Automatic optimization method for AES multi-stage flexible adaptive control parameters of variable-speed pumped storage units

Abstract

The present invention discloses a method for optimizing the multi-stage flexible adaptive control parameters of a variable-speed pumped-storage (AES) system, which at least solves the problem of optimizing the parameters of an AES controller. The method comprises: reading a working stage signal of a variable-speed pumped-storage unit to determine the current working stage of an AES-MSC controller; calling an AES-MSC controller parameter optimization algorithm to iteratively optimize the optimization objective function of the current working stage to optimize the parameters of the AES-MSC controller; calling an offline simulation model of the AES-MSC controller for the current working stage based on the finite state machine principle according to the current working stage of the AES-MSC controller; outputting the parameter optimization results of the AES-MSC controller when the maximum number of iterations is reached; substituting the parameter optimization results into the offline simulation model for verification; if the verification results are not as expected, performing iterative optimization again; if the verification results meet the expectations, saving the parameter optimization results of the AES-MSC controller.

Description

Technical Field

[0001] The present invention relates to the technical field of variable speed pumped storage AC excitation system control, and more specifically to a method for automatically optimizing multi-stage parameters of an AC excitation system controller under a variable speed pumped storage unit power generation condition. Background Art

[0002] With the increasing penetration of distributed power sources, fixed-speed pumped storage units can no longer meet the flexible regulation requirements of new power systems. Flexible control of variable-speed pumped storage (VSPS) units has become an urgent need. However, there is currently no reasonable and accurate controller parameter optimization method for the AC excitation system (AES), which is a key device for flexible control of variable-speed units.

[0003] AES controller parameter optimization is a crucial step in the design of VSPS unit excitation control systems. Its goal is to find an optimal or near-optimal set of controller parameters that achieves the desired performance goals for the excitation closed-loop control system. These goals include system stability, speed, accuracy, and robustness.

[0004] During the multi-stage flexible adaptive control of VSPS units, the physical characteristics and control objectives of the units vary across different stages, causing the AES control model and controller parameters to change as the units change stages. Therefore, the AES controller has the physical characteristics of adaptive control parameters. Optimizing these adaptive control parameters is a crucial step in designing the AES multi-stage flexible adaptive control strategy. Its rationality and accuracy are prerequisites for studying the transient processes of VSPS units.

[0005] On the other hand, the AES is a multivariable, strongly coupled, high-order nonlinear system, and its control performance is closely related to the operating parameters of the generator motor. Over time, the generator motor's electrical parameters change. If the AES controller parameters remain unchanged, its control performance will degrade or even malfunction.

[0006] Optimizing AES controller parameters is an iterative process requiring multiple adjustments and verifications. In practical applications, the optimization process can be complicated by factors such as the nonlinearity, uncertainty, and external interference of VSPS units. Therefore, selecting appropriate optimization strategies and algorithms is crucial for achieving satisfactory control results. Furthermore, with the advancement of computing power and optimization algorithms, the application of control parameter optimization in engineering is becoming increasingly widespread and effective.

[0007] In order to solve the above problems, the present invention proposes an automatic optimization method for multi-stage flexible adaptive control parameters of a variable speed pumped storage unit AES. Summary of the Invention

[0008] In a first aspect, the present invention proposes a method for dividing different stages of the power generation process and the pumping process of a VSPS unit.

[0009] Secondly, the present invention proposes a multi-stage objective optimization function for the AES controller applicable to the power generation process of the VSPS unit, which truly reflects the control objectives of the AES controller at different stages of the power generation process.

[0010] Thirdly, the present invention proposes a mathematical model and a control model applicable to the multi-stage AES controller of the VSPS unit power generation process, which finely restores the physical characteristics of the AES controller at different stages of the power generation process.

[0011] Fourthly, the AES-MSC controller parameter optimization method proposed in the present invention realizes the intelligent optimization function of the AES-MSC controller parameters.

[0012] In a fifth aspect, the present invention proposes a multi-stage flexible adaptive control parameter optimization method for an AES-MSC controller based on a finite state machine, thereby realizing an automatic parameter optimization function of the AES-MSC controller.

[0013] A variable-speed pumped-storage AES multi-stage flexible adaptive control parameter optimization method comprises the following steps: 1) reading a variable-speed pumped-storage unit working stage signal to determine the current working stage of an AES-MSC controller; 2) calling an AES-MSC controller parameter optimization algorithm to iteratively optimize an optimization objective function of the current working stage, so that the AES-MSC controller parameters reach the optimal value; 3) calling an offline simulation model of the current working stage of the AES-MSC controller based on a finite state machine principle according to the current working stage of the AES-MSC controller; 4) outputting the parameter optimization result of the AES-MSC controller when a maximum number of iterations is reached; 5) substituting the parameter optimization result into the offline simulation model for verification; 6) if the verification result is less than expected, jumping to step 2) and performing secondary optimization; 7) if the verification result is as expected, saving the AES-MSC controller parameter optimization result; 8) if the working stage of the variable-speed pumped-storage unit changes and is not a shutdown stage, jumping to step 1) and starting parameter optimization for the next working stage; 9) if the current working stage of the variable-speed pumped-storage unit is a shutdown stage, exiting the AES-MSC controller parameter automatic optimization program. The AES-MSC controller uses a PI structure. The AES-MSC controller parameters include the outer and inner loop control coefficients for the d-axis and q-axis at different stages of power generation.

[0014] The present invention has at least the following advantages or beneficial effects: automatic optimization of the multi-stage flexible adaptive control parameters of the AES controller in the VSPS unit power generation process based on a finite state machine can realize automatic optimization of the controller's multi-stage parameters according to the VSPS unit status signal; controller parameter optimization based on an intelligent algorithm improves the speed and accuracy of AES controller parameter optimization; and the dynamic and static performance of the AES multi-stage flexible control is used as the target optimization function of the intelligent algorithm to effectively ensure the dynamic and static performance of the AES controller at different stages. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a schematic diagram of the VSPS unit structure.

[0016] Figure 2 This is a schematic diagram of the multi-stage conversion process of the VSPS unit power generation process.

[0017] Figure 3 This is a schematic diagram of the multi-stage conversion of the VSPS unit's pumping process.

[0018] Figure 4 This is a control model block diagram of an AES-MSC controller in the no-load pressure building stage provided by an embodiment of the present invention.

[0019] Figure 5 This is a control model block diagram of an AES-MSC controller provided by an embodiment of the present invention during the flexible grid connection stage.

[0020] Figure 6 This is a control model block diagram of the AES-MSC controller provided by one embodiment of the present invention in the steady-state power generation stage.

[0021] Figure 7 This is a control model block diagram of the AES-MSC controller in the electrical braking stage provided by one embodiment of the present invention.

[0022] Figure 8 The present invention provides a flowchart of a multi-stage parameter automatic optimization method of an AES-MSC controller based on a finite state machine according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] Figure 1 The VSPS unit structure is shown. Its structure is extremely complex, including a monitoring system, coordinated controller, generator motor, AC excitation system (AES), speed governor and control system, reversible pump-turbine, upper and lower reservoirs, water diversion system, and surge tank. The VSPS unit is a coupled hydraulic, mechanical, electrical, and control system, characterized by multiple timescales, multiple stages, and strong coupling.

[0024] The AC excitation system (AES) consists of the generator-side converter (MSC) and its controller, and the grid-side converter (GSC) and its controller. The AES-GSC controller controls the capacitor DC voltage. The AES-GSC controller parameters remain constant across different phases, while the AES-MSC controller parameters vary across different phases, demonstrating multi-phase adaptive physical characteristics. The AES-GSC controller is also known as the GSC controller. The AES-MSC controller is also known as the MSC controller. The AES controller includes both the GSC and MSC controllers.

[0025] Since the parameter optimization methods of the MSC controller and the GSC controller are similar, this embodiment only introduces the automatic optimization method of the MSC controller parameters, and does not elaborate on the optimization method of the GSC controller parameters.

[0026] like Figure 2 As shown in Figure 1, the VSPS unit power generation process includes multiple stages: shutdown, power generation startup, no-load voltage buildup, flexible grid connection, load ramp-up, steady-state power generation, load shedding, and electrical braking. The MSC does not operate during the power generation startup and shutdown stages.

[0027] like Figure 3 As shown in Figure 1, the VSPS unit pumping process consists of multiple stages: shutdown, pumping start, electric synchronization, steady-state pumping, speed reduction, and electrical braking. When SFC is used for pumping start, the GSC begins charging the DC bus, but the MSC is inoperative. When auto-start is used for pumping start, both the GSC and MSC are operational. The MSC remains operational during all stages except shutdown and SFC start.

[0028] Based on the working principle of multi-stage flexible adaptive control of VSPS units, this paper proposes a method for automatically optimizing the multi-stage flexible adaptive control parameters of the MSC controller during the power generation process of a VSPS unit. The following is a detailed introduction to the MSC controller parameter optimization process.

[0029] (1) Optimize the objective function

[0030] The performance indicators ITAE (Integral Time Multiplied by Absolute Error) and THD (Voltage Harmonic Distortion) are selected as indicators for measuring MSC controller performance. This invention uses the ITAE (Integral Time Multiplied by Absolute Error) and THD (Voltage Harmonic Distortion) performance indicators based on the time sum of the errors and employs a weighted approach to optimize multiple performance objectives. Because the control objectives of the MSC controller vary at different stages, their objective optimization functions also differ. The objective optimization functions for each stage of the MSC controller are described below.

[0031] The control goal of the MSC controller in the no-load voltage building stage is to quickly build up the stator voltage, and its optimization objective function is:

[0032]

[0033] Where e0(t) is the error between the stator no-load voltage and the stator rated voltage.

[0034] The goal of the MSC controller in the flexible grid-connected stage is to make the inrush current of the generator motor close to zero at the moment of grid connection, and to minimize the impact of grid connection on the generator motor and the power grid. Therefore, the optimization objective function in the flexible grid-connected stage is:

[0035]

[0036] Where e1(t) is the error between the stator voltage and the grid voltage amplitude, e2(t) is the error between the stator voltage and the grid voltage phase, and a and b are weight coefficients. The weight coefficients are set according to actual needs. In one embodiment, a = 0.5 and b = 0.5.

[0037] The goal of the MSC controller in the steady-state power generation (load increase, load rejection) stage is to ensure that the amplitude and waveform of the stator output current of the generator motor meet the grid requirements and can quickly track the changes in the power command to achieve the purpose of peak and frequency regulation. Therefore, the optimization objective function in the steady-state power generation stage is:

[0038]

[0039] Where i sn is the current harmonic, n is the harmonic component, e p (t) is the error between the active power and the active power command, and a and b are weight coefficients. The weight coefficients are adjusted according to actual needs. In one embodiment, a=0.5 and b=0.5.

[0040] The goal of the MSC controller in the electrical braking stage is to generate sufficient braking torque to brake the generator motor rotor, so the objective optimization function in the electrical braking stage is:

[0041]

[0042] Where, e n (t) is the error between the rotational speed and zero.

[0043] (2) Mathematical model and control model of MSC controller:

[0044] Using the stator flux oriented vector control strategy, the rotor voltage of the generator motor can be expressed as:

[0045]

[0046] u rd is the rotor d-axis voltage component, u rq is the rotor q-axis voltage component, u′rd is the rotor d-axis voltage feedforward compensation term, Δu rd is the rotor d-axis voltage decoupling compensation term, u′ rq is the rotor q-axis voltage feedforward compensation term, Δu rq is the rotor q-axis voltage decoupling compensation term, R r Rotor resistance, i rd is the rotor d-axis current component, α is the motor leakage coefficient, l r Rotor inductance, p is the differential operator, ω s Slip angular frequency, i rq is the rotor q-axis current component, l m Stator and rotor mutual inductance, L s is the stator inductance, ψ sd Stator d-axis flux.

[0047] According to the generator motor rotor voltage formula, the AES-MSC control system can be designed as a double closed-loop architecture with feedforward compensation. The controller is a PI structure, and the voltage used for decoupling compensation is Δu r .

[0048] During the no-load voltage building phase, the stator d-axis and q-axis current components i sd =i sq =0, the rotor d-axis and q-axis currents can be expressed as:

[0049]

[0050] L m is the stator-rotor mutual inductance, ψ1 is the stator flux. According to the rotor current formula, the control model of the AES-MSC controller in the no-load voltage buildup stage is as follows: Figure 4 shown. Figure 4 middle, is the three-phase voltage of the power grid; is the rotor three-phase control voltage.

[0051] In the flexible grid connection stage, if the rotor d-axis current is controlled so that In theory, the amplitude of the stator open-circuit voltage can be guaranteed to be consistent with the grid voltage. In the actual grid connection process, due to the influence of voltage and current sampling errors, motor parameter changes and converter response characteristics, the amplitude of the terminal voltage cannot accurately track the grid voltage, and there is still a certain amplitude difference between the two. In order to solve the impact of parameter changes and enhance the robustness of the system, the rotor d-axis current i rd Add a compensation term Δi rd :

[0052]

[0053] Rotor d-axis current i during flexible grid connection rdIt consists of two parts: generalized excitation component and compensation amount:

[0054]

[0055] U g is the grid voltage, U s is the stator voltage, ω g is the grid voltage angular frequency, k p1 is the proportional coefficient, s is the differential operator, k i1 is the integral coefficient. According to the above rotor current formula, the control model of the AES-MSC controller in the flexible grid connection stage is as follows: Figure 5 shown. Figure 5 middle, is the three-phase voltage of the power grid, is the stator three-phase voltage, is the rotor three-phase control voltage, the rotor d-axis current i rd The compensation term Δi rd Also serves as a reference value for the rotor d-axis current is the rotor q-axis reference current.

[0056] In the steady-state power generation phase, the VSPS unit can automatically participate in system frequency regulation and be controlled by the AGC or the upper-level dispatching system, directly and decoupledly controlling the unit's active and reactive power. The unit operates in this phase for most of the time, and the generator motor output power is:

[0057]

[0058] P and Q are active power and reactive power respectively, u sq is the q-axis voltage component of the stator voltage, i sq is the q-axis current component of the stator current, i rq is the q-axis current component of the rotor current, i sd is the d-axis current component of the stator current, U g is the grid voltage, L m is the stator-rotor mutual inductance, L s is the stator inductance, ω g is the grid voltage angular frequency. According to the generator motor output power formula, the control model of the AES-MSC controller in the steady-state power generation stage is as follows: Figure 6 shown. Figure 6 In, P * , Q * These are the set values / command values ​​for active power and reactive power respectively.

[0059] During the electrical braking phase, the stator winding of the generator motor is short-circuited. Under the action of the excitation voltage (current), the generator motor will generate a braking torque. Due to the short circuit of the stator winding during electrical braking, the stator d-axis and q-axis voltage components u sd=u sq =0, according to the stator flux equation, the stator current can be expressed as:

[0060]

[0061] ψ sd Stator d-axis flux, L m is the stator-rotor mutual inductance, L s is the stator inductance, i rq is the q-axis current component of the rotor current, i rd is the d-axis current component of the rotor current.

[0062] Substituting the stator current equation into the stator voltage equation and torque equation, we can obtain:

[0063]

[0064] T s is the stator winding time constant, T s =L s / R s , r S is the stator resistance, T e Electromagnetic torque, p is the differential operator, n p is the number of motor pole pairs, ψ sd Stator d-axis flux, L m is the stator-rotor mutual inductance, L s is the stator inductance, i rd is the d-axis current component of the rotor current, i rq is the q-axis current component of the rotor current.

[0065] Control the rotor d-axis excitation current i rd The stator flux can be controlled and the d-axis excitation current reference value Setting it to a constant of zero can keep the stator flux constant. At this time, controlling the q-axis excitation current can control the electromagnetic torque and achieve rapid electrical braking of the unit. The control model of the AES-MSC controller in the electrical braking stage is as follows: Figure 7 shown.

[0066] (3) Selecting the controller structure: Based on the control model, system characteristics and design requirements of the AES-MSC controller, the control system is designed as a double closed-loop control architecture with feedforward, and the controller selects the PI structure.

[0067] (4) Parameterize the AES-MSC controller to prepare for the subsequent parameter optimization process. Taking the power generation process as an example, Figures 4 to 7 As shown in the table below, the control parameter optimization objects of the AES-MSC controller at different stages under power generation conditions are as follows:

[0068]

[0069] (5) Selection of AES-MSC controller parameter optimization algorithm: Grey wolf algorithm (GWO), particle swarm algorithm, fruit fly algorithm, etc. can be used to optimize and adjust the control parameters.

[0070] The steps of the multi-stage flexible adaptive parameter automatic optimization process of the AES-MSC controller based on the finite state machine are as follows:

[0071] 1) Initialize the system.

[0072] 2) Read the parameters of the generator motor, including the resistance, inductance, mutual inductance, etc. of the motor stator and rotor.

[0073] 3) Read the VSPS unit working stage signal to determine the current working stage of the AES-MSC.

[0074] 4) Call the AES-MSC controller parameter optimization algorithm.

[0075] 5) Based on the current AES-MSC operating phase, an offline simulation model of the AES-MSC controller is invoked based on the principles of a finite state machine. This offline simulation model is a simulation model of the VSPS unit. The method of the present invention does not require verification on a physical unit; verification on this simulation model is sufficient. This simulation model utilizes existing technology and will not be described in detail in this invention.

[0076] 6) When the maximum number of iterations is reached, the AES-MSC controller parameter optimization results are output.

[0077] 7) Substitute the parameter optimization results into the offline simulation model for verification.

[0078] 8) If the simulation results are not as expected, jump to step 4) for secondary optimization.

[0079] 9) The simulation results meet expectations, and the AES-MSC physical controller parameter optimization results are saved.

[0080] 10) If the unit's operating phase changes and is not a shutdown phase, jump to step 3) to start parameter optimization for the next operating phase.

[0081] 11) If the unit is currently operating in the shutdown phase, the AES-MSC controller parameter automatic optimization program will be exited.

Claims

1. A variable speed pumped storage AES multi-stage flexible adaptive control parameter optimization method, characterized in that: include: 1) Read the working stage signal of the variable-speed pumped storage unit to determine the current working stage of the AES-MSC controller; 2) Call the AES-MSC controller parameter optimization algorithm to iteratively optimize the optimization objective function of the current working stage to optimize the AES-MSC controller parameters; 3) According to the current working stage of the AES-MSC controller, the offline simulation model of the current working stage of the AES-MSC controller is called based on the finite state machine principle; 4) When the maximum number of iterations is reached, the parameter optimization results of the AES-MSC controller are output; 5) Substitute the parameter optimization results into the offline simulation model for verification; 6) If the verification result is not as expected, jump to step 2) for secondary optimization; 7) Verify that the results meet expectations and save the AES-MSC controller parameter optimization results; 8) If the variable-speed pumped storage unit's operating phase changes and is not in the shutdown phase, jump to step 1) and start parameter optimization for the next operating phase; 9) If the variable-speed pumped storage unit is currently operating in the shutdown phase, the AES-MSC controller parameter automatic optimization program will be exited; Among them, the performance index ITAE based on the time of error sum and the voltage harmonic index THD are adopted, and the weighted method is used to achieve the optimization of multiple performance objectives; The goal of the AES-MSC controller in the steady-state power generation stage is to ensure that the amplitude and waveform of the stator output current of the generator motor meet the grid requirements and can quickly track the changes in the power command to achieve the purpose of peak and frequency regulation. The optimization objective function in the steady-state power generation stage is: Where, is the current harmonic, n is the harmonic component, is the error between active power and active power instruction, a and b are weight coefficients, is the fundamental wave of current.

2. The method according to claim 1, characterized in that The power generation process of a variable-speed pumped storage unit includes: shutdown, power generation startup, no-load pressure buildup, flexible grid connection, load increase, steady-state power generation, load rejection and electrical braking. The MSC does not work during the power generation startup and shutdown stages.

3. The method according to claim 1, characterized in that The pumping process of a variable-speed pumped-storage unit includes: shutdown, pumping start, electric synchronization, steady-state pumping, speed reduction, and electric braking. When SFC is used for pumping start, the GSC begins to charge the DC bus, but the MSC does not work. When self-start is used for pumping start, both the GSC and MSC are in operation. The MSC is in operation in all stages except shutdown and SFC start.

4. The method according to claim 1, wherein The control goal of the AES-MSC controller in the no-load voltage building stage is to quickly build up the stator voltage, and its optimization objective function is: Where, It is the error between the stator no-load voltage and the stator rated voltage.

5. The method according to claim 1, wherein The goal of the AES-MSC controller in the flexible grid connection stage is to reduce the inrush current of the generator motor to zero at the moment of grid connection, thereby reducing the impact of grid connection on the generator motor and the power grid. The optimization objective function in the flexible grid connection stage is: Where, is the error between the stator voltage and the grid voltage amplitude, is the phase error between the stator voltage and the grid voltage, a and b are weight coefficients.

6. The method according to claim 1, characterized in that The goal of the AES-MSC controller in the electrical braking stage is to generate braking torque to brake the generator motor rotor. The objective optimization function of the electrical braking stage is: Where, is the error between the rotational speed and zero.

7. The method according to claim 1, characterized in that The AES-MSC controller selects the PI structure.

8. The method according to claim 7, characterized in that The AES-MSC controller parameters include the outer and inner loop control coefficients of the d-axis and q-axis at different stages of power generation conditions.

Images