Abstract:

The overall characteristics of structure evolution in the optimization of heat exchanger networks using random walk algorithm with compulsive evolution (RWCE) are studied by monitoring the optimization process of the heat exchanger units in the optimal structures. The results show that the heat exchanger units formed in the early stage turn to be impregnable, making it difficult to retain the newly generated units based on the local optimal structure and promote the global optimization. To solve the problem, this paper proposes a reconfiguration strategy of heat exchange units to enhance the ability of structural mutation for RWCE. First, the existing heat exchanger units in the optimal structure would be eliminated randomly to vacate the solution space for generating new units. Then, new heat exchange units with the same number of eliminated ones are reconfigured according to the constraint conditions, which can improve the structural diversity and expand the search domain. Meanwhile, combined with the indicator for monitoring optimization status, the mutation period is adjusted randomly to maintain high mutation activity and enhance the global optimization ability of the algorithm. The improved RWCE algorithm with the proposed strategy is applied to two cases and the mechanisms of the strategy are analyzed and summarized. The results obtained are better than those of the existing liferature, confirming the effectiveness of the strategy.

Key words: random walk algorithm with compulsive evolution, node-wise non-structural superstructure model, computer simulation, heat exchanger units reconfiguration strategy, structural diversity, optimal design

摘要：

应用强制进化随机游走算法优化换热网络的过程中，通过监测最优结构中换热单元优化进程，探究总体结构进化特点，发现存在换热单元先入为主的现象，主体结构稳固难以被突破，导致换热网络优化进入“停滞”状态。基于此，提出了换热单元重构策略，该策略旨在增强结构变异能力，经过一定阶段概率性消除结构中既有换热单元，腾出结构进化空间，再依据约束条件重新构建换热单元，既丰富了结构多样性又扩大了搜索域。此外建立优化状态监测指标，维持结构较高的变异活性，增强算法全局寻优能力。最后将该策略加入算法流程优化多个换热网络实例，总结分析其作用机理，所得结果均优于现有文献结果，证实了该策略的有效性。

关键词: 强制进化随机游走算法, 节点非结构模型, 计算机模拟, 换热单元重构策略, 结构多样性, 优化设计