摘要
针对同时存在整型变量和连续型变量的换热网络综合问题,提出一种双层优化方法。外层以换热网络的温差均匀性因子作为网络结构性能的评价指标,通过蒙特卡洛随机抽样技术产生试探结构,采用整型优化算法逐步进化外层结构;内层以最小年综合费用作为优化指标,采用动态更新子群的改进粒子群算法优化连续变量。优化结果表明,温差均匀性因子可以有效评价换热网络的结构性能,从而指导结构的进化;改进的粒子群算法具有更强的全局搜索能力,相关算例均找到了更优的网络设计,应用于工业生产实际,可以有效节约成本。
A bi-level simultaneous approach for heat exchanger network synthesis( HENS) is proposed to solve the problem containing complex integer and nonlinear variables. In the upper level,the structure optimization strategy is applied to optimize the test structures produced by Monte Carlo method,with the temperature uniformity factors of heat exchanger network as the performance evaluation indicators of each HEN structure. In the lower level,an improved particle swarm algorithm is proposed using dynamic update subgroups to optimize continuous variables,where the minimal total annual cost for each topology is set as the evaluation. The results show that the temperature uniformity of heat exchanger network can be used to evaluate the performance of the structure and guide the evolution of the structure effectively. The improved PSO algorithm has strong global searching ability. The proposed approach is applied to several cases from the literature,and the obtained optimal structures are better than those from other methods.
出处
《热能动力工程》
CAS
CSCD
北大核心
2017年第7期17-23,共7页
Journal of Engineering for Thermal Energy and Power
基金
国家自然科学基金(51176125)
沪江基金研究基地专项(D14001)
关键词
换热网络
粒子群算法
动态子群
温差均匀性因子
heat exchanger network synthesis
particle swarm optimization
dynamic update subgroup
temperature uniformity factor
