Data analytics for oil sands subcool prediction — a comparative study of machine learning algorithms. Issue 18 (2018)

• Record Type: Journal Article
• Authors: Li, Chaoqun; Jan, Nabil Magbool; Huang, Biao
• Is Part Of: IFAC-PapersOnLine. Volume 51:Issue 18(2018)
• Page Start: 886
• Page End: 891
• Publication Date: 2018
• Subjects: data analytics -- deep learning -- process application -- machine learning -- SAGD -- subcool; Automatic control -- Periodicals; 629.805
• DOI: 10.1016/j.ifacol.2018.09.234
• Languages: English
• ISSNs: 2405-8963

Abstract

Abstract: Steam Assisted Gravity Drainage (SAGD) is an efficient and widely used technology to extract heavy oil from a reservoir. The accurate prediction of subcool plays a critical role in determining the economic performance of SAGD operations since it influences oil production and operational safety. This work focuses on developing a subcool model based on industrial datasets using deep learning and several other widely-used machine learning methods. Furthermore, this work compares and discusses the out-of-sample performance of different machine learning algorithms using industrial datasets. In addition, we also show that care has to be taken when using machine learning algorithms to solve engineering problems. Data quality and a priori process knowledge play a role in their performance.