Computational fluid dynamics calculations for gas explosion safety have been widely used for doing risk assessments within the oil and gas industry for more than a decade. On the basis of predicted consequences of a range of potential accident scenarios a risk level is predicted. The development of applications using hydrogen as a clean energy carrier has accelerated in recent years, and hydrogen may be used widely in the future. Because of the very high reactivity of hydrogen, safe handling is critical. For most applications it is not realistic to perform an extensive risk assessment similar to what is done for large petrochemical installations. On the other hand, simplified methods, like venting guidelines, may have a questionable validity for hydrogen. The use of simple methods, if these actually are conservative, will in general predict too high consequences for the majority of scenarios, as these are not able to represent actual geometry and physics of the explosion. In this article a three-step approach is proposed. The initial approach will be to carry out a "worst-case" calculation evaluating the consequences if a full stoichiometric gas cloud is ignited. Mitigation measures can also be considered. As a second step, if potential consequences of the initial approach are not acceptable, the assumptions are refined and more calculations are performed to make the evaluations more realistic and reduce unnecessary conservatism of the chosen worst-case scenarios. Typically a number of dispersion calculations will be performed to generate likely gas clouds, which are subsequently ignited. If estimated consequences are still not acceptable, a more comprehensive study, including ventilation, dispersion, and explosion, is performed to evaluate the probability for unacceptable events.