STEAM is a program for performing the synthesis and optimization of utility plants that satisfy given electrical, mechanical and heating demands of industrial processes using a mixed-integer nonlinear programming (MINLP) model.

The optimal solution is selected from a superstructure containing all the major conventional utility plant equipment. The electrical demands can be satisfied by an electric generator driven by a gas turbine or a steam turbine. Steam can be generated in different types of boilers included in the superstructure:

• Heat Recovery Steam Generator, used to recover heat contained in gas turbine or furnace exhaust gases, generating superheated high pressure steam. Supplementary firing is allowed in this units.
• high pressure boiler fired by fuel.
• medium pressure boiler fired by fuel.
• Waste heat boiler working at medium pressure, recovering heat from process flue gases or from process units as chemical reactors. This unit can raise saturated or superheated steam.

Using STEAM, the mechanical power demands can be satisfied by different types of steam turbines working with inlet and outlets at different steam pressures, and at a variable efficiency depending on this working conditions and the mechanical power generated. The following configurations are considered for high pressure turbines:

• Back-pressure turbine exhausting to medium pressure.
• Back-pressure turbine exhausting to low pressure.
• Extraction back-pressure turbine exhausting to medium and low pressure.
• Condensing turbine.
• Condensing turbine with steam extraction to medium pressure.

For medium pressure turbines the considered configurations are:

• Back-pressure turbine exhausting to low pressure.
• Condensing turbine.

Also the option of using an electric motor for each power demand is considered. The operation and retrofit studies of utility plants are addressed by fixing some of the options available in the model to match the fixed plant equipment options. The final solution includes the flowrates and enthalpies for steam, water and gas, and also the steam turbine efficiencies for the optimal configuration using the selected operational parameters. Also model statistics are available.

This interface is written in language C and it is implemented on Windows NT and IBM/AIX operating system. The way this interface works is through the use of different menus. It is flexible and easy to use. The input data can be entered as an existing data file or introduced interactively. These data and the internal operating parameters can be modified using the main menu and other following submenus. Given the data of the problem, the program is asked to perform a new design or a retrofit study of a given plant. Next, STEAM creates a GAMS file containing the corresponding model equations and optimizes the problem using DICOPT++. MINOS is used as the NLP solver, and OSL as the MILP for the master problems.

DICOPT++ (DIscrete and Continuos OPTimizer) was developed by J. Viswanathan and Ignacio E. Grossmann at the Engineering Research Center (EDRC) of the Carnegie Mellon University . It is based on the extensions of the outer- approximation algorithm for the equality relaxation strategy.

STEAM has been developed by J.C. Bruno under the supervision of Ignacio E. Grossmann .

References for STEAM:

• Bruno, J.C., Fernandez, F., Castells, F. and I. E. Grossmann, MINLP Model for Optimal Synthesis and Optimization of Utility Plants , Under preparation.
  
  
• Bruno, J.C., Guidelines for STEAM Interface , Carnegie Mellon University, Engineering Design Research Center, Pittsburgh, PA.