基于神经元S型传递函数的磁流变阻尼器力学模型研究  

Research on a new mechanical model for MR damper based on neuron sigmoid transfer function

作　　者：梅德庆[1] 孔天荣[1] 陈子辰[1] 

机构地区：[1]浙江大学现代制造工程研究所,浙江杭州310027

出　　处：《功能材料》2006年第5期776-779,共4页Journal of Functional Materials

基　　金：国家自然科学基金资助项目(50405036);浙江省自然科学基金资助项目(Y104462);浙江省教育厅科研资助项目(20030262)

摘　　要：针对已有的一些磁流变阻尼器力学模型虽然能在一定程度上反映其实际动态特性,但对于控制应用却不甚方便、精确。本文通过引入神经网络理论中的神经元S型传递函数,并结合非线性Bingham塑性模型、非线性滞环双粘性模型和非线性粘弹塑性模型,提出了一种S型磁流变阻尼器力学新模型,并利用最小二乘法对该模型的参数进行识别。试验与仿真结果表明:该模型的动态滞环特性与实际测试值能很好地吻合,该模型在实际系统中将具有良好的可控性。Because of the nonlinearity of MR fluid, it is difficult to establish a accurate mechanical model for MR damper. Some existing mechanical models of MR damper can describe its actual dynamic characteristics approximately, but usually they are not convenient and accurate enough in the application of vibration control. Based on the compare and analysis of some main existing mechanical models of MR damper, a new sigmoid mechanical model is established with sigmoid transfer function in NN theory, and its parameters are identified by least square method. The results of experiment and simulation show that the hysteresis loop characteristic of the sigmoid mechanical model consist with the test data well, and the new model has good controllability in its application.

关 键 词：磁流变液 阻尼器 力学特性 参数识别 神经元 

分 类 号：TM27[一般工业技术—材料科学与工程]