摘要
针对罚函数法处理有约束问题时存在的不足,采用Lagrange乘子法优化换热网络。为求解Lagrange函数方程组,根据确定性方法,提出最速下降法求解策略以及Powell法求解策略。通过极小值判断机制,保证Lagrange函数方程组的解是原换热网络目标函数值的极小值。根据实际工况,提出结构进化策略,与Lagrange乘子法相结合,实现了换热网络全局最优化。通过经典算例验证了两种求解策略的有效性、准确性以及结构进化策略的通用性。与文献结果进行对比,结果表明本算法具有较强的局部搜索能力以及全局搜索能力,能够找到更优的换热网络结构,有利于在工业生产中节约成本。
In allusion to the deficiency of penalty functions for constrained problems,a Lagrange multiplier method was adopted to optimize the heat exchanger network. To solve the Lagrange function equations,the steepest-descent method and the Powell method solving strategy according to the deterministic approach were proposed. The minimum value judgment mechanism ensures that the Lagrange function equation solution equals the minimum objective function value of the original network. According to the actual working conditions,a structure evolution strategy combined with a Lagrange multiplier method was proposed to reach the aim of global optimization. The validity and accuracy of these two methods,as well as the universality of the structure evolution strategy were verified by two benchmark problems. Compared with literature results,the proposed approaches have both strong local and global search abilities to find better heat exchanger network structures,which is conducive to cost saving in industrial production.
出处
《化工进展》
EI
CAS
CSCD
北大核心
2016年第4期1047-1055,共9页
Chemical Industry and Engineering Progress
基金
国家自然科学基金(51176125)
沪江基金研究基地专项(D14001)项目
