换热网络全局优化的多维峰谷轮换法

Multidimensional peak-valley superseding approach for heat exchanger network global optimization

在峰谷轮换法的基础上提出了利用多维峰谷轮换法对以年综合费用最少为目标函数的无分流换热网络进行优化.在换热网络变量寻优过程中,首先将目标函数极小化得到一个局部极小值点,然后沿着多个方向进行变量搜索,找出最先跳出局部极小值点的变量组合,使得跳出局部极小值的搜索效率高于单变量的搜索效率,并可寻找到更好的换热网络结构,将极小化与跳出局部极小值点过程交替进行直至求得全局最优解.通过算例验证了该方法的可行性,且可找到比其它方法的优化结果更优的换热网络结构.

A multidimensional peak-valley superseding approach is presented for the optimization of comprehensive heat exchanger networks aimed at minimum cost. In the optimization process of heat exchanger network, the objective function is minimized to obtain a local minimum point, then the variables are searched along multiple directions. Finally, the first group of variables allows the solver to jump out of the local minimum is found. This algorithm can improve the search efficiency of jumping out the local minimum and looking for better structure until the optimal structure is found. Two examples are used to verify the feasibility of the method. The results show that this method can find better heat exchanger network structure compared to other methods.