For the creation of an experimental database related to physical phenomena relevant for LWR containment safety, tests are performed in MISTRA (CEA, France) and PANDA (PSI, Switzerland) facilities in the frame of the OECD/SETH-2 project. The specific purpose of these tests is to obtain data suitable to improve and validate advanced Lumped Parameter (LP) codes as well as codes with 3D capabilities with respect to the prediction of post-accident containment thermal-hydraulic conditions. The experimental data is related to hydrogen transport within containment compartments. In particular, the effect of mass sources (the release of steam and hydrogen), heat sources (hydrogen-oxygen recombiner), and heat sinks (condensation of steam caused by containment coolers and sprays or "cold" wall) on the break-up/erosion of an initially gas stratified configuration characterized by a layer with a high hydrogen content. Helium is used to simulate hydrogen in the PANDA facility. This paper presents the result of a series of SETH-2 PANDA tests attributed to "vertical fluid release" (plumes or jets). Two large containment compartments (∼180 m3) connected by a bended pipe of ∼1 m diameter are used for these tests. For all the tests, a helium-steam mixture having a thickness of 2 m is created in the upper volume of one compartment while the remaining volume is filled with steam. During the tests, steam jets or plumes are created by injecting steam from a vertical pipe located at the center of the vessel 2 m below the helium-steam mixture. The jet or plume is initially positively buoyant and becomes negatively buoyant once it reaches the helium-steam layer. These transient tests show the degradation of the helium-steam layer for different jet Reynolds numbers. The initial Froude number at the injection pipe varied in the range of ∼3 to ∼9, while the estimated Froude number at the helium-steam mixture/steam interface varied from ∼0.70 to ∼2.