Practical Aspects of Flowsheet Optimization

When solving a NLP, to optimize a flowsheet, still another alternative exists. In many cases, it is preferable to incorporate the design specifications as equality constraints, i, = 0, as shown in NLP3. Then, it is necessary to remove these design specifications when adding the optimization convergence unit. The latter usually replaces the recycle convergence units in the simulation flowsheet. 

Before implementing an infeasible path optimization, it is very helpfiil to carry out preliminary searches by varying the key decision variables, somewhat randomly, to gain insights into the key trade-offs. For these searches, it is probably best not to use optimization algorithms that require derivatives, or approximations to them, such as SQP. A common approach is to use the sensitivity analysis facilities of the process simulators referred to earlier. 

---