The capability of the RELAP5/MOD3 computer code to analyze water hammer transients due to water column rejoining and a water slug propelled by non-condensable gas is investigated. The code-calculated results have been compared with those obtained from simple ideal analytical models. Good agreement is obtained between the calculation and analytical results in the initial period of the transient during which the water column or slug retains its sharp interface and suffers from little breakup or dissipation. As the transient proceeds, the code-calculated hydrodynamic loads are generally less than those implied by the analytical models. This is most likely due to the breakup of the water phase, which is not taken into account in the analytical models. Effects of time step and mesh sizes have also been studied. The results show that the usual Courant time limit applies. Finally, a sample calculation, corresponding to a water hammer transient in a typical Westinghouse four-loop reactor head vent system piping, is presented. The transient is induced by the opening of a relief valve and accelerating a trapped water slug through the pipeline. Hydrodynamic loads (i.e., force-time curves) on various pipe segments have been evaluated by appropriate postprocessing of the transient results. The calculated peak forces at selected pipe segments compare favorably with those estimated from the analytical models.