Disjunctive multiperiod optimization methods for design and planning of chemical process systems

• Published in: Computers & chemical engineering 1999-08, Vol.23 (8), p.1075-1095
• Main Authors: van den Heever, Susara A., E. Grossmann, Ignacio
• Format: Article
• Language: English
• Subjects: Applications of mathematics to chemical engineering. Modeling. Simulation. Optimization; Applied sciences; Chemical engineering; Cost; Exact sciences and technology; Multiperiod optimization; Process systems
• ISSN: 0098-1354; 1873-4375
• DOI: 10.1016/S0098-1354(99)00273-2

In this paper, we present a general disjunctive multiperiod nonlinear optimization model, which incorporates design, as well as operation and expansion planning, and takes into account the corresponding costs incurred in each time period. This model is formulated with the use of logic and disjunctive programming, and includes Boolean variables for design, operation planning and expansion planning. We propose two algorithms for the solution of these problems. The first is a logic-based Outer Approximation (OA) algorithm for multiperiod problems. The second is a bilevel decomposition algorithm, that exploits the problem structure by decomposing it into an upper level design problem and a lower level operation and expansion planning problem, each of which is solved with the logic-based OA algorithm. Applications are considered in the areas of design and planning of process networks, as well as retrofit design for multiproduct batch plants. The results show that the disjunctive logic-based OA algorithm performs best for small problems, while the disjunctive bilevel decomposition algorithm is superior for larger problems. In both cases, a significant decrease in MILP solution time and total solution time is achieved compared to DICOPT++. Results also show that problems with a significant number of time periods can be solved in reasonable time.