Systematic building-testing of dynamic models for training simulators

Abstract A scheme to build and to test models of process units for power plant simulators is presented. The stages of this scheme consist of: (1) the mathematical model of each piece of equipment; (2) a set of standard procedures for every model: to set parameters, to evaluate the initial and working conditions, and to describe its dynamic or static behavior; (3) the automatic analysis of these models, with a scheme similar to the Automatic Differentiation, to verify the correct use of identifiers, to obtain the incidence matrix model, and to verify that no vulnerable arithmetic is presented. And finally, (4) the model is validated at different steady-state conditions and transients. With these stages, models are ready to be coupled to form a process flowsheet. An illustrative model example is shown.