Sensitivity Analysis in the Simulation of Complex Solids Processes

In the present work a method to quantify the range of uncertainty of selected stream values and to detect the most influential parameters within a given flowsheet has been developed for the flowsheeting of solids processes. By using heuristic optimization methods the range of uncertainty of relevant stream values can be determined. Distributions of attributes of solids can be handled and shape and location of distributions as well as characteristic parameters like the Sauter diameter can be predicted. Furthermore the most infuential parameters for the units or the models in the range of uncertainty can be detected by investigating the difference quotients in different operating points. With this information the time required to improve the accuracy of prediction of the simulation results can be reduced significantly. As an example the simulation of a hydrocyclone is presented. The influence of the parameters of an empirical separation model on the solids mass flow collected with underflow is demonstrated and the range of the product particle size distribution depending on uncertain parameters is examined.