在换热网络数学模型及分级超结构基础上,提出蒙特卡罗结合微分进化算法,双层优化换热网络。外层利用蒙特卡罗算法全局范围内搜索最佳的换热网络结构;内层用微分进化算法优化换热器面积;通过设定合理的预期换热器个数,控制Grossmann超结构下产生换热器的个数,不但能提高求解效率,而且可以有效地解决大规模换热网络问题。最后,算例结果表明,基于蒙特卡罗微分算法能保证在全局搜索能力的前提下,混合求解策略能够获得较优的结果。
On the basis of Grossmann stage-wise superstructure,presenting an approach which combines Monte Carlo with differential evolution algorithm to optimize heat exchanger networks( HENs) problem. Mento Carlo method was developed to sample the optimal heat exchanger networks structure in global range randomly. After getting a sample,it used differential evolution algorithm to optimize all areas of the fixed structure. Through setting the fixed respective heat exchanger number in Mento Carlo,it effectively restricts the heat exchanger number of sample structure,therefore,it can solve more larger HENs problem. Finally,the reliability of mixed approach was demonstrated by two cases.