The advent of swift increases in both energy costs and tighter environmental regulations give rise to a necessity for redesigning existing engineering processes such as power generation plants. These new designs aim to lower capital input and maximize energy output. These goals can be attained by solving a variety of optimization problems, such as minimum capital cost (MCC), minimum total annualized cost (MTAC), minimum utility cost, maximum power generation and so on. Conventional methods approach the optimization design as a combinatorial problem by creating a pre-established network (superstructure) that is then solved. Such superstructure-based methods cannot guarantee global optimality.

In this work we apply the infinite-dimensional state-space (IDEAS) framework with the MCC and MTAC objective functions to the design of power generating plants. The aforementioned objective functions are chosen so as to reflect economies of scale and are thus reverse convex functions of the corresponding characteristics of the corresponding units (heat exchanger's heat exchange area, compressor's power consumption, turbine's power generation, etc). An infinite number of linear equations is used to describe the feasible region. Global solution of this infinite-dimensional optimization problem is pursued through a truncation, underestimation procedure that is guaranteed to converge in the limit to the global optimum. Software developed at UCLA will then be used to automatically generate a globally optimal flowsheet of a power generating plant.

Keywords: IDEAS, Power plant, minimum capital cost, minimum total annualized cost, infinite-dimensional linear program