结构试验机的神经网络内模自适应跟踪控制系统

针对结构试验系统的非线性和不确定性特性，研究了一种基于神经网络的非线性内模自适应加载控制方法。引入的神经网络内模可跟踪学习对象的时变动力学，控制器的设计较少依赖于对象的先验知识，控制参数的调整是基于被控过程的测量信息，利用导出的神经网络算法来实现的。实验结果证明该系统具有良好的控制效果。

Neural Networks Based Internal Model Adaptive Tracking Control System for Structural Testing Machine

A nonlinear internal model adaptive control method based on neural networks, with respect of the complex nonlinearities and uncertainties in electrohydraulic servo structural testing system, is presented and an internal model adaptive loading control system for the structural testing machine is designed in this paper. A feedforward neural network is defined to learn timevarying system dynamics as the internal model, which can be modified on line. The controller can be designed with little priori information required about the controlled plant and regulated on line by using the measured input/output data with the neural network learning method derived from the BP algorithm. The effectiveness of the tracking control system is verified by experimental results.