The techniques for sensitivity analysis employed in the field of building energy performance can be classified into local and global methods [7–10]

The techniques for sensitivity analysis employed in the field of building energy performance can be classified into local and global methods 7–10. Local sensitivity analysis belongs to the class of the one-factor-at-a-time (OFAT) methods 7. The OFAT method studies the behavior of model outputs with the changes in model inputs. According to this method, each time one design variable is changed over its entire range, while all other parameters are held fixed at their initial mode. This technique can be repeated with other variables 9–12. It should be noted that OFAT method is often evaluated using gradients or partial derivatives of the output variables when the function of outputs is available 13. Global sensitivity analysis methods evaluate output variations due to one input variable by varying entire parameters over their over their whole input space at the same time 14–17. Thus, global methods assess the interaction of variables and provide robust sensitivity measures for the nonlinear models. The variance-based method is an appropriate sensitivity analysis method for complex nonlinear and non-additive models 7.
OFAT sensitivity analysis methods are straightforward to establish parameter dependency of the solutions, and useful to study problems with a few uncertain parameters. However, these methods only explores the variation of energy performance around a single point or a base case, and the interaction of other design parameters are not considered 7,10,13. In addition, the fatal limitation of local sensitivity analysis methods is that they are unwarranted when the model input is uncertain or when the model is of unknown linearity 14,15. In this respect, local sensitivity analysis methods give quite misleading results for nonlinear models. On the contrary, global sensitivity analysis approaches use a representative set of samples to explore the design space, which provides robust sensitivity measures in the presence of nonlinearity and interactions among the parameters compared to local sensitivity analysis methods. As a result, global methods are considered as more reliable and precise techniques for building analysis 7. The main drawback of global sensitivity analysis methods is their higher computational costs than local methods as complexity and number of design parameters increase.