And another one from the crowd;
Fault Analysis

Hi Vinomarky,
Thanks very much for great contributions in this forum.
My question is this:
I have a static model which has faults in it. How do i know the transmisibility value to assign to these faults in order to determine if they are conducting or non-conducting ones. Or do i just assign any value and wait till history matching phase to determine the actual value. Or better still, conduct material balancing first to know if there is an extra energy or communication and regress the transimibility multiplier. What do you think?
You will not 'know' the transmissibility value prior to simulation - and in fact, the reality is that you probably wont know for sure after simulation either, but if done correctly you should get a better understanding as to the possible range(s)

In short, sensitizing on fault transmissibilities during history matching is the way to go, BUT I caution against going for a single deterministic match as being 'correct' and encourage you to try a number of different fault transmissibility combinations that yield similarly good matches and better understand you range of possible outcomes. If you are going to do a simulation anyway, unless your model is a beast that takes ages to run (and so you can quickly narrow your range) there is little additional benefit in regressing via MBal to start with