The paper extends the Conceptual Programming technology of Part I for retrofit design problems. Similar to the grassroots designs, the methodology includes a targeting and an optimisation stage and accounts for the simultaneous use of thermodynamic principles and mathematical programming. It discusses the development of two conceptual models which enable an automated auditing of the existing network (HEAT model), and a screening of its most promising modifications (TAME model). Both models are formulated and optimised as Mixed Integer Linear Programming problems whose sizes, unlike previous mathematical programming developments, remain very manageable even for large scale industrial problems. The optimisation of these models gives rise to the retrofit Hypertargets which involve, in addition to the energy, all other costing information available. The new methodology reports improvement of up to 40% against the established techniques and its application is illustrated in several examples.The paper extends the Conceptual Programming technology of Part I for retrofit design problems. Similar to the grassroots designs, the methodology includes a targeting and an optimisation stage and accounts for the simultaneous use of thermodynamic principles and mathematical programming. It discusses the development of two conceptual models which enable an automated auditing of the existing network (HEAT model), and a screening of its most promising modifications (TAME model). Both models are formulated and optimised as Mixed Integer Linear Programming problems whose sizes, unlike previous mathematical programming developments, remain very manageable even for large scale industrial problems. The optimisation of these models gives rise to the retrofit Hypertargets which involve, in addition to the energy, all other costing information available. The new methodology reports improvement of up to 40% against the established techniques and its application is illustrated in several examples.
