摘要
针对RWCE（random walk algorithm with compulsive evolution，RWCE）算法优化有分流换热网络时，由于变量维数过多造成的算法全局搜索能力下降问题，提出了一种主辅强制进化随机游走算法（master-salve random walk algorithm with compulsive evolution，MS—RWCE）。算法通过将换热量和分流比的优化分开处理，建立一种以换热量优化为主导、分流比优化为辅从的新的进化机制。同时，为了进一步保证换热量优化的主导地位，引入分流比隔代优化策略，减少换热量和分流比在优化中的相互制约。采用9SP和15SP算例对MS-RWCE算法的有效性进行验证，MS-RWCE算法优化9SP和15SP算例，较采用RWCE算法的年综合费用分别下降了4395和18 420$/a，证明MS-RWCE算法可以实现换热网络优化质量的提升。
When the RWCE algorithm is applied to optimize heat exchanger network with split streams, there is a significant reduction in global search ability due to the excessive variables. In view of this, a MS-RWCE algorithm （master-salve random walk algorithm with compulsive evolution） is proposed in this paper. The algorithm separates the optimization of heat loads and split ratio and establishes a new evolu- tion mechanism where the heat loads are optimized dominantly, and meanwhile both the selection of the step size and the optimization are dependent on the heat loads. Further, the atavistic evolution is applied to optimize the split ratio and reduces the restriction between the heat loads and split ratio. This strategy en- sures the dominant position of the heat loads optimization by reducing the optimization frequency of the split ratio. Finally,9SP and 15SP case studies are used to verify the effectiveness of the MS-RWCE algo- rithm. The results of 9SP and 15SP obtained by MS-RWCE algorithm are better than most of the best re- sults published; they are reduced by 4 395 and 18 420 $/a, respectively, compared to the TAC opti- mized by RWCE algorithm. Hence,it is proved that the MS-RWCE algorithm can improve the overall opti-mization quality.
作者
孙涛
崔国民
陈家星
赵兵涛
SUN Tao;CUI Guo-min;CHEN Jia-xing;ZHAO Bing-tao(Institute of New Energy Science and Technology,University of Shanghai for Science and Technology,Shanghai,China,Post Code ： 20009)
出处
《热能动力工程》
CAS
CSCD
北大核心
2018年第10期8-15,共8页
Journal of Engineering for Thermal Energy and Power
基金
国家自然科学基金(51176125)
上海市科委部分地方院校能力建设计划(16060502600)~~
关键词
强制进化随机游走算法
换热网络优化
换热量
分流比
random walk algorithm with compulsive evolution
heat exchanger network optimization
heatloads
split ratio
