Adaptive Partition\nLinearization Global Optimization\nAlgorithm and Its Application on the Simultaneous Heat Exchanger Network\nand Organic Rankine Cycle Optimization

The simultaneous optimization problem of the heat exchanger\nnetwork\nand organic Rankine cycle (HEN-ORC) poses significant challenges due\nto its highly nonconvex and nonlinear equations. We develop an adaptive\npartition linearization global optimization algorithm which is suitable\nfor a wide range of mixed integer nonlinear programming (MINLP) problems\nand specially customized for HEN-ORC. The algorithm identifies convex\nequations of the logarithmic mean temperature function and the power\nfunction within the HEN-ORC model, which are relaxed by the first\nTaylor expansion and piecewise linearization. A multilevel McCormick\nrelaxation is applied for the bilinear/multilinear functions derived\nfrom the HEN-ORC energy balance equations. The algorithm achieves\nglobal optimality by solving mixed integer linear programming and\nNLP submodels iteratively, enhancing the lower bound adaptively. Tested\non seven heat exchanger networks and waste heat power generation cases,\nit outperforms two mainstream MINLP global optimization solvers (Baron\nand Couenne). The current best solutions are obtained for both a HEN\nand a HEN-ORC case, respectively.