Sensitivity analysis studies how the output variables are affected by input variables of a
statistical model. In other words, this model systematically changes the inputs of a statistical
model so that the effects can be predicted in the output. Using this method, the
effect of different variables including key variables in the model was determined. In this
project, 9 independent variables were evacuated for sensitivity analysis of hydrogen
release in the environment. These variables are divided into process variables (supply
temperature, pressure, leakage diameter, and height of release point) and environmental
variables (temperature, wind speed, pressure, humidity and surface roughness). In this
study, fractional factorial experiment design 293
IV and analysis of variance were used to
specify the conditions of different scenarios and statistical analysis of results for 68 scenarios,
respectively. All results were examined in stable conditions F and D. To consider
the possible extreme points in results, 4 scenarios were considered for central conditions.
Results showed that the model is more sensitive to process variables than environmental
variables. Leakage diameter, supply pressure and supply temperature have the highest
impact on the rate of discharge while supply diameter and surface roughness have the
highest impact on dispersion.
Copyright © 2016, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights