序列数编码的遗传算法柔性优化多股流板翅式换热器通道排列

Flexible optimization of passage arrangement for multi-stream plate-fin heat exchangers using genetic algorithm with ordinal number encoding

实数编码的遗传算法在优化多股流板翅式换热器通道排列中进行个体间交叉、变异操作后,存在无法保证各流体的通道数恒定不变的问题。对此,以通道的热负荷累积均方差为目标函数,建立通道排列的优化模型,提出基于序列数编码方式的改进遗传算法进行多股流板翅式换热器通道优化设计,实现了个体间交叉和变异遗传操作。对一包含4个流股57个通道的板翅式换热器通道排列进行优化,并通过多工况点设计对系统柔性进行分析。结果表明:本文优化设计方案的累积热负荷均方差为3562.9W,比文献经验值小2.7%,相比于实数编码遗传算法得到的结果,减少了15.1%,且累积热负荷在零线上下均匀波动。文章表明序列数编码的遗传算法增加了遗传算法种群中个体的多样性,提高了搜索效率具有更好的全局搜索能力。

There is a problem that the passage number cannot remain the same,when the crossover and the mutation operation are performed to optimize the passage arrangement of multi-stream plate-fin heat exchanger（MS-PFHE） by genetic algorithm（GA） using the real coding. Hence,an improved GA using ordinal number encoding was proposed to optimize the passage arrangement of MS-PFHE. On this basis,the crossover and the mutation operation can be successfully performed between individuals,and an optimization model of passage arrangement was built with the mean square deviation of heat load accumulation in the passage as the objective function. The passage arrangement of MS-PFHE was optimized,which contained 4 streams and 57 passages,and flexibility of the system was analyzed by several operation points. The results showed that,the mean square deviation of heat load accumulation of the optimized design scheme in this study was 3562.9W,which was 2.7% lower than the empirical method in the literature,and it was 15.1% lower than the result obtained by GA using the real coding. In addition,the heat load accumulation fluctuated to the zero line. It indicated that GA using ordinal number encoding can increase population individuals variety in GA and improve searching efficiency,so that the algorithm can have better performance in reaching the global optimum.