Variable speed pumped storage aes multi-stage flexible adaptive control parameter automatic online identification method

Abstract

The application discloses a variable-speed pumped storage AES multi-stage flexible adaptive control parameter automatic online identification method, which comprises the following steps: S1, reading a unit working stage signal to determine the current working stage of the MSC; S2, based on the current working stage of the MSC, a three-phase pseudo-random sequence is used to stimulate the secondary measurement signal of the MSC controller hardware-in-the-loop; S3, a particle swarm algorithm is used to identify the proportional and integral parameters of the MSC controller; S4, when the unit working stage changes and is not in the shutdown stage, jumping to S1 to start the parameter identification of the next working stage; and S5, when the unit working stage is in the shutdown stage, the identification program is exited. The three-phase pseudo-random sequence is used to stimulate the secondary measurement signal of the MSC controller, the operation on the intermediate variable of the controller is avoided, and the engineering practicability is enhanced; and the control parameter identification method based on the simultaneous stimulation of the inner loop and the outer loop of the controller effectively solves the step-by-step identification problem of the cascade PI controller.

Description

Technical Field

[0001] The present invention relates to the technical field of variable speed pumped storage AC excitation system control, and more particularly to an online automatic identification method for multi-stage flexible adaptive control parameters of a variable speed pumped storage AC excitation controller. Background Art

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

[0003] During the VSPS multi-stage flexible control process, the varying hydraulic heads and geographical conditions of different units result in varying electrical parameters and AES controller parameters. Furthermore, due to the varying physical characteristics and control objectives of the AES at different stages within a single unit, the AES control model and controller parameters change with each stage. Given the physical characteristics of the AES controller, which exhibits adaptive control parameters, adaptive control parameter identification is a crucial step in designing the AES multi-stage flexible adaptive control strategy. Its rationality and accuracy are prerequisites for both AES controller design and detailed unit simulation.

[0004] On the other hand, the AES is a multivariable, strongly coupled, high-order nonlinear system, and its control performance is closely related to the generator motor's operating parameters. Long-term operation or aging of the generator motor can cause changes in its electrical parameters. Without timely monitoring of AES controller parameter changes, using outdated control parameters will make it impossible to accurately simulate the unit's transient processes. Therefore, automatic identification of AES controller parameters is a prerequisite for achieving precise unit simulation research.

[0005] AES controller parameter identification is the process of determining the AES controller model parameters through experiments and data analysis in the design of VSPS unit automatic control systems. These parameters are used in the design of the AES controller to ensure that the system operates according to the predetermined performance indicators.

[0006] The key to AES controller parameter identification lies in selecting an appropriate model structure and identification algorithm, as well as ensuring data quality. In practical applications, AES is subject to nonlinearity, multivariable variables, time-varying characteristics, uncertainty, and external interference, all of which complicate identification. Therefore, AES controller parameter identification requires online automatic adjustment and optimization based on the specific conditions of the VSPS unit. Summary of the Invention

[0007] To solve the above problems, the present invention applies for a method for automatic online identification of multi-stage flexible adaptive control parameters of the AES controller of a variable-speed pumped storage unit. The method is not only suitable for the AES controller parameter identification of units with different heads and geographical conditions, but also suitable for the online automatic identification of flexible adaptive control parameters of a single unit AES at different stages.

[0008] In the first aspect, the present invention proposes an online identification method based on superimposing an excitation signal on the secondary measurement signal side, and the inner loop and outer loop of the AES controller simultaneously superimpose a three-phase pseudo-random sequence excitation.

[0009] In a second aspect, the present invention proposes a method for dividing different stages of the power generation process and the pumping process of a VSPS unit.

[0010] In a third aspect, the present invention proposes an AES-MSC controller parameter identification method based on the least squares method to achieve intelligent identification of AES-MSC controller parameters.

[0011] In a fourth aspect, the present invention proposes a multi-stage flexible adaptive control parameter identification method for an AES-MSC controller based on a finite state machine, thereby realizing an automatic parameter identification function of an AES-MSC controller.

[0012] In a fifth aspect, the present invention proposes a multi-stage flexible adaptive control parameter identification method for an AES-MSC controller based on hardware-in-the-loop, thereby realizing the identification function of the AES-MSC controller parameters under hardware-in-the-loop.

[0013] An automatic online identification method for multi-stage flexible adaptive control parameters of a variable-speed pumped storage system (AES) is disclosed. Except for the MSC controller, which is a physical device, all other parameters are implemented by a real-time simulator. The method comprises: S1, reading the unit's working stage signal to determine the current working stage of the MSC; S2, based on the current working stage of the MSC, using a three-phase pseudo-random sequence to stimulate the secondary measurement signal of the MSC controller's hardware-in-the-loop; S3, using a particle swarm algorithm to identify the proportional and integral parameters of the MSC controller; S4, if the unit's working stage changes and is not a shutdown stage, jumping to S1 and starting parameter identification for the next working stage; S5, if the unit's working stage is a shutdown stage, exiting the identification program.

[0014] After identifying the proportional and integral parameters of the MSC controller, the communication between the MSC controller and the simulator is interrupted and the MSC controller is shielded. According to the unit working stage signal, the hardware-in-the-loop simulation model of the MSC current working stage is called based on the finite state machine principle to verify the parameter identification results. If the simulation result is wrong, jump to S3 for secondary identification. If the simulation result is correct, save the MSC controller parameter identification results.

[0015] The present invention has at least the following advantages or beneficial effects: it is not only suitable for the AES controller parameter identification of units with different water heads and geographical conditions, but also suitable for the automatic identification of flexible adaptive control parameters of AES of a single unit at different stages, and has good compatibility; the superimposed three-phase pseudo-random sequence excitation method based on the secondary side avoids the operation of the intermediate variables of the controller, thereby enhancing the engineering practicality; the control parameter identification method based on the simultaneous excitation of the inner loop and the outer loop of the AES controller effectively solves the step-by-step identification problem of the cascade PI controller; the AES multi-stage flexible control parameter automatic identification method based on the finite state machine effectively solves the problem of segmented identification; the online identification of control parameters based on hardware-in-the-loop avoids the disturbance operation of the unit and improves the safety of parameter identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a schematic diagram of the VSPS unit structure.

[0017] Figure 2 This is a schematic diagram of the multi-stage conversion process of the VSPS unit power generation process.

[0018] Figure 3 This is a schematic diagram of the multi-stage conversion of the VSPS unit's pumping process.

[0019] Figure 4 This is the block diagram of the AES-MSC control model for the VSPS unit during the steady-state power generation stage.

[0020] Figure 5 This is a schematic diagram of hardware-in-the-loop simulation for automatic identification of AES-MSC multi-stage flexible adaptive control parameters for VSPS units.

[0021] Figure 6 Schematic diagram of a three-phase pseudo-random sequence excitation method for an AES-MSC controller of a VSPS unit provided by an embodiment of the present invention.

[0022] Figure 7 The present invention is a flow chart of automatic multi-stage parameter identification of the AES-MSC controller of the VSPS unit based on the finite state machine. DETAILED DESCRIPTION

[0023] Figure 1 The structure of the VSPS unit is shown. The unit is extremely complex and includes a monitoring system, a coordinated controller, a generator motor, an AC excitation system (AES), a speed regulator and its control system, a reversible pump-turbine, an upper reservoir, a lower reservoir, a water diversion system and a surge tank. The VSPS unit is a water-machine-electrical-control coupled system with multi-timescale, multi-stage and strongly coupled physical characteristics.

[0024] The AC excitation system (AES) consists of the generator-side converter (AES-MSC) and its controller, and the grid-side converter (AES-GSC) and its controller. The AES-GSC controller targets the capacitor DC voltage, so its parameters remain constant across different phases. The VSPS unit's control objectives vary across different phases, so the AES-MSC controller parameters also vary, demonstrating multi-stage adaptive physical characteristics.

[0025] Since the parameter identification methods of the AES-MSC controller and the AES-GSC controller are similar, this embodiment introduces the parameter identification method of the AES-MSC controller, and the parameter identification method of the AES-GSC controller is not repeated.

[0026] like Figure 2 As shown in Figure 1, the VSPS unit power generation process includes multiple phases: shutdown, power generation startup, no-load voltage buildup, flexible grid connection, output ramp-up, steady-state power generation, load shedding, and electrical braking. During the power generation startup phase, the AES-GSC begins charging the DC bus, but the AES-MSC is inoperative. Except for shutdown and power generation startup, the AES-MSC remains operational throughout all other phases.

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

[0028] According to the working principle of multi-stage flexible adaptive control of VSPS unit, the present invention proposes an online automatic identification method of multi-stage flexible adaptive control parameters of AES-MSC controller based on finite state machine. The method is described in detail below. According to the multi-stage working principle of AES-MSC controller of VSPS unit, in the present invention, AES-MSC controller adopts a typical double closed-loop PI control architecture with feedforward. The identification target of AES-MSC is the controller parameters of each stage. Taking the steady-state power generation stage as an example, the controller parameters to be identified are k p1 ~k p4 and k i1 ~k i4 , a total of 8 parameters, Figure 4 This is the control model block diagram of the AES-MSC steady-state power generation stage of the VSPS unit. The coefficients of the d-axis outer and inner loop PI controllers in the steady-state power generation stage are: p1 ~k p2 , k i1~k i2 ;Q-axis outer and inner loop PI controller coefficients: k p3 ~k p4 , k i3 ~k i4 . Figure 4 Where P and Q are active power and reactive power respectively, P * , Q * These are the set values / command values ​​for active power and reactive power respectively.

[0029] In the laboratory, you can build a Figure 5 The variable-speed pumped storage unit hardware-in-the-loop experimental platform shown performs online automatic identification of the multi-stage flexible adaptive control parameters of the AC excitation controller. For example, RT-LAB can be used to establish a hardware-in-the-loop experimental platform. The hardware-in-the-loop simulation test platform has a real-time scheduling platform, a real-time simulator (such as OP5700), a central controller and a host computer. Optical fiber communication is used between the machine-side converter controller and the simulator, and the communication protocol is Aurora. The present invention performs online identification of the AES-MSC multi-stage flexible adaptive control parameters. Except for the machine-side converter controller (AES-MSC controller) of the VSPS unit, which is a physical object, all others (including the grid-side converter controller) are implemented by a real-time simulator.

[0030] Considering that the AES-MSC controller takes action because the secondary side measurement circuit detects the change of the primary side electrical quantity, the controller response is stimulated only after the measurement signal is sent to the AES-MSC controller. Figure 6 As shown in the figure, the present invention adopts a method of identifying the parameters of the AES-MSC controller by superimposing a three-phase pseudo-random sequence signal on the secondary side measurement signal. This method can be directly applied to actual engineering. m1 and m2 are two disturbances in the rotating coordinate system (three-phase pseudo-random sequence signals), which are transformed into disturbances in the three-phase stationary coordinate system (stator excitation current i sf , rotor excitation current i rf ), the perturbations of the inner and outer loops are the same. Figure 6 in,i s is the stator current, i f is the rotor current, i s ′ is the stator excitation current i sf and stator current i s The value input to the AES-MSC controller after superposition, i′ r is the rotor excitation current i rf and rotor current i f The value input to the AES-MSC controller after superposition, u sq is the q-axis voltage component of the stator voltage, u sd is the d-axis voltage component of the stator voltage, is the rotor q-axis reference current, Δu rq is the rotor q-axis voltage compensation term, The simulation results are shown in Figure 2. It is necessary to identify the proportional and integral (kp, ki) of the PI controller in this model.

[0031] The parameters of the AES-MSC controller are identified by superimposing the secondary side measurement signals of the inner and outer loops of the AES-MSC controller on three-phase pseudo-random sequence signals, avoiding the problem of step-by-step identification caused by the cascade of the inner and outer loop PI controllers.

[0032] In the simulation environment, the three-phase pseudo-random sequence signals of the inner and outer loops of the AES-MSC controller are both realized by a simulator (such as OP5700).

[0033] Considering that the response time of the outer loop PI control of the AES-MSC controller is several milliseconds to tens of milliseconds, and the response time of the inner loop PI control is hundreds of microseconds to several milliseconds, and to ensure that the signal meets the continuous excitation condition, the frequency of the three-phase pseudo-random sequence signal can be 800-1200Hz.

[0034] Considering that the actual controller output limit design value cannot be obtained, in order to avoid the saturation nonlinearity of the controller output caused by signal excitation, the present invention adopts a method of determining the excitation signal size by 4% to 8% of the maximum value of the controller feedback input during the startup phase of the VSPS unit.

[0035] The least square method is used to identify the parameters of the AES-MSC controller. The multi-stage flexible adaptive parameter automatic identification process of the AES-MSC controller based on the finite state machine is as follows: Figure 7 As shown, the working phase signal of the VSPS unit is sent to this program by the coordination controller, and the electrical parameters of the VSPS unit itself are sent to this program by the monitoring system.

[0036] The steps of the automatic identification process of the multi-stage flexible adaptive control parameters of the AES-MSC controller based on the finite state machine are as follows:

[0037] 1. Initialize the system.

[0038] 2. Read the parameters of the generator motor, including the resistance, inductance, mutual inductance, etc. of the motor stator and rotor.

[0039] 3. Read the unit working stage signal to determine the current working stage of AES-MSC.

[0040] 4. Use a three-phase pseudo-random sequence to stimulate the secondary measurement signal of the AES-MSC physical controller hardware in the loop. That is, apply a disturbance (three-phase pseudo-random sequence signal) to the identification model and obtain the simulation results. The three-phase pseudo-random sequence signal is superimposed with the stator current measurement signal and the rotor current measurement signal to obtain the stator current superposition signal i s ′ and rotor current superposition signal i′ r , change i s ′、i′ r The actual stator current and rotor current input signals of the AES-MSC controller are used to obtain the simulation results of the actual model. The PI controller parameters (kp, ki) in the identification model (AES-MSC controller) are used as the variables to be identified.

[0041] 5. Call the particle swarm algorithm to identify the kp and ki of the AES-MSC controller. Initialize the particle velocity and position, calculate the fitness function for each particle, calculate and update the local and global optimal values ​​of the particles, update the particle position, velocity, and inertia function, calculate the population fitness variance, determine whether the algorithm has premature convergence, use the logistic mapping equation to perform chaotic processing on prematurely converged particles, update the particle position and velocity according to the fitness ranking, and continue the iterative search until the maximum number of iterations is reached. The iteration is terminated and the optimal fitness function value is output.

[0042] 6. Interrupt the communication between the physical controller and the simulator, and shield the AES-MSC physical controller.

[0043] 7. According to the working phase signal of the VSPS unit, the hardware-in-the-loop simulation model of the current working phase of AES-MSC (including the AES-MSC controller) is called based on the finite state machine principle to verify the parameter identification results.

[0044] 8. If the simulation result is wrong, jump to step 5 for secondary identification.

[0045] 9. The simulation results are correct, and the AES-MSC physical controller parameter identification results are saved.

[0046] 10. If the unit's operating phase changes and is not a shutdown phase, jump to step 3 and start parameter identification for the next operating phase.

[0047] 11. The unit's working phase is the shutdown phase, and the AES-MSC controller parameter automatic identification program is exited.

Claims

1. A method for automatic online identification of multi-stage flexible adaptive control parameters of variable speed pumped storage AES, characterized by: Except for the MSC controller, which is a physical device, all other components of the unit are implemented by a real-time simulator. The MSC controller adopts a double-closed-loop PI control architecture with feedforward. The method includes: S1. Read the unit working stage signal to determine the current working stage of the MSC; S2. Based on the current working stage of the MSC, a three-phase pseudo-random sequence is used to stimulate the secondary measurement signal of the MSC controller hardware in the loop to avoid the problem of step-by-step identification caused by the cascade of the inner and outer loop PI controllers: the three-phase pseudo-random sequence signal is superimposed with the stator current measurement signal and the rotor current measurement signal to obtain a stator current superposition signal and a rotor current superposition signal, and the stator current superposition signal and the rotor current superposition signal are used as the actual stator current and rotor current input signals of the MSC controller; S3, using particle swarm optimization to identify the proportional and integral parameters of the MSC controller; S4, after identifying the proportional and integral parameters of the MSC controller, interrupting the communication between the MSC controller and the simulator, and shielding the MSC controller; S5. Based on the unit working stage signal, the hardware-in-the-loop simulation model of the MSC current working stage is called based on the finite state machine principle to verify the parameter identification results: if the simulation result is wrong, jump to S3 and perform secondary identification; if the simulation result is correct, save the MSC controller parameter identification results; S6: If the unit's working stage changes and it is not the shutdown stage, jump to S1 and start parameter identification for the next working stage; S7. The unit's working stage is the shutdown stage, and the identification program exits.

2. The method according to claim 1, characterized in that The pulse frequency of the pseudo-random sequence signal is 800-1200 Hz.

3. The method according to claim 1, characterized in that The pulse frequency of the pseudo-random sequence signal is 1000 Hz.

4. The method according to claim 1, wherein The size of the three-phase pseudo-random sequence signal is determined by using 4%~8% of the maximum value of the MSC controller feedback input during the unit startup phase.

5. The method according to claim 1, wherein 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.

Images