changes that let the problems run longer, the cyclic example features now a sin-curve in injection rather than a hard on/off. Numerical robustness is affected by strong changes in pressure and temperature, mainly when the steam chamber is suddenly reduced in size due to a stop of injection. Condensation leads to strong volume changes, primary var-switches, etc. The same holds, when injection recommences. Steam re-exists quickly and demands for volume. The IAPWS properties seem to be the Achilles heel, since validity limits are exceeded often or regions are changed suddenly.
The scenarios have another crucial point to pass. That is when the front breaks through at the production well and well cooling sets in. The sudden cooling also leads to sudden strong changes.