Spe-192002-ms case Study Applied Machine Learning to Optimise pcp completion Design in a cbm field

SPE-192002-MS
Case Study Applied Machine Learning to Optimise PCP Completion Design
in a CBM Field
Charles Prosper and David West, Origin Energy
Copyright 2018, Society of Petroleum Engineers
This paper was prepared for presentation at the SPE Asia Pacific Oil & Gas Conference and Exhibition held in Brisbane, Australia, 23–25 October This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the authors. Contents of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the authors. The material does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.
Abstract
The key objective of multiple Coal Bed Methane (CBM) development operations is to determine cost effective methods to allow sustainable economic production and maximum reserves recovery. The cost of workovers as well as the associated deferred production may overwhelm the economic viability of the field.
The primary reason for workovers is progressive cavity pump (PCP) system failures.
Here, we demonstrate the use of a machine learning framework that can be used to customise each workover configuration such that it optimises PCP run-life, respecting the well's heterogeneity and age.
The framework can be generalised into three major parts 1) converting the dynamic production data into a stationary surrogate model for the well 2) the use of Gaussian process regression to create a function that estimates runlife; 3) an optimiser that will search the functional space to recommend the best completion design.
A telemetry and completion dataset for PCP run-lives from years 2014-2018 was obtained across the
Surat and Bowen basins. After filtering data for completeness, 1499 PCPs remained in the cohort, of which failed during the observation period. A small portion of the original data was used as a test set.
Our work suggests that PCP run-life can be extended by taking a multivariate statistical approach to provide recommendations for customised completions and production strings per well that respect the wells’
geology and production history and thereby improve life of field economics.