[olrac]

Hi everyone, I'm trying to simulate PCM melting in a parallelepiped enclosure (100x20x100 mm) heated from the side (100x100 face) with a costant heat flux. Since it seems a natural convection problem to me, I've activated gravity in the Y direction and used the Boussinesq approximation. The liquid density is 760 kg/m3 at 80°C and the solid density is 880 kg/m3 at 15°C, melting temperature being 41-43 °C. In order to use the Boussinesq approximation I've used a mean density and under "operating conditions" a mean temperature. I've tried different mesh size,as well as different timestep and iterations per time step. Other settings are as default. Evey time I get weird results, as you can see in attachment #2: when the liquid fractions get to 0.9, the PCM just melts instantly. If I use costant density, the I get small steps as you can see in attachment #3, so a less smooth rise but at least I don't get that jump at 0.9. Can you help me? Thank you