Projects per year
Abstract
With dwindling easily accessible oil and gas resources, more and more exploration and production activities in the oil industry are driven to technically challenging environments such as unconventional resources and deeper formations. The temperature and pressure can become extremely high, e.g., up to 250 °C and 2400 bar, in the deep petroleum reservoirs. Furthermore, many of these deep reservoirs are found offshore, including the North Sea and the Gulf of Mexico, making the development even more risky. On the other hand, development of these high pressure high temperature (HPHT) fields can be highly rewarding if successfully produced. This PhD project is part of the NextOil (New Extreme Oil and Gas in the Danish North Sea) project which is intended to reduce the uncertainties in HPHT field development. The main focus of this PhD is on accurate description of the reservoir fluid behavior under HPHT conditions to minimize the production risks from these types of reservoirs. In particular, the study has thoroughly evaluated several noncubic Equations of State (EoSs) which are considered promising for HPHT fluid modeling, showing their advantages and short comings based on an extensive comparison with experimental data. In the course of the evaluation, we have developed new petroleum fluid characterization procedures, built large databases for welldefined mixtures and reservoir fluids, and improved the evaluation software and made it more suitable for efficient and large scale comparison. We have made a comprehensive comparison between cubic and noncubic EoSs to evaluate whether advanced EoS in noncubic forms, including both the SAFTtype EoS with strong theoretical basis (e.g. the PCSAFT EoS) and the empirical BWRtype EoS (e.g. the SoaveBWR EoS), can be advantageous for describing the physical properties and phase equilibrium of reservoir fluids over a wide temperature and pressure range. In addition, we have also compared these models in calculation of heat capacities and JouleThomson coefficients for pure components and multicomponent mixtures. JouleThomson coefficients are of special interest to the oil industry because of the so called reverse JouleThomson effect commonly observed in HPHT fields, where a decrease in pressure results in an increase in temperature, which is just the opposite to the effect at low pressure. In the comparative studies between cubic and noncubic vmodels, we also included GERG2008 which is a widerange EoS developed for 21 components of natural gases and their binary mixtures and is regarded as the mostaccurate EoS model for natural gas mixtures. It was found that the noncubic models are much better than the cubics in density, compressibility, heat capacity and JouleThomson coefficient calculation of the well defined light and heavy components in reservoir fluids over a wide temperature and pressure range, GERG2008 being the best with the lowest deviation among all EoS models. GERG2008 however gives very large deviations for bubble point pressure calculation of some heavy and asymmetric binary systems such as nbutane + nnonane system. This suggests that this EoS and its binary interaction parameters could still be improved for some of the binary pairs. SoaveBWR gives the closest prediction of the thermal properties to that of GERG2008 among other EoSs tested in this study. The binary VLE calculation showed that PCSAFT and SoaveBWR are similar to SRK and PR in correlating the important binary pairs in reservoir fluids. Although SoaveBWR and PCSAFT give smaller average kij values than SRK and PR, they are more sensitive to the change in kij . Phase envelope prediction of synthetic gases showed that all the EoS models were similar for not too asymmetric synthetic gases,with or without the optimal kij values for SRK, PR, PCSAFT and SoaveBWR. For highly asymmetric synthetic mixtures, SoaveBWR and GERG2008 tend to predict phase envelopes different from other models where as none of the tested models give satisfactory predictions. For heat capacity and JouleThomson coefficients, GERG2008 and SoaveBWR give the closest predictions. All the evaluated EoS models tend to predict a nearly constant JouleThomson coefficient at high pressures. For typical reservoir temperatures, the constant is around 0.5 K/MPa. For noncubic models like PCSAFT the characterization method is less mature than the cubic models. A reservoir fluid characterization method for PCSAFT has been proposed by combining Pedersen’s method with a newly developed set of correlations for the PCSAFT model parameters m, mε/k and mσ3. In addition, we further improved the characterization method for PCSAFT by adjusting the correlations with a large PVT database. We have further improved the correlations and more importantly, we have established a general approach to characterizing reservoir fluids for any EoS. The approach consists in developing correlations of model parameters first with a database for welldefined components and then adjusting the correlations with a large PVT database. The adjustment is made to minimize the deviation in key PVT properties like saturation pressures, densities at reservoir temperature and Stock TankviOil (STO) densities, while keeping the nalkane limit of the correlations unchanged. Apart from applying this general approach to PCSAFT, we have also shown that the approach can be applied to classical cubic models like SRK and PR. In addition, we discussed how to develop a PNA based characterization for PCSAFT and also utilize a large PVT database to further improve the characterization. With the developed characterization methods, we have made a comparison in PVT calculation involving 17EoScharacterization combinations and 260 reservoir fluids. PCSAFT with the new general characterization method is shown to give the lowest AAD% and maximum deviation in calculation of saturation pressure, density and STO density, among all the tested characterization methods for PCSAFT. Application of the new characterization method to SRK and PR improved the saturation pressure calculation in comparisonto the original characterization method for SRK and PR. Using volume translationtogether with the new characterization approach for SRK and PR gives comparable results for density and STO density to that of original characterization for SRK and PR with volume translation. For the PVT database used in this study, cubic EoSs seem to have better performance than PCSAFT in calculation of saturation pressure; PCSAFT and cubics with volume translation show comparable results in calculation of density and STO density. As a preliminary attempt to integrate more analytical information in characterization, we discussed how to modify the existing algorithms to utilize data from both simulated distillation and true boiling point distillation, and in particular, the component distribution information from the simulated distillation. Some analyses have been made on the impact of including more detailed analytical information. Finally, to improve SoaveBWR for mixture calculation, we have tried to develop several new sets of mixing rules for this EoS. The new mixing rules were developed based on some theoretical considerations as well as the previous mixing rules for noncubic EoS models. In addition, it was tried to create some hybrid mixing rules by combining a new set of mixing rules and the original mixing rules for SoaveBWR. It was shown that some problems with the original SoaveBWR mixing rules can be fixed by the new mixing rules although the overall performance is not significantly improved. Development of mixing rules for noncubic EoS models is still a semiempirical process, requiring extensive testing to evaluating their performance. We have developed the code in a structured manner so that the new mixing rules can be quickly tested. It can facilitate further extensive screening of new mixing rules for SoaveBWR or even other noncubic EoS models.
Original language  English 

Place of Publication  Kgl. Lyngby 

Publisher  DTU Chemistry 
Number of pages  305 
Publication status  Published  2017 
Fingerprint Dive into the research topics of 'Modeling Study of High Pressure and High Temperature Reservoir Fluids'. Together they form a unique fingerprint.
Projects
 1 Finished

Study of High Pressure and High Temperature Reservoir Fluids
Varzandeh, F., Yan, W., Stenby, E. H., Møller, K. B., Lindeloff, N. & Montel, F.
Eksternt finansieret virksomhed
15/12/2013 → 25/08/2017
Project: PhD