Estimation Methods for Solution Gas Venting/Flaring Volume
Introduction
The accurate calculation of greenhouse gas emissions is increasingly important in Alberta and throughout the world as government regulations require companies to reduce emissions or pay a penalty for every tonne of GHGs emitted over a specified target (in Alberta, facility emissions over 100,000 tonnes/year of CO2Eq). Process Ecology is in the process of developing an easy-to-use but rigorous tool, “FlareAdvisor”, to help operators accurately estimate the volumes of gas flared or vented from oil and gas facilities. The focus of this article is on the different approaches used for volume estimation of solution gas as a source of GHG emissions.
At oil production facilities, the bulk of the gas produced along with the hydrocarbon liquids is separated from the liquid phase at the inlet separator. A certain amount of gas remains trapped in solution at high pressures and it is referred to as solution gas. The amount of this gas depends on the temperature and pressure of the separation vessel, as well as the oil composition. The solution gas is subsequently released through a flashing process when the hydrocarbon liquids are further processed or stored at lower pressures and thermodynamic equilibrium between the vapour and liquid phases is established. This gas is also a source of VOC emissions and it may be vented, flared or conserved depending on the quantity and quality of gas, regulatory requirements and economics [1].
There are several approaches for estimating flashing/solution gas losses. Some of the more common methods recommended in the Canadian Association of Petroleum Producers (CAPP) [1] and American Petroleum Institute (API) [2] documents are:- Empirical correlations (EUB rule-of-thumb, Standing correlation, and Vasquez-Beggs Correlation)- Rigorous thermodynamic (flash) calculations using a process simulator
The main objective of this article is to compare solution gas losses from storage tanks calculated using empirical correlations and flash calculations, which is the basis in FlareAdvisor for solution gas volume estimation. The intent is to show that flash calculations can help companies more accurately report GHG emissions, thus reducing the risk of over-estimating volumes and associated financial penalties. It is noted that working and breathing losses can be an important component of tank emissions but are not addressed in this paper.
Case Study
In order to compare the various estimation methods and to perform solution gas loss calculations, an oil battery was selected as a case study. It was assumed that the facility has an inlet separator and oil storage tank and that crude oil with an API gravity of 45API is processed; oil deliveries are assumed to be at 500 Sm3/month. The temperature and pressure of the separator are 30C (86 F) and 450 kPag (65 psig), while those in the storage tank are 15.56C (60 F) and 101.325 kPaa (14.7 psia). The molecular weight of the solution gas is 44 kg/kmole.
Empirical Correlations and Flash Calculations
Both the CAPP and API Compendium documents include the EUB rule of thumb, Standing, and Vasquez-Beggs empirical correlations to estimate flashing losses. If sufficient input data are available (composition of the oil entering the storage tank, or compositions of sales oil and solution gas) flash calculations can be used for volume estimation of solution gas. In this review, the flash calculation was performed using Aspentech’s HYSYS v7.3 process simulator and equations shown in Table 1 were used in calculations. Note that the Standing and Vasquez-Beggs presented in the CAPP document are different from original ones. Please download the complete report to see the effect of those differences on the flashing losses estimation.
Figure 1. Shows the process flow diagram for the particular case study and the amount of solution gas volume calculated using the Peng-Robinson equation of state.
The solution gas volume flowrate from the storage tank estimated by different approaches against the separator temperature and pressure are shown in Figure 2. As indicated in Figure 2(a), the EUB rule predicts flashing losses 5.4 to 8 times of the volume estimated by HYSYS, while standing and Vasquez-Beggs predictions are similar and around two times the HYSYS results. CAPP reports that the EUB approach tends to estimate high flashing losses and is recommended for facilities with low oil production rate and some heavy oil production facilities. Figure 2(c) shows the amount of solution gas vented/flared from the tank for the case studied in this work. Although the volume estimated by Standing and Vasquez-Beggs correlations is closer to HYSYS result, they over-predict the volume of solution gas by over 100%.
Figure 2. Comparison of empirical correlation approaches to HYSYS.
Conclusion
The usual approach to evaluate solution gas volumes is using the empirical correlations which are typically generated for a specific region and hence based on a specific oil composition. Using an inappropriate correlation for the region of interest may cause unacceptably high errors in the estimation of the solution gas volume such as when using the EUB approach for light oils. If the thermodynamic equilibrium between two liquid and vapour phases is rigorously modeled through a flash calculation method and the compositions of the oil and gas are taken into account, more accurate estimates of the solution gas volume can be provided. This is the approach Process Ecology has implemented in the FlareAdvisor tool for the calculation of the solution gas vented or flared from emulsion treaters and storage tanks.
References
[1] Canadian Association of Petroleum Producers (CAPP), “Guide for Estimation of Flaring and Venting Volumes from Upstream Oil and Gas Facilities”, May 2002.
[2] American Petroleum Institute (API), “Compendium of Greenhouse Gas Emissions Methodologies for the Oil and Natural Gas Industry”, August 2009.
[3] Ahmed T., “Reservoir Engineering Handbook”, Gulf Professional Publishing, 2006.
[4] Danesh, Ali, “PVT and Phase Behaviour of Petroleum Reservoir Fluids”, Elsevier Science, 1998.
[5] Dandekar, A.Y., “Petroleum Reservoir Rock and Fluid Properties”, CRC Press, 2nd ed, 2013
.[6] Lyons, W.C., “Standard Handbook of Petroleum and Natural Gas Engineering”, Elsevier, 2005.
[7] Vasquez, M., Beggs, H.D., “Correlations for Fluid Physical Property Prediction”, Journal of Petroleum Technology, June 1980, 968-970.
Do you have questions or comments regarding this article? Click here to contact us.
