Title

Optimization of chemical processes under uncertainty in the case of insufficient process data at the operation stage

Date of Completion

January 2004

Keywords

Engineering, Chemical

Degree

Ph.D.

Abstract

Recent advances in computer hardware and computationally efficient algorithmic developments in process system engineering have increased the tendency for many large companies to use computer-aided design of chemical processes to complement and reduce the load on experimental research. The use of simulations allows obtaining optimal design and operating regimes of the process. A simulation is based on the construction of an appropriate process model, which can then be employed in an optimization model. Satisfaction of design specifications during the design of a chemical process is complicated by the presence of uncertainty in the process model. Process simulations are often carried out in commercial software packages such as ASPEN, HYSIS and gProms. The problem with these simulators is that they do not take uncertainty into account. Instead nominal or fixed values of process parameters are employed for simulations, followed by the use of empirical or heuristic-based over-design factors. The consequence of this is that we cannot guarantee the satisfaction of process constraints during the operation stage. A number of flexibility analysis methods have been developed to deal with uncertainty. However all these strategies were based on the following assumption: During the operation stage all values of uncertain parameters can be obtained exactly at each time instant . This assumption is very restrictive since relevant sensors may not be available and/or hundred percent accurate. Therefore it is fair to say that the assumption is often not satisfied in practice. ^ In this dissertation the optimization of chemical processes under uncertainty in the case of insufficient process data at the operation stage will be tackled in a rigorous framework. In connection with this we postulate the presence of three groups of uncertain parameters. This leads to the need to develop an entirely new set of flexibility analysis sub problems. ^

COinS