What data is necessary for WellDone?

The minimal required set of data includes historical production rates, locations of wells relative to each other, and injection rates, in case if waterflooding is performed. Besides, historical and planned values of pressure at the bottomholes of production wells and planned injection rates are needed to perform history matching and prediction. All other pieces of data are optional (performed workovers, surface facilities, etc.).

Can WellDone predict the type of necessary well workover?

Yes, it can. First, the 2-level classifier identifies the wells that require workovers to be performed. After that, different types of workovers are ranked for each of the selected wells, depending on the identified probability of their effectiveness.

Is it possible to use WellDone for offshore fields?

For now, we focus on oil, dry gas, and gas condensate onshore fields. However, offshore fields might be the case as well.

What does the accuracy of the forecast depend on?

The accuracy of the forecast depends mostly on the quantity and quality of the input data. In addition, the performance of the adequate well clustering and the correct choice of the predictive model contribute to the quality of the obtained predictions.

Can the system work with new fields?

Yes, it can. In WellDone, a user can find both models that require training for the current project and pre-trained models. Models that require training adjust to the uploaded field data, while for pre-trained models to be used in different cases training samples have been compiled the way covering a wide variety of field types.

Can the system be used for gas fields?

Yes, it can. Separate models are provided for gas fields that consider the physical processes associated with the gas inflow to the well.