A Guide to Uncertainty and Global Sensitivity Analysis in Lumped-Parameter Models of the Cardiovascular System

Abstract

In this paper, a general 5-steps framework of uncertainty analysis (UA) and sensitivity analysis (SA) in lumped parameter models of the cardiovascular system (partial or complete) is presented. In order to conduct proper UA and SA, a high number of model simulations is required. Therefore, lumped parameter (0D) models of the cardiovascular system (CVS) are suitable for UA and SA as compared to high or multi-dimensional models (3D,2D,1D). As an example, a linear elastic lumped-parameter model of arm arteries is considered and a 5-steps framework is applied to quantify the impact of input parameters on output variables. The framework uses the method of Sobol which is easy to implement and computationally efficient as compared to other methods, like FAST, eFAST, Sparse grid method, etc. Additionally, the framework algorithm could be applied on all types of compartment models present in different fields of science and technology.