In this simulation I've tried to optimize the cooling methane flow by introducing several baffles. This makes the heat transfer from the shield to the cooling methane more efficient "trapping" the methane longer in the shield jacket. As a result the fresh 100K cold methane gas is heated up to 330K, which is about 7 KW cooling power in a 50 mm shield slice. If we approximate this to the full Starship length we get minimum necessary cooling power of  about 8400KW. Same reentry conditions - V=6000m/s and Patm.=15 Pascals. If we consider the methane liquid-vapor phase transfer energy I think that 10 tons of methane should be enough for a 15 min. long reentry cooling. However this is on optimistic side and 15 tons methane should provide enough safety margin.