Hybrid Inferential Modeling for Vapor Pressure of Hydrocarbon Mixtures in Oil Production

详细信息    查看全文

• 作者：Yu C. Pan ; S. Joe Qin ; Phi Nguyen ; Michael Barham
• 刊名：Industrial & Engineering Chemistry Research
• 出版年：2013
• 出版时间：September 4, 2013
• 年：2013
• 卷：52
• 期：35
• 页码：12420-12425
• 全文大小：352K
• 年卷期：v.52,no.35(September 4, 2013)
• ISSN：1520-5045

文摘

In gas and oil production processes, both the true vapor pressure (TVP) and the Reid vapor pressure (RVP) are important and, yet, difficult-to-measure physical properties. Many upstream process units, such as floating roof tanks, have specifications given in terms of the TVP. Careful estimation and control in terms of the RVP will ensure that no extra gas other than necessary is removed from high-valued oil. Laboratory analyses alone are not effective for this purpose because they are time-consuming and involve large time delays. On the other hand, online analyzers are expensive and difficult to install and maintain. In this paper, we propose a hybrid inferential sensor based on a combination of the modified vapor pressure equation proposed by Korsten [Ind. Eng. Chem. Res. 2000, 39, 813鈥?20] and a partial least-squares (PLS) model. With only one measurement of temperature and pressure at an equilibrium state, Korsten鈥檚 vapor pressure equation is able to produce a rough estimate of the vapor pressure curve, and PLS is utilized to model the residuals resulting from this simple Korsten model. An industrial case study is reported to show that the proposed method gives an improved model with the average error of less than 4%.