Fig. 1.

Working steps of sensitivity analysis are listed concisely in the diagram, and described below: (a) Choose a sensitivity analysis method based on the number of model evaluations, the correlation structure of input-output parameters and an experimental design. (b) Identify simulation input-output parameters of interest and what input factors are needed to include in the analysis. (c) Define upper and lower bound of input parameters by taking respectively 15 − 20% more and less from the mean value of each parameter and generate input sets of parameters by LHS method and Sobol sequence. An input set has the form of N strings of input factor values on which the model is evaluated. (d) Evaluate the model on the generated input samples and produce the outputs, which contain N output values in the desired form. (e) Analyse the model outputs and draw conclusions depending on PRCC and Sobol indices calculated using the outputs. (f) Based on the observation of model outputs parameter sample size is determined. More detail on sample size can be found in Appendix D.