可主动变形波纹蒙皮驱动控制及神经网络建模  

Drive control and neural network modeling of corrugated skin with initiative deformation

作　　者：周春华[1] 王帮峰[1] 刘曌[2] 陈珩[1] 牟常伟 

机构地区：[1]南京航空航天大学机械结构强度与振动国家重点实验室,江苏南京210016 [2]南京航空航天大学自动化学院,江苏南京210016 [3]南京科瑞达电子装备有限公司,江苏南京211100

出　　处：《兵器材料科学与工程》2012年第1期25-28,共4页Ordnance Material Science and Engineering

基　　金：国家自然科学基金资助项目(51075207;10772081)

摘　　要：为满足变体机翼蒙皮连续光滑以及大变形要求,提出一种可主动变形的波纹型蒙皮构型,并进行驱动试验。试验中使用恒流源控制激励电流强度,在驱动试验数据的基础上建立一个以电流强度为输入参量,驱动位移为输出参量的RBF神经网络模型。该模型逼近曲线较为准确和贴近实际驱动特性,最大预测相对误差小于6%。A type of corrugated skin structure with active deformation was proposed for meeting requirements of continuous smooth and large deformation. And test samples and driving experiments were made. In the drive configuration test, the constant current source was used to control incentive current strength. A RBF neural network model was established based on the data of driving test, in which the current strength parameters were regarded as the input parameters and the driver displacement as output parameter. The approximation curve in the RBF neural network model is more accurate and close to the practical driving characteristics, and the biggest relative prediction error is less than 6%.

关 键 词：主动变形 波纹蒙皮结构 形状记忆合金增强复合材料 人工神经网络 

分 类 号：TB381[一般工业技术—材料科学与工程]