摘要
针对汽轮发电机调速系统,提出了一种新的神经网络逆推控制方法,实现了汽门开度控制。首先,构造汽轮发电机的动态模型状态空间表达式,使之适合于逆推控制设计方法;假设模型系统动态完全已知,得到逆推控制器并证明了其稳定性;提出了基于神经网络的逆推控制律,利用神经网络补偿系统模型中的不确定项。该方法无需回归矩阵计算及参数线性化的假设,神经网络连接权值可在线调节,能够保证跟踪误差和网络连接权一致有界。仿真结果证明了其有效性。
A neural networks backstepping control technique is presented for turbo generator speed governor to realize turbine steam valve control. The state space equations are built up and fit for backstepping control design. With the assumption that the deduced equations can describe the system completely, the backstepping controller is designed and its stability is proved. A backstepping control rule based on neural networks is proposed,which uses neural networks to compensate the uncertain items. Neither regression matrix calculation nor assumption of parameter linearization is needed,and the weights of neural networks can be on-line regulated. The designed controller can guarantee the boundness of tracking error and weight updates. Simulative result shows its effectiveness. This project is supported by National Natural Science Foundation（60274009） and Research Fund for the Doctoral Program of Higher Education（20020145007）.
出处
《电力自动化设备》
EI
CSCD
北大核心
2005年第10期13-16,共4页
Electric Power Automation Equipment
基金
国家自然科学基金(60274009)
高等学校博士学科点专项科研基金(20020145007)
