Heat exchanger networks (HENs) in separated plants can be simultaneously synthesized through process streams directly or intermediate fluids indirectly. However

the direct and indirect methods are studied separately in most existing researches. As the result

conventional designs are probably suboptimal

because optimal solutions may call for hybrid approach wherein direct and indirect integration methods are used synchronously. To circumvent this drawback aforementioned

we propose in this research a novel methodology to synthesize multi-plant HENs considering both direct and indirect approaches. The methodology employs a novel superstructure covering most potential topologies for both interplant and intra-plant heat integration. We also take into account multiple kinds of intermediate fluids for indirect integration and this has not been fully addressed in previous research. A mixed-integer nonlinear programming (MINLP) is formulated to optimize multi-plant HENs involving indirect and direct methods. One example from existing literature and one industrial problem are solved to demonstrate the methodology's capability.