Hi, I have a problem with convergence in my reservoir simulation model.

I have one well in a thin oil column. The well cones in gas and water.
The well rates are controlled by either liquid rate, or by tubing head pressure and lift curve.

Linear convergence errors occur mainly when both gas and water are coned in, less when only gas is coned in, and least when only water is coned in.

I would be greatful for any input on how to improve the convergence.

Regards,
Abulfaz