中文摘要：

针对复杂产品设计模型中设计变量与设计目标间关联关系复杂且存在变关联关系的问题，提出一种基于变关联的模型简化方法．该方法根据敏感度分析和信息熵理论将设计变量划分为强关联变量、弱关联变量和变关联变量，将弱关联变量简化后的设计模型分解为层次模型；在层次模型的每个计算步求解中，根据关联矩阵法计算变关联变量与设计目标的变关联强度，若为强关联则计算该变关联变量，若为弱关联则简化该变关联变量．最后以换热器结构优化设计为例，验证了文中方法的有效性．

英文摘要：

Correlations between design parameters and design objectives, especially the changeable correlations, are complicated for the design models of complex products. Hence, this paper presents a novel simplification method of design models for complex products based on the changeable correlation analysis. Design variables are partitioned into strong correlative variables, weak correlative variables and changeable correlative variables according to sensitivity analysis and information entropy theory, and then the design model is decomposed into the hierarchical model after simplifying weak correlative variables. In each calculation step of the hierarchical model, correlation strengths of changeable correlation variables are calculated according to the changeable correlation matrix. For each changeable correlation variable, if its correlation strength is strong enough in a calculation step, it is calculated in this step. Otherwise it is simplified. The structure optimization of a heat exchanger is taken as an example to verify the effectiveness of the approach.