在用强制进化随机游走算法优化换热网络时，优化后期由于换热单元之间存在耦合关系，换热网络结构已基本稳定，难以新增加或者消去换热单元，因此结构变异较为缓慢。为了促进结构变异，提出了一种割裂结构耦合关系的高效优化方法。该方法基于耦合关系随机消去换热单元，使得匹配已稳定的冷热流股重新获得一定的换热潜能，从而具有匹配新换热单元的潜力，快速寻找到更多匹配方式。通过20SP和15SP算例进行有效性验证，分别取得1 395 461 $和5 115 557 $的年综合费用，表明割裂结构耦合关系的方法能进一步促进结构进化，增强算法的全局搜索能力。

When random walk algorithm with compulsive evolution is used to optimize the heat exchanger network,in the later stage of optimization,due to the coupling relationship between heat exchange units,the heat exchange network structure is basically stable,and it is difficult to add or eliminate heat exchange units,so the structure variation is relatively slow.Therefore,in order to promote the structural variation,an efficient optimization method is proposed to split the structural coupling relationship.Based on the coupling relationship,the heat exchange units are randomly eliminated by using this method,so that the matching stable cold and hot streams can regain a certain heat transfer potential again,which has the potential to match new heat exchange units and quickly finds more matching methods.Through 20SP and 15SP examples,the effectiveness verification is carried out,the annual comprehensive costs of 1 395 461 $ and 5 115 557 $ are obtained respectively.It shows that the method of splitting the structural coupling relationship can further promote the structural evolution and enhance the global search ability of the algorithm.

TK124

国家自然科学基金（21978171，51976126）；中国博士后科学基金（2020M671171）