Wind turbine generator resonance suppression system and method based on multi-target H infinity / generalized H additional damping controller

Abstract

The invention discloses a wind turbine generator resonance suppression system and method based on a multi-target H infinity / generalized H additional damping controller, and the system and method achieve the optimization of control parameters through the combination of pitch-torque control and a shrinkage coefficient particle swarm optimization gravitational search algorithm, and reduce the output power fluctuation and structural load of a wind turbine generator. And the resonance effect in a specific frequency range is weakened. The system is suitable for dynamic stability optimization of the large-scale wind turbine generator, improves the operation reliability and prolongs the service life.

Description

Technical Field

[0001] The present invention belongs to the technical field of wind power generation control, and specifically relates to a method based on multi-objective H ∞ / A wind turbine resonance suppression system and method based on a generalized H2 additional damping controller is used to suppress output power fluctuations and structural load accumulation caused by the resonance effect due to wind speed disturbances. Background Art

[0002] Wind turbines are susceptible to wind speed fluctuations during operation, leading to resonance within a specific frequency range, unstable power output, and fatigue damage to key components. While traditional control methods such as PID and model predictive control can partially mitigate power fluctuations, they struggle to effectively suppress resonance effects. In existing technologies, the single pitch angle control of the RC controller lacks adaptability to multivariable coupled dynamics, and the combined pitch angle-torque control lacks targeted damping optimization for critical frequencies. Furthermore, parameter adjustment relies on empirical trial and error, limiting control accuracy and robustness. Summary of the Invention

[0003] The present invention proposes a multi-objective H ∞ / The wind turbine resonance suppression system and method based on the generalized H2 additional damping controller solves the above problems through the following technical solutions:

[0004] 1. Linearized modeling: Generate the linearized state equation of the wind turbine, expressed as:

[0005]

[0006] Where, and are state variables and their derivatives; is the wind speed disturbance input; is the control input (in variable pitch control, it indicates the pitch angle change; in torque control, it indicates the torque change); and are the control outputs under H∞ and H2 performance indicators respectively; , , , , , , , and is a constant matrix.

[0007] 2. Multi-objective controller design: combining H ∞ Robust stability and generalized H2 performance are optimized, and a pitch-torque joint controller is designed. Its comprehensive performance function is expressed as:

[0008]

[0009] Where, For input To output The closed-loop transfer function H ∞ Criterion number paradigm; For input To output The closed-loop transfer function The generalized H2 criterion numerical paradigm; and are weight coefficients, corresponding to H ∞ Robust stability and generalized H2 control accuracy are determined through CPSOGSA algorithm optimization.

[0010] 3. Intelligent parameter optimization: The controller parameters are optimized using the contraction coefficient particle swarm optimization gravitational search algorithm (CPSOGSA) with the objective function of minimizing the time-weighted integral of absolute error (ITAE) of the state variables.

[0011] 4. Additional damping superposition: An optimized damping signal is superimposed on the output of the baseline controller (RC) to suppress the resonance peak in the 1P-2P frequency band, reducing output power fluctuations and fatigue loads.

[0012] Beneficial effect: The present invention adopts a multi-objective H ∞ A wind turbine resonance suppression system and method using a generalized H2 supplementary damping controller attenuates the resonance peak amplitude in the 1P-2P frequency band, reducing output power fluctuations and mechanical fatigue. This reduces power generation and rotor speed fluctuations by 19.97% and 11.90%, respectively, and lowers the standard deviations of tower top displacement and tower base rolling torque by 18.9% and 30.13%, respectively. The CPSOGSA algorithm improves parameter optimization efficiency to ensure a globally optimal solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 It is a multi-objective H ∞ / Generalized H2 additional damping controller (H ∞ / GH(2) C) Control block diagram showing the joint control logic of pitch angle and torque;

[0014] Figure 2 It is a flow chart of pitch angle and torque optimization, describing the iterative process of CPSOGSA algorithm;

[0015] Figure 3 This is the fitness curve of the algorithm during the pitch angle and torque optimization process, which is compared with the contraction coefficient particle swarm optimization algorithm (CPSO), showing the superior performance of the CPSOGSA algorithm;

[0016] Figure 4is the turbulent wind speed curve;

[0017] Figures 5 to 11 It is a comparison of controller performance and power spectrum density under turbulent wind conditions to verify the suppression effect of output power fluctuations and structural load accumulation caused by resonance effects. DETAILED DESCRIPTION

[0018] In order to explain the technical solution of the present invention in detail, the following is further described in conjunction with the accompanying drawings and specific embodiments. It should be understood that the described embodiments are only used to explain the present invention and are part of the embodiments of the present invention, rather than all the embodiments.

[0019] Step 1: Model linearization

[0020] Under the rated wind speed (11.4m / s) condition, a linearized model of the wind turbine is generated and the state matrix is extracted.

[0021] Step 2: Controller Construction

[0022] Design H ∞ / Generalized H2 output feedback controller, and optimize the controller gain matrix.

[0023] Step 3: Parameter Optimization

[0024] Set ITAE as the objective function and use the CPSOGSA algorithm to iteratively optimize the weight coefficients α, β and controller gains. The specific process includes:

[0025] (1) Initialize the particle swarm: set the population size to 50 and the number of iterations to 200;

[0026] (2) Gravity calculation and velocity update: introduce dynamic inertia weight (0.4~0.9) and velocity limit (±10% search range);

[0027] (3) Iteration termination condition: ITAE change rate <1% or the maximum number of iterations is reached.

[0028] Step 4: Simulation Verification

[0029] Comparison of RC and MOH in turbulent wind conditions ∞ / GH2 The control effect of C is used to verify the suppression effect of output power fluctuation and structural load accumulation caused by the resonance effect.

[0030] Example 1

[0031] Taking the NREL 5MW wind turbine as the object, the average wind speed of turbulent wind is 18m / s and the turbulence degree is 0.05.

[0032] See also Figure 1A wind turbine resonance suppression system based on a multi-objective H∞ / generalized H2 additional damping controller comprises a linearized modeling module, a multi-objective H∞ / generalized H2 additional damping controller, a parameter optimization module, and a superposition module. The linearized modeling module uses the wind turbine's measured data (tower displacement change, tower displacement change rate, generator speed change, drive system angle change, and drive system angle change rate) as state variables, wind speed as a disturbance input, and pitch angle change and torque change as control inputs to generate a linearized state equation. The multi-objective H∞ / generalized H2 additional damping controller, comprising a pitch angle control module and a torque control module, suppresses resonance effects by jointly controlling pitch angle and torque. The parameter optimization module optimizes control parameters using a contraction coefficient particle swarm optimization gravitational search algorithm (CPSOGSA) to minimize the time-weighted integral of absolute error (ITAE) of the state variables. The superposition module is used to superimpose the additional damping signal onto the output of the baseline controller (RC) to achieve dynamic suppression of critical frequency resonance peaks.

[0033] See also Figure 2 The pitch angle and torque optimization flow chart shows that 200 iterations were performed on the pitch angle and torque respectively. The algorithm fitness curve is shown in the figure below. Figure 3 As shown in the figure, compared to the shrinking coefficient particle swarm optimization (CPSO), the CPSOGSA algorithm experiences a rapid decline in fitness during the initial iterations (within 50 iterations), indicating that the algorithm is rapidly approaching the optimal solution. Fitness stabilizes after 130 and 50 iterations, respectively, reaching the optimal solution. The entire algorithm iteration process demonstrates that the CPSOGSA algorithm demonstrates superior global optimization performance, both in speed and efficiency, compared to the CPSO algorithm.

[0034] See also Figure 4 The turbulent wind speed curve has an average wind speed of 18m / s and a turbulence degree of 0.05. Under turbulent wind conditions, the wind turbine's power generation, rotor speed, tower base rolling moment, tower top displacement, and blade root load curves for the three blades are shown in the following table. Figures 5 to 9 Compared with the basic controller (RC), the multi-objective H ∞ / Generalized H2 additional damping controller (H ∞ / GH(2) C) reduces the standard deviation of generated power from 33.50kW to 26.81kW (a decrease of 19.97%), the rotor speed fluctuation from 0.084rpm to 0.074rpm (a decrease of 11.90%), the standard deviation of the tower base rolling moment from 1856.18kNm to 1296.84kNm (a decrease of 30.13%), the tower top displacement from 0.0127m to 0.0103m (a decrease of 18.90%), and the standard deviation of the average load of the three blades from 1344.69kNm to 1341.92kNm (a decrease of 0.2%).

[0035] See also Figure 10 The power spectrum density of the rolling moment, yaw moment and tower top displacement of the wind turbine. ∞ / Generalized H2 additional damping controller (H ∞ / GH(2) The peak frequencies of the controller (C) and the basic controller (RC) are basically the same, but their ability to suppress fluctuations at the resonant frequency is very different. ∞ / GH(2) C reduces the resonance influence through additional damping and improves the control accuracy of the system, thereby enhancing the additional damping effect. It significantly reduces the amplitude peaks of the tower base rolling moment power spectrum density, yaw moment power spectrum density and tower top displacement power spectrum density in the 1P to 2P frequency range, showing excellent load fluctuation suppression performance.

[0036] The above results show that the wind turbine resonance suppression system and method based on the multi-objective H∞ / generalized H2 additional damping controller described in the present invention optimizes the control parameters by combining pitch-torque control and the contraction coefficient particle swarm optimization gravity search algorithm, thereby weakening the resonance effect caused by wind speed disturbances within a specific frequency range and reducing the output power fluctuation and structural load of the wind turbine.

Claims

1. A multi-objective H ∞ / A wind turbine resonance suppression system based on a generalized H2 additional damping controller is characterized in that: include: Linear modeling module, multi-objective H ∞ / Generalized H2 additional damping controller, parameter optimization module and superposition module.

2. The system according to claim 1, wherein: The linearized modeling module is used to generate the linearized state equation of the wind turbine; the multi-objective H ∞ / The generalized H2 additional damping controller includes a pitch angle control module and a torque control module, which suppresses the resonance effect by jointly controlling the pitch angle and torque; the parameter optimization module uses a contraction coefficient particle swarm optimization gravitational search algorithm (CPSOGSA) to optimize the control parameters to minimize the time-weighted integral of absolute error (ITAE) of the state variables; the superposition module is used to superimpose the additional damping signal on the output of the baseline controller (RC), reducing the output power fluctuation and structural load of the wind turbine and achieving dynamic suppression of the critical frequency resonance peak.

3. The system according to claim 1, wherein: The linearized state equation of the wind turbine generated by the linearized modeling module is: Where, and are state variables and their derivatives; is the wind speed disturbance input; is the control input (in variable pitch control, it indicates the pitch angle change; in torque control, it indicates the torque change); and are the control outputs under H∞ and H2 performance indicators respectively; , , , , , , , and is a constant matrix.

4. The system according to claim 1, wherein: The H ∞ The performance index is used to ensure the robust stability of the system, and the generalized H2 performance index is used to optimize the control accuracy. Its comprehensive performance function is expressed as: Where, For input To output The closed-loop transfer function H ∞ Criterion number paradigm; For input To output The closed-loop transfer function The generalized H2 criterion numerical paradigm; and are weight coefficients, corresponding to H ∞ Robust stability and generalized H2 control accuracy are determined through CPSOGSA algorithm optimization.

5. The system according to claim 1, wherein: In the parameter optimization module, the CPSOGSA algorithm includes the following steps: S1: Initialize the position and velocity of the particle swarm; S2: Calculates the inter-particle attraction and updates the velocity, introducing dynamic inertia weight and velocity limit mechanism; S3: Iterate the optimization until the minimum ITAE objective function is achieved and the optimal control parameters are generated.

6. The system according to claim 1, wherein: The additional damping control signal acts on the resonant frequency range (1P~2P frequency band) of the tower base rolling moment, yaw moment and tower top displacement to reduce the output power fluctuation and fatigue load of the wind turbine and dynamically suppress the critical frequency resonance peak.

7. A multi-objective H ∞ / A wind turbine resonance suppression method based on a generalized H2 additional damping controller is characterized in that: The following steps are involved: S1: Establish a linearized model of the wind turbine and describe its dynamic behavior through state equations; S2: Using the existing reference open source controller (RC) as the basic controller, design a multi-objective H ∞ / Generalized H2 additional damping controller, which jointly controls pitch angle and torque to suppress resonance effects; S3: Optimize the control parameters using the CPSOGSA algorithm described in claim 2 to minimize the time-weighted absolute error (ITAE) of the state variables; S4: Generate an additional damping signal based on the optimized parameters and superimpose it on the baseline controller output to reduce the output power fluctuation and fatigue load of the wind turbine and dynamically suppress the critical frequency resonance peak.

Images