The energy systems in large industrial plants are often very complex involving hundreds of items of equipment such as furnaces, turbines, boilers, generators, etc., and numerous energy forms such as oil, natural gas, steam, electricity and so on. It is usually not obvious how to operate the system to minimize energy consumption, thereby minimizing fuel expenditures. Computer models can be effective tools for the plant manager in tackling this problem. This paper presents the results of the application of a new modeling procedure to the energy system in an existing Canadian petrochemicals plant. The new procedure identified an estimated $600,000 per annum in additional energy savings over other modeling techniques that had been applied to the plant. The procedure includes second-law measures in a convenient and easily-applied form.