Integrated Online Emulsion Management System

Abstract

A full suite of integrated online emulsion management system (IOEMS) transforming the handling of decades old crude oil emulsion production issue at field from reactively onsite to proactively online. This technology is made possible with insights on emulsion formation from physics-based molecular models, access of huge database on crude oil properties, emulsion toughness and demulsifier chemistries, coupling with statistical and supervised machine learning application. Intriguingly, this innovation journey began with designing an enhanced oil recovery (EOR) technology in mind. Study on generating stable emulsion for oil recovery was the aim of our pioneering research initially. We successfully developed physics-based models to assess emulsion stability at molecular level. We then applied these models retrospectively for produced emulsion management, with advancement in data science and computational power. The technology concept is to design and plan demulsification strategy based on predicted emulsion stability. The robustness of IOEMS lies in the combination of the goods of accurate interpolated data based on machine learning, with that of extrapolated data from physics-based model. Firstly, mathematical models of relationships between crude properties and emulsion stability index (ESI) were established using statistical method. This led to a good 90% match with laboratory ESI data. Secondly, a demulsifier selection functionality was developed based on machine learning, covering dozens type of demulsifier. We used operating conditions, fluid and demulsifier properties as training data input, with the corresponding lab bottle tests outcomes as training data output to build a classification model via supervised learning algorithms. Its predictive accuracy is at 87%. By bringing the produced emulsion assessment from on-site to online, offshore emulsion sampling and the associated lab bottle tests are minimized. Health safety and environment (HSE) risks are reduced accordingly with the decrease of human intervention in field sampling. The emulsion stability predictive functionality enables operation to prepare early in anticipation of sudden spike of emulsion production and thus, avoiding unplanned well shut in. Furthermore, this function is especially useful when emulsion samples or historical data are not available during field development stage. Meanwhile, the recommended demulsifers from IOEMS are at 17% lower cost than the incumbent demulsifiers used at fields in Malayia, in addition to 90% manhour reduction from conventional trial and error demulsifier screening in lab. Ultimately, the IOEMS has successfully enabled step-change in oilfield emulsion management via an efficient and reliable scientific based digital platform.