Simulation based optimisation

The benefits captured by optimisation are very process-specific. They can match or exceed those captured by constraint control or they can be a very minor addition. The case illustrated in Figure 13.10 is one where a rigorous simulation-based optimiser would not be justified but perhaps a much lower cost empirical technique, capturing a portion of the available benefits, might be worthwhile. Since the choice of technology and implementer are likely to be influenced by the scope of the project then it is important to decide upon this first. 

Chapter 10 discusses the simulation-based optimisation experiments which are undertaken using the MOGA tool under parameterised operational strategies, which investigate the impact of the coordinated management mechanism on SCR... 

This chapter discusses research methods. The multiple case study method is reviewed and is followed by an assessment of the appropriateness of a case study approach. The data collection methods in general and the rationale for using interview and observation in this study to collect primary data are addressed. Archived data are outlined and the issues of reliability and validity are considered. Finally, simulation and simulation-based optimisation methods are introduced. 

In this research, due to the complexity of the SC system, the simulation-based optimisation method is employed to deal with decision making issues. The SC system and the GA optimisation programs are described in Chapter 7. In the next chapter, data collection and the background to the case companies are introduced in accordance with the designed research method. 

In this work, new developments were achieved through the use of new examples, one of which the optimisation of a real crude distillation unit involving 19 decision variables. The performance of the metamodel-based optimisation is compared with results obtained with the optimisation based on a first-principles model, embedded in a sequential-modular process simulator. It is shown that metamodel-based optimisation with adaptation of the metamodels during the optimisation procedure provides results with good accuracy and significant reduction of computational effort. The performance comparison between neural networks and kriging models for chemical processes is another contribution of this work. 

The study demonstrates that particle swam optimisation is a powerful optimisation technique, especially when the objective function has several local rninirna. Conventional optimisation techniques could be trapped in local minima but PSO could in general find the global rninimrun. Stacked neural networks can not only given better prediction performance but also provide model prediction confidence bounds. In order to improve the reliability of neural network model based optimisation, an additional term is introduced in the optimisation objective to penalize wide model prediction confidence bormd. The proposed technique is successfully demonstrated on a simulated fed-batch reactor. 

A model reduction-based optimisation framework for large-scale simulators using iterative solvers... 

A novel gradient-based optimisation framework for large-scale steady-state input/output simulators is presented. The method uses only low-dimensional Jacobian and reduced Hessian matrices calculated through on-line model-reduction techniques. The typically low-dimensional dominant system subspaces are adaptively computed using efficient subspace iterations. The corresponding low-dimensional Jacobians are constructed through a few numerical perturbations. Reduced Hessian matrices are computed numerically from a 2-step projection, firstly onto the dominant system subspace and secondly onto the subspace of the (few) degrees of freedom. The tubular reactor which is known to exhibit a rich parametric behaviour is used as an illustrative example. 

A Model Reduction-Based Optimisation Framework for Large-Scale Simulators Using Iterative Solvers... 

Unphysical quenching rate is not the only limitation of MD. Since the potential used enable computations of only central forces, it is suitable for simulations of glasses, which are significantly ionic. It is also successful for the simulations of metallic glasses where use is made of optimised pseudo potentials obtained from first-principle calculations. But in largely covalent materials, MD cannot be of much use imless suitable effective potential functions are developed which take care of non-central nature of the forces as well. In the next section we discuss further advances in MD simulations based on the use of quantum mechanical calculations, which optimise the local geometries and therefore provide more accurate simulations of structure. 

The potential roles that computer simulations and software have in the choice, sizing, simulation and optimisation of solid/liquid separation equipment are shown schematically in Figure 5.6. Despite the relative complexity of the flowsheet, it should be realised that Figure 5.6 represents only some of the potential scenarios that an end user may encounter. The paths indicated differ mainly in the degree of experimentation and use of computer software such as FDS the latter is based upon the concepts proposed by Wakeman and Tarleton (1991a, 1993), Wakeman (1995) and Tarleton and Wakeman (2003, 2005 a,b) ... 

In this paper a dynamic model for a semibatch polymerisation process was presented. It was validated with experimental data from the industrial site and used for simulating the process. The simulation results show that the model can adequately describe the process and therefore constitute the base for the optimisation. The flowsheet simulator CHEMCAD has proved an efficient and powerful tool for the modelling, simulation and optimisation of semibatch polymerisation processes. 

The simulations based on this model make it possible to compare multiple scenarios on the basis of a group of performance and cost indicators. Consequently, they provide important and reliable decision-making support for maintenance optimisation. 

The issues of evolutionary library design in the frame of material and catalyst discovery were described. Concepts of diversity management on material library to enhance the efficiency of the optimisation were proposed. The diversity monitoring is implemented via two different approaches. The first deals with a dynamic monitoring of mutation and crossover rates, whereas the second involves a selection step based on sample distance . Simulations of optimisation were performed on a surface response, which was designed to mimic realistic data. Algorithm performances were compared in terms of both efficiency and reUability. 

SiNx/Au/Si02), similar to the device in Fig. 16 C [49] Simulations based on a slightly more refined model than the one above are shown on the right (panel B), and use, i.a., experimental data for i ct( icR> 2cr) and a least-square-fit optimised Rpore (see [49] for further details). 

Jorgensen et al. has developed a series of united atom intermolecular potential functions based on multiple Monte Carlo simulations of small molecules [10-23]. Careful optimisation of these functions has been possible by fitting to the thermodynamic properties of the materials studied. Combining these OPLS functions (Optimised Potentials for Liquid Simulation) with the AMBER intramolecular force field provides a powerful united-atom force field [24] which has been used in bulk simulations of liquid crystals [25-27],... 

---