1.) Geological & Seismic Data : Reservoir “Architecture” & 3D framework.
2.) Petrophysical Data: Rock & fluid Properties and interaction of fluids with rocks.
3.) Fluid Flow Behavior: Forces operating the reservoir and there effect on fluid flow.
4.) Well construction & Completion: what you observe at the surface, maybe is completely different then what it looks like at subsurface. (In engineer’s language: effect of pressure & temperature)
5.) Reservoir management: Effect of pressure depletion on all of the above properties.
Questions that we want to answer :
1. How much Hydrocarbons are there 2. How much can we recover (Recovery Factor)
2. How much can we recover (Recovery Factor)
3. Can we increase the recovery if we operate the reservoir in a certain way
And all these questions have to be answered with the use of noisy and uncertain dataset.
Use of big data has provided a much needed push towards making sense of data but oil industry which is known for its late adoption of IT technologies is still lagging behind to apply this concept. One of the major hurdle in applying this is unstructured and inaccuracy of data from different sources which has to be refined and prepared before applying any data mining techniques. And the product has to ensure the creation of new knowledge which is not an obvious process.