a method for rapid no-load grid connection of a variable speed pumped storage unit based on an improved ADRC controller. Figure 15 The figure shows the simulation results of the three-phase voltage error between the generator terminal and the grid segment under the classic PID, classic ADRC, and ADRC controllers based on the improved Kalman state observer in Example 2, which is a method for rapid no-load grid connection of a variable speed pumped storage unit based on an improved ADRC controller. Figure 16 This is a flowchart of the RTLAB hardware-in-the-loop experimental platform for a method of rapid no-load grid connection of a variable speed pumped storage unit based on an improved ADRC controller, as shown in Example 2. Figure 17 This is a simulation result of the voltage amplitude output at the oscilloscope terminal and the grid terminal of the RTLAB hardware-in-the-loop experimental device for a method of rapid no-load grid connection of a variable speed pumped storage unit based on an improved ADRC controller in Example 2. Figure 18 This is a simulation result of the voltage phase output at the oscilloscope terminal and the grid terminal of the RTLAB hardware-in-the-loop experimental device for a method of rapid no-load grid connection of a variable speed pumped storage unit based on an improved ADRC controller in Example 2. Figure 19 This is a simulation result of the voltage and frequency output of the oscilloscope at the machine end and the grid end of the RTLAB hardware-in-the-loop experimental device for a method of rapid no-load grid connection of a variable speed pumped storage unit based on an improved ADRC controller in Example 2. Figure 20This is a simulation result of the single-phase voltage tracking output of the oscilloscope at the machine end and the grid end of the RTLAB hardware-in-the-loop experimental device for a method of rapid no-load grid connection of a variable speed pumped storage unit based on an improved ADRC controller in Example 2. Figure 21 This is a simulation result of the three-phase voltage error output between the oscilloscope terminal and the grid terminal of the RTLAB hardware-in-the-loop experimental device for a method of rapid no-load grid connection of a variable speed pumped storage unit based on an improved ADRC controller in Example 2. Figure 22 This is a structural diagram of a rapid no-load grid connection system for a variable speed pumped storage unit based on an improved ADRC controller, as shown in Example 3. Detailed Implementation

[0036] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this application. To better illustrate this embodiment, some parts in the accompanying drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions; It will be understood by those skilled in the art that certain well-known structures and their descriptions may be omitted in the accompanying drawings.

[0037] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0038] Example 1 A method for rapid no-load grid connection of variable-speed pumped-storage units based on an improved ADRC controller, such as... Figure 1 As shown, it includes the following steps: S1: Simplify the variable speed pumped storage unit into a single AC excitation asynchronous generator, and establish the state-space model of the AC excitation asynchronous generator under no-load grid-connected conditions. S2: Construct an ADRC controller for the state-space model, and improve the ADRC controller by introducing an extended Kalman state observer to obtain the EKO-ADRC controller; S3: Using the EKO-ADRC controller, the stator voltage amplitude, phase and frequency of the AC excitation asynchronous generator are controlled in a closed loop until the stator voltage amplitude, phase and frequency of the AC excitation asynchronous generator are consistent with the voltage amplitude, phase and frequency of the grid to be connected, so as to achieve zero inrush current grid connection.

[0039] In the specific implementation process, the variable-speed pumped-storage unit is first simplified into an AC-excited asynchronous motor. The current state equations and voltage state equations of the AC-excited asynchronous motor under no-load conditions at the stator end are obtained. Then, the obtained current and voltage state equations are transformed into the state-space expression of the AC-excited asynchronous motor under no-load grid connection. Next, a classic ADRC controller is designed based on ADRC technology. Subsequently, based on the shortcomings of the classic ADRC control theory, the state-space equation of the AC-excited asynchronous motor is linearized using a first-order Taylor expansion. Finally, an improved ADRC control method based on an extended Kalman state observer is proposed. To address the shortcomings of traditional open-loop control strategies for no-load voltage building, a closed-loop control method was designed from three perspectives: grid voltage amplitude, phase, and frequency. This involves using the three fundamental characteristics of the grid voltage signal to perform closed-loop control of the generator stator output voltage. Finally, the designed closed-loop control strategy based on stator voltage and the improved ADRC control strategy based on an extended Kalman state observer were tested on the Simulink platform and the RTLAB hardware-in-the-loop experimental platform. An AC-excited asynchronous generator no-load grid-connected experimental circuit was built, and simulation and semi-physical experiments were conducted to verify the feasibility of the two designed control strategies.

[0040] Example 2 A method for rapid no-load grid connection of a variable-speed pumped-storage unit based on an improved ADRC controller includes the following steps: S1: Simplify the variable speed pumped storage unit into a single AC excitation asynchronous generator, and establish the state-space model of the AC excitation asynchronous generator under no-load grid-connected conditions. S2: Construct an ADRC controller for the state-space model, and improve the ADRC controller by introducing an extended Kalman state observer to obtain the EKO-ADRC controller; S3: Using the EKO-ADRC controller, the stator voltage amplitude, phase and frequency of the AC excitation asynchronous generator are controlled in a closed loop until the stator voltage amplitude, phase and frequency of the AC excitation asynchronous generator are consistent with the voltage amplitude, phase and frequency of the grid to be connected, so as to achieve zero inrush current grid connection.

[0041] In step S1, the state-space model of the AC excitation asynchronous generator under no-load grid-connected conditions is represented as follows:

[0042] in, The rotor voltage vector serves as the control vector. The stator voltage vector serves as the output vector. The rotor current vector serves as the state vector. This is the voltage across the d-axis of the generator rotor. This is the q-axis voltage of the generator rotor. This refers to the d-axis voltage of the generator stator. This refers to the q-axis voltage of the generator stator. This refers to the d-axis current of the generator rotor. Let be the q-axis current of the generator rotor; A, B, C, and D are the first to fourth state space matrices, respectively; The state space matrices A, B, C, and D are represented as follows:

[0043]

[0044] in For generator stator resistance, For generator stator inductance, For generator rotor inductance, The generator's self-inductance coefficient. For the generator rotor speed, For grid voltage angle, generator slip speed It can be represented as: .

[0045] In step S2, the ADRC controller is a second-order ADRC controller, which includes a TD module, an ESO module, and an NLSEF module.

[0046] The TD module adopts a single-input dual-output architecture, including a nonlinear differential tracker to avoid input jumps and smooth the input. The nonlinear differential tracker in the discrete state is represented as follows:

[0047] in For target tracking signals; Let be the input signal at time k; This is the input signal at time k+1; Input signal The differential signal; for The state at the next moment; Let be the error between the input signal and the tracked signal; r be the velocity factor; h be the sampling period; and fhan(·) be the fastest control synthesis function, expressed as:

[0048] in This represents the boundary of the linear region. The boundary of the switching zone is y; y is the predicted future error value. The differential form of the input signal ; For intermediate control quantities; It is a linear control variable; the Sgn(·) function is a sign function, and its return value is based on the sign of the input value. If the input value is positive, it returns 1; if the input value is 0, it returns 0; if the input value is negative, it returns -1. The ESO module adopts a dual-input, three-output port architecture, including an ESO extended state observer. The inputs to the ESO module are the output value and the input observation value of the controlled system, respectively. The outputs of the ESO module are the output estimate of the controlled system, its first derivative, and its second derivative, respectively. The output estimate and its first derivative of the controlled system are fed back to the nonlinear differential tracker, and the second derivative of the output estimate is fed back to the input of the controlled system as a compensation perturbation. The ESO extended state observer in the discrete state is represented as follows:

[0049] in This represents the output value of the controlled system at time k. The input observation value of the controlled system at time k; for The differential form; The observed value of the derivative is input to the controlled system at time k; for The differential form; Let be the observed value of the total disturbance of the controlled system at time k; for The differential form; , , These are the parameters for the first, second, and third ESO observers, respectively. Here are the control gain parameters for the ESO extended state observer; fal(·) is the saturation function, expressed as:

[0050] Where e is the error between the output value of the controlled system and the input signal; Indicates the nonlinear factor; The preset linear interval length; These are the nonlinear factors of the first and second fal(·) functions, respectively; The NLSEF module adopts a dual-input single-output architecture and includes a nonlinear state error feedback controller; in discrete states, the nonlinear state error feedback controller is represented as:

[0051] in This refers to the error between the preset intermediate process and the observed state; The error between the preset differential signal and the observed differential state; , These are the parameters for the fourth and fifth ESO observers, respectively. , These are the nonlinear factors of the third and fourth fal(·) functions, respectively; To control the gain; For preliminary control purposes; It is the final control variable of the controlled system.

[0052] In step S2, the state-space model is linearized using a first-order Taylor expansion, and expressed as:

[0053]

[0054] in Let k be the state vector at time k. Let this be the state vector at time k-1; Let k be the control vector at time k. This is the control vector at time k-1; This represents the process noise at time k-1. The measurement error noise at time k; Let T' be the output vector at time k; T' be the total sampling time. Let be the identity matrix, e be the natural logarithm; F, H, M, and G are the first to fourth linearization parameter matrices, respectively; The covariance matrix of the noise is constructed as follows:

[0055] Where Q is the process noise covariance matrix, matrix R is the measurement error noise covariance matrix; T represents matrix transpose; w and v represent process noise and measurement error noise, respectively; The observation estimation error of the state-space model at time k is:

[0056] in Let be the observed estimate of the state vector at time k. The error between the state vector at time k and the observed estimate of the state vector; Further construct the covariance matrix of the observation estimation error of the state-space model at time k, which is expressed as: .

[0057] In step S2, the iterative process of the extended Kalman state observer includes: Calculate the prior state estimate:

[0058] in, This is the prior state estimate of the state vector at time k; This is the observed estimate of the state vector at time k-1; Calculate the prior estimate of the covariance matrix:

[0059] The prior state estimate is the covariance matrix of the observation estimation error at time k. Let be the covariance matrix of the observation estimation error at time k-1; Calculate the Kalman gain value :

[0060] Calculate the posterior state estimate:

[0061] Finally, update the covariance matrix: .

[0062] In step S3, the EKO-ADRC controller is used to perform closed-loop control of the stator voltage amplitude of the AC excitation asynchronous generator, including: The voltage amplitude to be connected to the grid is used as the input of the TD module in the EKO-ADRC controller, and the stator voltage amplitude is used as the input of the ESO module in the EKO-ADRC controller. Inside the EKO-ADRC controller, the error of the two input amplitude signals and the error of their derivatives are fed back to the NLSEF module to form the control signal of the generator rotor excitation current, thereby controlling the amplitude of the generator stator voltage.

[0063] In step S3, the EKO-ADRC controller is used to perform closed-loop control of the stator voltage phase of the AC excitation asynchronous generator, including: The voltage phase to be connected to the grid is used as the input of the EKO-ADRC controller TD module, and the stator voltage phase is used as the input of the EKO-ADRC controller ESO module. In the EKO-ADRC controller, the phase difference between the two phases is calculated to form a phase difference, which is fed back to the NLSEF module together with the error of the phase derivative of the two phases. Finally, the compensation angle control signal is obtained and converted into a radian compensation angle to directly compensate the phase of the excitation current, and thus indirectly compensate the phase of the generator stator voltage.

[0064] In step S3, the stator voltage frequency of the AC excitation asynchronous generator is controlled in a closed loop using the EKO-ADRC controller, including: The voltage frequency to be connected to the grid is used as the input of the TD module of the EKO-ADRC controller, and the stator voltage frequency is used as the input of the ESO module of the EKO-ADRC controller. Inside the EKO-ADRC controller, the error between the two input frequencies and the differential error are fed back to the NLSEF module to generate a control signal for the stator voltage frequency, thereby compensating for the frequency of the generator stator voltage.

[0065] This embodiment also designs an S-Function simulation model based on the state-space model of the AC excitation asynchronous generator in Simulink, and verifies the effectiveness of the method through the RTLAB hardware-in-the-loop semi-physical simulation platform.

[0066] In practical implementation, this embodiment operates on a simplified model of a variable-speed pumped storage unit, as shown in the following figure. Figure 2 As shown, the no-load grid-connected system of a variable-speed pumped storage unit is first simplified to multiple AC-excited asynchronous generators connected to the grid under no-load conditions. Then, based on the stator voltage state equation and the rotor current state equation, a no-load grid-connected method and a stator flux-oriented vector control method are adopted, wherein the stator flux-oriented control is as follows: Figure 3 As shown, a mathematical model of a single AC-excited asynchronous generator connected to the grid under no-load conditions is obtained. An ADRC control theory method based on an improved Kalman state observer is applied to design a controller for the state-space model of the AC-excited asynchronous generator. The classic second-order ADRC control system framework diagram is shown below. Figure 4 As shown, the improved ADRC controller based on the Extended Kalman State Observer is as follows: Figure 5 As shown, to address the shortcomings of traditional no-load pressure build-up open-loop control strategies, the no-load pressure build-up open-loop control process based on EKO-ADRC is as follows: Figure 6 As shown, a closed-loop control strategy based on stator voltage amplitude, phase, and frequency is adopted to ensure that the generator terminal voltage is consistent with the grid voltage in terms of amplitude, phase, and frequency, thereby achieving zero-current-impact grid connection.

[0067] In this embodiment, an improved ADRC controller is mounted at the front end of the back-to-back converter and rotor current sensor. The EKO-ADRC (an improved ADRC based on a Kalman state observer) controller controls the SVPWM pulse signal of the back-to-back converter, thereby controlling the rotor excitation current of the AC-excited asynchronous generator. When the rotor speed of the variable-speed pumped storage generator reaches the predetermined rotor speed, the rotor-end current short-circuit is closed, and the stator-end current grid-connected circuit breaker is opened, achieving an open-circuit state at the stator end. The rotor excitation current generated by the back-to-back converter is then input to the rotor end of the generator, thereby controlling the generator's terminal voltage. The Kalman state observer is constructed using a first-order Taylor expansion to linearize the AC-excited asynchronous generator no-load grid-connected system, and the linearized system is then discretized to derive difference equations. Taylor's formula in A first-order Taylor expansion at this point can be expressed as:

[0068] By taking only the first two items, we can achieve the following. exist Linearized representation of points near the given location:

[0069] like Figure 7 As shown, a closed-loop control strategy based on the amplitude, phase, and frequency of the electronic voltage is adopted, enabling the generator terminal voltage amplitude, phase, and frequency to quickly track the amplitude, phase, and frequency of the grid voltage within a short time, thus meeting the grid connection requirements. This includes: like Figure 8 As shown, the voltage amplitude to be connected to the grid is used as the input of the TD module in the EKO-ADRC controller, and the stator voltage amplitude is used as the input of the ESO module in the EKO-ADRC controller. Inside the EKO-ADRC controller, the error of the two input amplitude signals and the error of their derivatives are fed back to the NLSEF module to form the control signal of the generator rotor excitation current, thereby controlling the amplitude of the generator stator voltage.

[0070] like Figure 9 As shown, the voltage phase to be connected to the grid is used as the input of the EKO-ADRC controller TD module, and the stator voltage phase is used as the input of the EKO-ADRC controller ESO module. In the EKO-ADRC controller, the phase difference between the two phases is used to form a phase difference, which is fed back to the NLSEF module together with the error of the phase derivative of the two phases. Finally, the compensation angle control signal is obtained and converted into a radian compensation angle to directly compensate the phase of the excitation current, and thus indirectly compensate the phase of the generator stator voltage.

[0071] like Figure 10As shown, the voltage frequency to be connected to the grid is used as the input of the EKO-ADRC controller TD module, and the stator voltage frequency is used as the input of the EKO-ADRC controller ESO module. Inside the EKO-ADRC controller, the error between the two input frequencies and the error of their derivatives are fed back to the NLSEF module to generate a control signal for the stator voltage frequency, thereby compensating for the frequency of the generator stator voltage.

[0072] Subsequently, an S-Function simulation model based on the state-space model of the AC-excited asynchronous generator was designed in Simulink. The Simulink model of the AC-excited asynchronous generator fast no-load grid connection control system is as follows: Figure 11 As shown, the effectiveness of the proposed method was verified using the RTLAB hardware-in-the-loop semi-physical simulation platform. Simulations were performed on the amplitude, phase, and frequency output of the terminal voltage under open-loop and closed-loop control, respectively, and the results are shown in the figures below. Figure 12 , 13 As shown in Figures 14 and 15, simulations were performed on the three-phase voltage errors between the generator terminal and the power grid using classic PID, classic ADRC, and an ADRC controller improved based on a Kalman state observer, respectively. The results were then verified using the RTLAB hardware-in-the-loop semi-physical simulation platform. The flowchart of the RTLAB hardware-in-the-loop experimental platform is shown below. Figure 16 As shown, the simulation results of the voltage amplitude, phase, and frequency at the oscilloscope terminal and the power grid terminal of the RTLAB hardware-in-the-loop experimental setup are as follows: Figure 17 , 18 As shown in Figure 19; the single-phase voltage tracking output of the oscilloscope terminal and the power grid terminal of the RTLAB hardware-in-the-loop experimental setup is as follows: Figure 20 As shown; the three-phase voltage error output between the oscilloscope terminal and the power grid terminal of the RTLAB hardware-in-the-loop experimental setup is as follows: Figure 21 As shown, the generator terminal voltage amplitude, phase, and frequency can quickly track the voltage amplitude, phase, and frequency of the grid segment within a short time, thus meeting the grid connection requirements.

[0073] In this implementation case, compared with traditional control methods, such as classic PID control, which takes about 0.46s to complete the grid connection between the generator and the grid, the control method in this implementation case can make the three-phase voltage error between the generator and the grid close to 0 within 0.26s, meeting the no-load grid connection conditions and achieving rapid grid connection. This minimizes the grid connection time between the generator and the grid, allowing the generator to quickly complete the grid connection and enter the next operating condition.

[0074] Example 3 This embodiment also provides a rapid no-load grid connection system for variable-speed pumped-storage units based on an improved ADRC controller, such as... Figure 22 As shown, it includes: AC excitation asynchronous generator 301: Its stator side is connected to the power grid via a grid-connected circuit breaker, and its rotor side is connected to the power end of a back-to-back converter. It is used to establish the stator voltage at the generator terminal by the rotor excitation current under no-load conditions. Signal acquisition unit 302: includes a phase sequence analyzer, a phase-locked loop and an optical encoder, used to acquire the amplitude, phase, frequency and generator rotor angular velocity signals of the grid voltage and the stator voltage at the generator terminals in real time; EKO-ADRC controller 303: It consists of an extended Kalman state observer and an ADRC controller cascaded together; the ADRC controller sequentially includes a nonlinear differential tracker, an extended state observer and a nonlinear state error feedback controller; Back-to-back converter 304: includes two voltage source converters sharing a DC bus, used to receive the rotor two-phase voltage command output by the EKO-ADRC controller, generate switching signals through 2-3 coordinate transformation and SVPWM module, and inject excitation current into the rotor of the AC excitation asynchronous generator. Grid-connected circuit breaker 305: Used to close when the amplitude, phase and frequency errors of the stator voltage at the generator terminal and the grid voltage are all less than a set threshold, thus completing zero-impact current grid connection.

[0075] In the specific implementation process, the AC-excited asynchronous generator 301 operates under no-load grid-connected conditions. The stator is disconnected from the grid via the grid-connected circuit breaker 305, and the rotor is connected to the back-to-back converter 304. The stator voltage at the generator terminal is determined by the rotor excitation current. The signal acquisition unit 302 extracts the grid voltage amplitude, phase, and frequency using a phase sequence analyzer and a phase-locked loop; the voltage / current sensor measures the generator terminal stator voltage; the photoelectric encoder measures the rotor angular velocity; and all signals are synchronously sent to the EKO-ADRC controller 303. In the TD module of the EKO-ADRC controller 303, the nonlinear differential tracker smooths the grid voltage amplitude, phase, and frequency to generate reference commands; in the ESO module, the extended state observer takes the generator terminal stator voltage and rotor control voltage as dual inputs, outputs the stator voltage estimate and its first and second derivatives, and estimates the total disturbance in real time; in the NLSEF module, the nonlinear state error feedback controller compares the reference command with the estimated value to generate the rotor two-phase voltage command. The back-to-back converter 304 receives the command, generates a switching signal through 2-3 coordinate transformation and SVPWM module, drives the dual VSCs to inject excitation current into the rotor, enabling the stator voltage at the generator terminal to quickly track the grid voltage. When the amplitude, phase, and frequency errors between the stator voltage at the generator terminal and the grid voltage are all less than the set threshold, the grid-connected circuit breaker 305 closes, completing the zero-inrush current grid connection.

[0076] The same or similar labels correspond to the same or similar parts; The terms used to describe positional relationships in the accompanying drawings are for illustrative purposes only and should not be construed as limiting this application. Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A method for rapid no-load grid connection of a variable-speed pumped-storage unit based on an improved ADRC controller, characterized in that, Includes the following steps: S1: Simplify the variable speed pumped storage unit into a single AC excitation asynchronous generator, and establish the state-space model of the AC excitation asynchronous generator under no-load grid-connected conditions. S2: Construct an ADRC controller for the state-space model, and improve the ADRC controller by introducing an extended Kalman state observer to obtain the EKO-ADRC controller; S3: Using the EKO-ADRC controller, the stator voltage amplitude, phase and frequency of the AC excitation asynchronous generator are controlled in a closed loop until the stator voltage amplitude, phase and frequency of the AC excitation asynchronous generator are consistent with the voltage amplitude, phase and frequency of the grid to be connected, so as to achieve zero inrush current grid connection.

2. The method for rapid no-load grid connection of a variable-speed pumped-storage unit based on an improved ADRC controller according to claim 1, characterized in that, In step S1, the state-space model of the AC excitation asynchronous generator under no-load grid-connected conditions is represented as follows: in, The rotor voltage vector serves as the control vector. The stator voltage vector serves as the output vector. The rotor current vector serves as the state vector. This is the voltage across the d-axis of the generator rotor. This is the q-axis voltage of the generator rotor. This refers to the d-axis voltage of the generator stator. This refers to the q-axis voltage of the generator stator. This refers to the d-axis current of the generator rotor. Let be the q-axis current of the generator rotor; A, B, C, and D are the first to fourth state space matrices, respectively; The state space matrices A, B, C, and D are represented as follows: in For generator stator resistance, For generator stator inductance, For generator rotor inductance, The generator's self-inductance coefficient. For the generator rotor speed, For grid voltage angle, generator slip speed It can be represented as: .

3. The method for rapid no-load grid connection of a variable-speed pumped storage unit based on an improved ADRC controller according to claim 1, characterized in that, In step S2, the ADRC controller is a second-order ADRC controller, which includes a TD module, an ESO module, and an NLSEF module.

4. A method for rapid no-load grid connection of a variable-speed pumped storage unit based on an improved ADRC controller, as described in claim 3, is characterized in that... The TD module adopts a single-input dual-output architecture, including a nonlinear differential tracker to avoid input jumps and smooth the input. The nonlinear differential tracker in the discrete state is represented as follows: in For target tracking signals; Let be the input signal at time k; This is the input signal at time k+1; Input signal The differential signal; for The state at the next moment; Let be the error between the input signal and the tracked signal; r be the velocity factor; h be the sampling period; and fhan(·) be the fastest control synthesis function, expressed as: in This represents the boundary of the linear region. The boundary of the switching zone is y; y is the predicted future error value. The differential form of the input signal ; For intermediate control quantities; It is a linear control variable; the Sgn(·) function is a sign function, and its return value is based on the sign of the input value. If the input value is positive, it returns 1; if the input value is 0, it returns 0; if the input value is negative, it returns -1. The ESO module adopts a dual-input, three-output port architecture and includes an ESO extended state observer. The inputs of the ESO module are the output value and the input observation value of the controlled system, and the outputs of the ESO module are the output estimate of the controlled system and its first and second derivatives. The output estimate of the controlled system and its first derivative are fed back to the nonlinear differential tracker, and the second derivative of the output estimate of the controlled system is fed back to the input of the controlled system as a compensation disturbance; the ESO extended state observer in the discrete state is represented as follows: in This represents the output value of the controlled system at time k. The input observation value of the controlled system at time k; for The differential form; The observed value of the derivative is input to the controlled system at time k; for The differential form; Let be the observed value of the total disturbance of the controlled system at time k; for The differential form; , , These are the parameters for the first, second, and third ESO observers, respectively. Here are the control gain parameters for the ESO extended state observer; fal(·) is the saturation function, expressed as: Where e is the error between the output value of the controlled system and the input signal; Indicates the nonlinear factor; The preset linear interval length; These are the nonlinear factors of the first and second fal(·) functions, respectively; The NLSEF module adopts a dual-input single-output architecture and includes a nonlinear state error feedback controller; in discrete states, the nonlinear state error feedback controller is represented as: in This refers to the error between the preset intermediate process and the observed state; The error between the preset differential signal and the observed differential state; , These are the parameters for the fourth and fifth ESO observers, respectively. , These are the nonlinear factors of the third and fourth fal(·) functions, respectively; To control the gain; For preliminary control purposes; It is the final control variable of the controlled system.

5. A method for rapid no-load grid connection of a variable-speed pumped storage unit based on an improved ADRC controller, as described in claim 1, is characterized in that... In step S2, the state-space model is linearized using a first-order Taylor expansion, and expressed as: in Let k be the state vector at time k. Let this be the state vector at time k-1; Let k be the control vector at time k. This is the control vector at time k-1; This represents the process noise at time k-1. The measurement error noise at time k; Let T' be the output vector at time k; T' be the total sampling time. Let e ​​be the identity matrix and e be the natural logarithm. F, H, M, and G are the first to fourth linearization parameter matrices, respectively; The covariance matrix of the noise is constructed as follows: Where Q is the process noise covariance matrix and matrix R is the measurement error noise covariance matrix; T represents matrix transpose; w and v represent process noise and measurement error noise, respectively; The observation estimation error of the state-space model at time k is: in Let be the observed estimate of the state vector at time k. The error between the state vector at time k and the observed estimate of the state vector; Further construct the covariance matrix of the observation estimation error of the state-space model at time k, which is expressed as: 。 6. A method for rapid no-load grid connection of a variable-speed pumped storage unit based on an improved ADRC controller, as described in claim 3, is characterized in that... In step S2, the iterative process of the extended Kalman state observer includes: Calculate the prior state estimate: in, This is the prior state estimate of the state vector at time k; This is the observed estimate of the state vector at time k-1; Calculate the prior estimate of the covariance matrix: The prior state estimate is the covariance matrix of the observation estimation error at time k. Let be the covariance matrix of the observation estimation error at time k-1; Calculate the Kalman gain value : Calculate the posterior state estimate: Finally, update the covariance matrix: 。 7. A method for rapid no-load grid connection of a variable-speed pumped storage unit based on an improved ADRC controller, as described in claim 1, is characterized in that... In step S3, the EKO-ADRC controller is used to perform closed-loop control of the stator voltage amplitude of the AC excitation asynchronous generator, including: The voltage amplitude to be connected to the grid is used as the input of the TD module in the EKO-ADRC controller, and the stator voltage amplitude is used as the input of the ESO module in the EKO-ADRC controller. Inside the EKO-ADRC controller, the error of the two input amplitude signals and the error of their derivatives are fed back to the NLSEF module to form the control signal of the generator rotor excitation current, thereby controlling the amplitude of the generator stator voltage.

8. A method for rapid no-load grid connection of a variable-speed pumped storage unit based on an improved ADRC controller, as described in claim 1, is characterized in that... In step S3, the EKO-ADRC controller is used to perform closed-loop control of the stator voltage phase of the AC excitation asynchronous generator, including: The voltage phase to be connected to the grid is used as the input of the EKO-ADRC controller TD module, and the stator voltage phase is used as the input of the EKO-ADRC controller ESO module. In the EKO-ADRC controller, the phase difference between the two phases is calculated to form a phase difference, which is fed back to the NLSEF module together with the error of the phase derivative of the two phases. Finally, the compensation angle control signal is obtained and converted into a radian compensation angle to directly compensate the phase of the excitation current, and thus indirectly compensate the phase of the generator stator voltage.

9. A method for rapid no-load grid connection of a variable-speed pumped storage unit based on an improved ADRC controller, as described in claim 1, is characterized in that... In step S3, the stator voltage frequency of the AC excitation asynchronous generator is controlled in a closed loop using the EKO-ADRC controller, including: The voltage frequency to be connected to the grid is used as the input of the TD module of the EKO-ADRC controller, and the stator voltage frequency is used as the input of the ESO module of the EKO-ADRC controller. Inside the EKO-ADRC controller, the error between the two input frequencies and the differential error are fed back to the NLSEF module to generate a control signal for the stator voltage frequency, thereby compensating for the frequency of the generator stator voltage.

10. A method for rapid no-load grid connection of a variable-speed pumped storage unit based on an improved ADRC controller, as described in any one of claims 1 to 9, characterized in that, An S-Function simulation model based on the state-space model of the AC-excited asynchronous generator was designed in Simulink, and the effectiveness of the method was verified by the RTLAB hardware-in-the-loop semi-physical simulation platform.

Images