叶红玲, 李耀明, 张颜明, 隋允康. 基于对数型Heaviside近似函数作为过滤函数的动力响应结构拓扑优化ICM方法[J]. 工程力学, 2014, 31(6): 13-20. DOI: 10.6052/j.issn.1000-4750.2013.03.0247
引用本文: 叶红玲, 李耀明, 张颜明, 隋允康. 基于对数型Heaviside近似函数作为过滤函数的动力响应结构拓扑优化ICM方法[J]. 工程力学, 2014, 31(6): 13-20. DOI: 10.6052/j.issn.1000-4750.2013.03.0247
YE Hong-ling, LI Yao-ming, ZHANG Yan-ming, SUI Yun-kang. STRUCTURAL TOPOLOGY OPTIMIZATION OF FREQUENCY RESPONSE PROBLEM BY APPROXIMATELY LOGARITHMIC HEAVISIDE FUNCTION BASED ON ICM METHOD[J]. Engineering Mechanics, 2014, 31(6): 13-20. DOI: 10.6052/j.issn.1000-4750.2013.03.0247
Citation: YE Hong-ling, LI Yao-ming, ZHANG Yan-ming, SUI Yun-kang. STRUCTURAL TOPOLOGY OPTIMIZATION OF FREQUENCY RESPONSE PROBLEM BY APPROXIMATELY LOGARITHMIC HEAVISIDE FUNCTION BASED ON ICM METHOD[J]. Engineering Mechanics, 2014, 31(6): 13-20. DOI: 10.6052/j.issn.1000-4750.2013.03.0247

基于对数型Heaviside近似函数作为过滤函数的动力响应结构拓扑优化ICM方法

STRUCTURAL TOPOLOGY OPTIMIZATION OF FREQUENCY RESPONSE PROBLEM BY APPROXIMATELY LOGARITHMIC HEAVISIDE FUNCTION BASED ON ICM METHOD

  • 摘要: 应用ICM(Independent Continuous and Mapping)方法, 建立了以重量最小为目标函数, 以连续频率带或离散点频率的简谐激励下的响应振幅为约束的拓扑优化模型. 引入了对数型Heaviside近似函数作为过滤函数, 并做了敏度分析, 利用对偶二次规划进行优化模型的求解, 并运用敏度过滤的方法处理动力响应数值不稳定的问题. 数值算例比较了利用对数型函数和幂函数作为过滤函数时对拓扑结构的影响, 结果显示利用对数型函数较幂函数结构优化迭代次数更少, 收敛更快.

     

    Abstract: A topology optimization model for frequency response was established based on ICM (Independent Continuous and Mapping) Method, in which the minimization of structure weight subject to the response amplitude of a harmonic excitation with a continuous frequency band or a discrete frequency. The approximately logarithmic Heaviside function was introduced as a filter function and design sensitivity was analyzed. The dual-quadratic programming was used to solve this optimization problem. The sensitivity filtering method was applied to deal with the numerical instability problems of frequency response. The numerical examples show that the modified logarithmic function and the power exponent function have different effects on a topology structure. In addition, the optimal results by using the modified logarithmic function is better than that by using the power exponent function in iteration number and convergence precision.

     

/

返回文章
返回