Good day Ahmed,
Let's step back so I get your point.
If you History Matched the OOIP this is not called history match but initialization of the model.
The volumes expressed are compared with the static model volumes usually and a tolerance of 5%is considered acceptable between the two.
The real history match starts when you run your model with the real data sets in production/Injection schedule (usually in schedule the keywords will have an additional H at the end).
Then you see whether your model is able to reproduce this peace of work.
The next step in HM is to plot the historical data versus the calculated ones:
- oil production, water cut and GOR
- pressure trend vs. Time (measured data represented as points/dots)

With the 4 plots in hand you start analysing how model matches or not the reality. The issues are usually water breakthrough mismatch and pressure off trend.

The real work will start as you'll have to find reasons and how to overcome the mismatch (missing volume, PI of the well, barriers, rel perm curves. ..etc )

CAHM (computer assisted history match - like MEPO) is a tool that nowadays is common to use, I am not that fond of as this use a kind of statistical approach to vary the Reservoir parameters in a so called "objective function" that you tend to minimise to have the best match but this has to be used carefully otherwise you end up with a non realistic yet perfectly matched model -> my point here is yes I believe in statistics and yes with a random infinite number of trials you can match almost anything but then ask yourself is this physically possible or not take my advice of the day (which actually is from a famous statistician) "if you torture the data long enough it will confess to anything"

So... good luck

Envoyé de mon SM-A500F en utilisant Tapatalk