基于RBF的电液变量施肥控制系统PID参数整定

PID Parameter Setting of Electro- hydraulic Variable Rate Fertilization System Based on RBF

电液变量施肥控制系统的非线性和PID算法的局限性,导致常规PID控制已不能满足控制系统性能要求。为此,提出了基于RBF神经网络整定PID参数的方法,利用自适应RBF神经网络辨识被控对象Jacobian信息,采用梯度下降法计算PID参数Δk_p、Δk_i、Δk_d,对系统进行增量式PID控制。与采用增量式PID的系统阶跃响应曲线对比可知,利用RBF-PID算法的系统具有良好的动态性能及较强的自适应性。

Nonlinearity of electro- hydraulic variable rate fertilization system and limitations of PID algorithm lead that normal PID control cannot meet the performance requirements of control system. Aimed at this,a method of setting PID parameter based on RBF neural network has been proposed,which takes advantage of adaptive radial basis function neural network( ARBFNN) to identify the Jacobian information of controlled object,adopts gradient descent to calculate PID parameters Δk_p,Δk_i,Δk_d,and conducts incremental PID control on system. Compared with the step response curves of incremental PID algorithms,it can be learnt that the system which adopts RBF- PID algorithm has excellent dynamic performance and strong adaptivity.