Uncertainty and sensitivity analysis of thermodynamic models using equal probability sampling (EPS)

A novel approach called equal probability sampling (EPS) is used for analyz ing uncertainty and sensitivity in thermodynamic models. Uncertainty and se nsitivity analysis for simulation and design of industrial processes are be coming increasingly important. The (EPS) method produces more realistic res ults in uncertainty analysis than methods based on other sampling technique s such as Latin hypercube sampling (LHS) or shifted Hammersley sampling (SH S), when parameters are highly correlated. When parameters are not correlat ed, EPS reduces to the LHS method. The EPS method is based on resampling to obtain uniform coverage over level sets of the objective function used to obtain the parameters of the model. The existence of unfeasible situations is substantially reduced with EPS. It can be extended to any regression mod el describing other kinds of physical applications and can be used as a bet ter tool to estimate more reliable safety factors in the design and simulat ion of industrial chemical processes. (C) 2000 Elsevier Science Ltd. All ri ghts reserved.