Page 62 - Design of Simple and Robust Process Plants
P. 62

46  Chapter 3 Design Philosophies
                 conceptual design of a process plant will ultimately not be solved as one problem,
                 on the basis that:

                   .  The alternatives for the design will still be subject to the inventiveness of the
                      designer. There is no doubt that the quality of the synthesis tools including
                      generation of design alternatives will improve, but (as learns the examples
                      for simplification in Chapter 5), there is still a long way to go.
                   .  Designers always want to keep track of the decisive steps in an optimization
                      effort. They tend to follow a step-wise approach, with intermediate validation
                      of results and reconsideration of active constraints.
                   .  Optimization software is currently not sufficiently robust to solve major
                      optimization problems, nor to include the effects of integration and controll-
                      ability.
                   .  Designers will (in agreement with business) always try to minimize the risk
                      of a design, and therefore often take a more conservative route. This does not
                      mean they do not want to quantify the best design ± but they may decide on
                      another route.
                 Based on the above conclusion, process synthesis have to be carried out at different
                 hierarchical levels. The detailed levels that are discussed follow an interactive onion
                 model, which is based on the conventional onion model of Smith and Linnhoff,
                 1988.
                   Between these different hierarchical levels, interaction and process simplification
                 need to take place in order to optimize the total plant design, as is illustrated in
                 Figure 3.3.
                  The optimization will be done in a layered approach. It starts with the evaluation
                 of many process alternatives. The number of alternatives will gradually be reduced
                 while the modeling details will enlarge. Ultimately one overall process flowsheet




                                                  Site integration
                          Utilities                Controllability
                        Heat exchange               Integration
                         Separations/               Separations/
                          Recycle                    Recycle
                                                                 Idea generation
                          Reactor                    Reactor      Optimization
                                                                  Interaction




                      Conventional                   Interactive

                 Fig. 3.3. The interactive onion model versus the conventional onion model.
   57   58   59   60   61   62   63   64   65   66   67