Machine learning-driven production allocation

NavPort’s patented technology fixes ‘disappointing and unreliable’ production allocation models.

‘Data intelligence’ service provider NavPort has been testing its machine learning derived technology on the thorny problem of production allocation. Production allocation involves the estimation of individual well’s production from what is in general a partial set of flow information from tests, meters, tank levels and other secondary sources. NavPort reports that many Texas wells suffer from a lack of data and these make most production allocation models ‘disappointingly inaccurate and unreliable.’ Moving from lease-level production reporting to single well production is error prone. Enter NavPort’s patented technology that combines machine learning and completion-related variables such as completion date, frac design and proppant selection.

Trials on a client-supplied set of well data compared the actual production with the ‘completion-based’ model. Three months of production data from three wells showed ‘only a 5-8% variance’ from the actual production.

Click here to comment on this article

Click here to view this article in context on a desktop

© Oil IT Journal - all rights reserved.