Development scheme for simulation studies

In the validation phase the behaviour of the simulation model is compared with the real system s behaviour. If errors are revealed, the modelling and programming steps have to be re-started to ensure the required level of accuracy. 

All sub-processes of the development process, beginning with the perception of deficiencies, require the analysis of process data. The data characterizes the system s performance and behaviour. This is necessary to parametrize the simulation model and build a credible basis for comparing the model s behaviour and reality. A common database maintained by both modellers and managers allows linking necessary and available information. A lack of data may lead to uncertainty, more complexity, and a loss in accuracy. Uncertainty is reflected e.g. by stochasticity at variables which in turn may lead to a larger number of simulation experiments and more volatile performance measures. On the other hand, the information collected in the project database needs to be relevant for the project. Otherwise, a flood of data to be evaluated and analysed hinders an efficient development process as it is time and resource consuming. 

The analysis of available and relevant data typically requires a set of statistical methods. They are applied to extract deeper knowledge about the processes to be modelled. The set of methods is vast and the choice of methods depends on the needs of the speciflc project. Time series methodology is one prominent branch of methods. 

The following example describes the conceptual model of an artificial chemical SC analysed in this section. 

Both production sites produce basic and intermediate chemicals based on Naphtha. However, both sites focus on different final products. They also vary in their plants capacities. The Naphtha is. steam-cracked at both sites with similar plant configurations. Hence, the ratios of output flows are. similar but the heights of the flow rates are not. 

---