Introducing Data-Driven Virtual Viscosity Measurements
-
Published:2022-10-31
Issue:
Volume:
Page:
-
ISSN:
-
Container-title:Day 3 Wed, November 02, 2022
-
language:
-
Short-container-title:
Author:
Karpan Volodimir1, Al Farsi Samya2, Al Sulaimani Hanaa2, Al Mahrouqi Dawood2, Al Mjeni Rifaat2, van Batenburg Diederik3
Affiliation:
1. Shell Development Oman LLC 2. Petroleum Development Oman 3. Shell International Exploration and Production
Abstract
AbstractPolymer-based chemical flooding is a mature enhanced oil recovery technology that has proven to result in significant incremental oil recovery that is both cost and GHG emission-competitive compared to the oil recovered by conventional waterflooding. For such chemical flooding projects, controlling the viscosity of injected polymer solution is critical because the polymer cost is one of the most significant cost elements in the project economics. The polymer viscosity is routinely measured in the laboratory using fluid samples taken manually at different sampling points (i.e., polymer preparation facilities, injecting lines, and well heads). However, in the case of large-scale projects, such viscosity monitoring becomes time-consuming and requires dedicated field staff. Moreover, the quality of laboratory-measured viscosity is questionable due to the potential viscosity degradation caused by the oxygen ingress or polymer shearing during sampling, storage, and measurement. The inline viscometers were introduced to improve the reliability of viscosity measurements and have a better quality of viscosity monitoring. Such viscometers are relatively simple devices readily available on the market from several vendors. However, the device comes at additional costs and requires modifications at the tie-in point (bypass line, drainage, and (sometimes) communication and power lines). On top of it, operational costs include regular maintenance that the inline viscometer requires to ensure good data quality.This study introduces a data-driven Virtual Viscosity Meter (VVM) as a tool to augment the inline and laboratory viscosity measurements. Standard injector wells in a field are equipped with gauges that report injection rate, well/tubing head pressure, and temperature of the injected fluid. With such well data and viscosity measurements, calculating the viscosity becomes a machine learning regression problem. Training the machine learning regression methods on the actual inline and laboratory-measured polymer viscosity has demonstrated that VVM is a promising, high-accuracy solution with a low computational cost. The possibility of further implementing this approach to calculate the viscosity of an injected fluid was investigated using the data from several projects. Finally, the application of the VVM tool for viscosity monitoring and the limitations of VVM were discussed.
Reference20 articles.
1. Al Hashimi, A., Al Kindi, O., Al Farsi, S., Al Sawafi, M., Al Mawali, S., Al Abri, M., Al Hinai, S. (2022) Automation and Digitalization of Polymerflood Viscosity Surveillance Through Machine Learning– A Case Study from Marmul. Paper SPE-211015 presented at the SPE Abu Dhabi International Petroleum Exhibition and Conference. Abu Dhabi, UAE, 31 October–3 November. 2. Al Mahrouqi, D., Sulaimani, H., Farajzadeh, R., Svec, Y., Farsi, S., Baqlani, S., …Karpan, V. (2021). Advancement Towards the Full-field Implementation of Marmul Alkaline-surfactant-polymer In the Sultanate of Oman. Paper SPE-190382 presented at the SPE Abu Dhabi International Petroleum Exhibition and Conference. Abu Dhabi, UAE, 15–18 November. 3. Al-Shuaili, K., Svec, Y., Guntupalli, S., Al-Amri, M., Al-Hinai, G., Al-Shidi, M., Mohammed, S., Al-Hadhrami, H., van Batenburg, D. W., de Kruijf, S. (2018). Field Piloting of Alkaline Surfactant Polymer in Sultanate of Oman, SPE-190392 presented at the EOR Conference at Oil and Gas West Asia held in Muscat, Oman, 26–28 March 2018 4. Machine Learning and Data Science in the Oil and Gas Industry: Best Practices, Tools, and Case Studies, Elsevier Science;Bangert,2021 5. First principles and machine learning virtual flow metering: a literature review;Bikmukhametov;Journal of Petroleum Science and Engineering,2020
|
|