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



Download 1.89 Mb.
View original pdf
Page2/8
Date22.08.2023
Size1.89 Mb.
#61902
1   2   3   4   5   6   7   8
spe-192002-ms
Introduction
Coal Bed Methane (CBM) is a prominent source of energy. The mega CBM to LNG projects in Australia require thousands of wells as feedstock. Fields of this scale pose different problems to their conventional cousins namely due to the sheer well count and areal extent of the fields. Maintenance of such a large fleet of wells, many of which require artificial lift, poses a challenge for CBM operators.
Typical machine learning (ML) algorithms require millions of data points for successful pattern discovery and prediction. This is in stark contrast to the available hundreds to thousands of complete runlife periods available over the life of a CBM field. Here again, the production periods are from inhomogeneous pockets of coal seam at different stages of depletion.
Downloaded from http://onepetro.org/SPEAPOG/proceedings-pdf/18APOG/2-18APOG/D021S016R002/1220497/spe-192002-ms.pdf/1 by Vedanta Limited - Cairn Oil & Gas user on 28 June 2023


2
SPE-192002-MS
The purpose of this paper is to propose a representation and optimisation framework in CBM wells to recommend an optimal completion design. This method is novel and has the following advantages. Surrogate well model representation from time series production data. The representation is stationary and reflects key features of a coal seam gas well such as tank volume, decline and gas liquid ratio. Use of a Gaussian Regression Optimisation framework that provides both a runlife estimate and an uncertainty estimate with a low number of observations. Utilise the exploration and exploitation of the regression function to design new completion types that have not been tried before and also recommend completion designs that offer the most runlife with the lowest uncertainty.
PCP Pump Failures
There are many causes of PCP failure and table is comprehensive list is managed by C-FER which is a joint industry project created to enhance the performance of PCP pumps.
Table 1
Downloaded from http://onepetro.org/SPEAPOG/proceedings-pdf/18APOG/2-18APOG/D021S016R002/1220497/spe-192002-ms.pdf/1 by Vedanta Limited - Cairn Oil & Gas user on 28 June 2023


SPE-192002-MS
3
The focus of this modeling approach is isolated to the PCP design given the reservoir fluids and completion type. This is a practical consideration given that the operator is only able to choose the best pump design once the well is drilled and completed.
A central premise of this modeling approach is that pumps operating in the same environment will have similar runlives. However, the operating environment introduces variability in runlife of PCPs, such as the differential pressures across the pump, intake pressure, flowrates, gas liquid ratios, gas volume fractions,
solids volume fractions, pump off control, and water chemistry.

Download 1.89 Mb.

Share with your friends:
1   2   3   4   5   6   7   8




The database is protected by copyright ©ininet.org 2024
send message

    Main page