改进型PRP-BP-PID算法在风力机变桨距控制中的应用

Application of the Improved PRP-BP-PID with in Variable Pitch Controller of Wind Turbine

针对变桨距风力机存在非线性、时变性、抗干扰性和滞后性等问题,并在分析双馈风力发电机组系统特性和变桨距控制要求的基础上,设计了一种改进型共轭梯度BP神经网络PID算法控制器。该控制器通过引入具有附加动量项的PRP共轭梯度法进行搜索,提高和改善传统BP神经网络的收敛速度和性能。在Mat-lab/Simulink仿真软件上搭建仿真平台,验证该控制器在额定风速之上的控制效果。与传统的变桨距PI控制相比,在相同风速变化条件下,所设计的改进型变桨距PID控制器能够较好地保持风轮转速,使输出功率基本保持在额定功率附近,具有较强的鲁棒性。

In order to solve the problems of non-linear, time-varying, anti-disturb and time-lag, a variable pitch PID controller with improved BP neural network based on conjugate gradient was proposed after analyzing the characteristics of a doubly-fed induction wind generator and the requirement of variable pitch control Using the Polakribiere- polyak （PRP） conjugate gradient method with additional momentum term, this controller could make the convergence of neural network faster and more stable. By using Matlab/Simulink software as a simulation platform, the control effect of the designed controller for the wind speed above rated value was verified. Comparing with traditional variable pitch PI control, at the same wind speed condition, the proposed controller has the advantage of better robustness. And then, it can make rotating speed of the wind wheel more stable, and maintain the output power at near rated power.