Determining Oil Field Recovery by Using Analog

#recoveryfactor #reserversestimation #analog #oilfieldrecovery #prediction #estimation

One of the challenges in determining the recoverable volume from a new field, is we do not have production data to do prediction. Therefore, one of the common ways to do this estimation, besides doing a very detailed modelling work, is to look at nearby reservoirs or fields which have similar characteristics and look at their performance. By doing so we can use the performance to make an estimate of the recovery of the new field based on the expected reservoir conditions and parameters. Let us demonstrate on how we can do that.


Walkthrough


Firstly, we need to get the data. We can use the handy Sample icon and retrieve the recovery factor dataset.


Click Sample and Retrieve this Dataset

So here, we have an analog dataset of reservoir properties, rock properties and fluid properties from 150 reservoirs along with its recovery factor. The inputs are: STOOIP = Stock Tank Original Oil In Place, net pay interval, porosity of the formation, initial water saturation, permeability of the formation, API of oil, viscosity of fluid and initial reservoir pressure. We can use this dataset to predict the recovery factor based on the inputs that AiSara learned from.


Let us take out one row of data to predict the recovery factor. In this case, we took out row 41, which has the following reservoir properties and conditions:

  • STOIIP = 184 MMSTB

  • Net pay = 469.8 ft

  • Porosity = 0.12

  • Initial Water Saturation = 0.24

  • Permeability = 35mD

  • API = 33°

  • Viscosity = 0.25 cp

  • Initial Reservoir Pressure = 8406 psi

This reservoir produces light oil, has moderate initial reservoir pressure, as well as has low permeability and porosity. Knowing how our reservoir parameters look like, we can insert the reservoir properties and obtain the prediction from AiSara. The test data is placed somewhere accessible, so we can get back to it soon.

Use the learn function and let AiSara observe the pattern between the variables with the recovery factor.
Click Learn and simply select your input, output and an anchor cell

Then we can proceed on using the predict() function on the reservoir properties and conditions as listed previously.

It is also worth noting that the results depend on the quality of the data.

AiSara functions significantly better with good data. Eventhough there could potentially be more input data to determine recovery factor with AiSara, we continue to use this dataset for demonstration purposes.

By using the predict function(), we can estimate the recovery factor of the reservoir with the displayed inputs

Bearing in mind there are more possible factors that could contribute to the value of recovery factor in an oil field. For example, current contacts, the temperature of the reservoir, formation type, formation volume factor of oil and much more. There are also economical, operational and geographical factors that can further determine the recovery factor of a field development project. Such as the technical limit to produce probable reserves due to harsh weather conditions to operate in deep sea offshore.

However, for this article the recovery factor is focused on the reservoir volume. Let us proceed to the results of the recovery factor that AiSara has predicted.

OBJECTIVE ACHIEVED


AiSara managed to predict the recovery factor of a reservoir to be 29.62%, whereas the actual recovery factor is 31.7%. This is a 6.56% prediction error.

We hope this article has been beneficial for you. Apart from AiSara’s capability to process analog reservoir data to determine the recoverable volume of a new field, you can imagine how AiSara can offer reliable estimations across various line of work.

71 views

No. A-8-3, Floor 8, Block A,

Megan Avenue 1,

 Jalan Tun Razak,

50400 Kuala Lumpur

  • Twitter
  • Instagram
  • Facebook
  • YouTube
  • LinkedIn