Management of Integrated Production System Models (IPSM)

I was recently asked to provide some thoughts on a forum on the management of Integrated Production System Models (IPSM). I thought I would share based on my experience leading the roll-out and improvements of Petex IPM models in Shell.

How to establish that a model is fit for general use?
There should be some criteria, for example I said that a model is calibrated if all the wells are withing +/- 5% of their actual rate. And overall, the model is within +/-10% of the actual rate.  You need to establish a rule like this, based on the questions that the model is going to answer, i.e. what decisions will be made with the model.  Ask if your decision would be different if the model was more or less matched to the current conditions.

How often do you check and calibrate/update a model?
Calibration is different from updating.  If new wells are online or old wells off line, or different equipment is installed, this needs to be captured on a regular schedule.
Calibration should follow the rules above. If the rate from the model is close enough, no change is necessary.

What do you change in order to calibrate the model?
Have a fixed and unchanging method for calibration.  For example, use IPR Shift in GAP to match the rate, given that you don't really know reservoir pressure.  Change the friction factor in the pipelines given that you may build up more liquids or scale in the pipes.  Don't go changing randomly each month or you won't know what is going on or how to get back to something more sensible.

Who is responsible for updating what part?
Reservoir Engineers are responsible for reservoir pressures, WC, GOR, WGR, CGR and the match of these with cumulative production. Their monthly updates need to include updates of cumulatives for each reservoir so forecasts can start from the right point.
Production Engineers are responsible for inflow productivity, artificial lift values, THP, choke dP, and updates to well equipment .  Sometimes they are responsible for overall calibration and management of the model.
Facility Engineers are responsible for calibration of flowlines, compressors, and routing of flow.  They may have less work than most in the update cycle since less changes in this area.

How do you share models and control updates?
Please don't put large folders of models on  a shared network.  Especially GAP will run slowly as it has to access all the files over the network where read/write speeds are significant compared to its calculation time.  Also, what happens if someone inadvertantly breaks a model or erases it?  What happens if two people update the same model at the same time?
Better idea: Put the model on your own local drive. Create a GAP Archive file (.gar) and put that on the network.  When you need a fresh version, copy the .gar file to your machine.  Update it and put a revised .gar file back.  
Best idea: Invest in IFM. IFM comes with Model Catalog which archives multiple versions and controls access to one person at a time.

How do you minimize rework in the update cycle?
It is best to have one person who is in charge of the official model version.  Once per month (or whatever is decided) they coordinate the update cycle, starting with the Resevoir Engineer's updates, then the Produciton Engineer's, then the Facility Engineer's.  Thus, the PrE has the right reservoir pressure, WC, etc. to calibrate the well model.  And the Facility Engineer has the right rates on each trunkline to calibrate the pressure drops.

How complicated should a model be?
A model should be no more complicated that required to help make the necessary decisions for which it is constructed.   A simple model is one that requires less effort and is more often ready to be used.