This thesis provides a systematic approach for redesigning flowsheet structures of existing chemical processes to improve the flexibility of operating over a range of uncertain parameters. Two basic problems are addressed: (1) Determine the parametric and/or structural modifications so as to achieve a desired degree of flexibility at minimum investment cost for the retrofit. (2) Establish the trade-off between the retrofit cost and the revenue that can be expected from having a higher flexibility in the process so as to determine the optimal level of flexibility.; Basic analytical properties of flexibility are presented for linear models with which very efficient reduced formulations are developed, and which have the unique feature of including explicit constraints for flexibility. Based on these formulations, trade-off curves relating cost to flexibility can be generated. Further, by considering probability distribution functions for the uncertain parameters, an efficient stochastic optimization method is presented for estimating the expected revenue. With this, the curve relating the expected revenue to flexibility can be generated which allows one to determine the level of flexibility that maximizes the total profit of the process.; Redesign methods for the case of nonlinear models are also proposed. A novel computational strategy is presented for determining minimum cost modifications for a fixed degree of flexibility, which relies on the iterative solution of an optimal redesign formulation that features as constraints a relaxation of the feasibility function for the region of flexibility. Special structures of nonlinear models, such as bilinear in terms of the uncertain parameters and the control variables, are also analyzed. Finally, the problem of establishing optimal trade-offs between cost and expected revenue for nonlinear models is also presented.
展开▼
