Optimization of heat exchanger networks by coupled simulated annealing and NLP procedures

This paper addresses die problem of systematic generation of heat exchanger networks for minimizing the global cost. To an upper level, the approach implements simulated annealing to propose HEN configurations. This information is then used in a lower level within a nonlinear programming (NLP) frame to optimize die operating variables. In this paper, the automatic generation, initialization and optimization of these nonlinear problems are only presented. The proposed method provides minimum size problems subject to a minimum number of variables and constraints.