中文摘要：

基于神经网络方法和状态预测补偿法,提出应用于含时滞空间结构振动控制的新方法。运用单个神经网络代替由两个神经网络组成的控制系统,以减少时滞产生环节和系统时滞。构造了基于神经网络的多步预测时滞补偿策略。研究表明：该文时滞补偿策略可有效运用于受不确定性外界荷载作用的多自由度大跨空间结构的风振控制,且不依赖于实际中难以获得的被控系统的精确数学模型;经时滞补偿后的结构振动控制效果优于未经时滞补偿的控制效果。

英文摘要：

An improved neural network time delay compensation method for the wind-induced vibration control of spatial latticed structures was developed, based on the neural network strategy and state predictive compensation technique. Instead of the combined system consisting of two neural networks, the single neural network system in the structural control was utilized to reduce the structural system time delay. An n-step predictive time delay compensation method was developed. The study reveals that the present compensation strategy is available to the wind-induced vibration control of spatial latticed structure subjected to uncertain external loads and does not depend on precise model of the controlled structural system which is almost impossible to be obtained. Finally, the numerical simulation shows that the vibration control effect on the structural system by this time delay compensated control strategy is better than that by the control strategy without time delay compensation.