An experimental investigation of some capillary pressure-relative permeability correlations for sandstone reservoir rocks.

Bashir, Abdulaadem A.

Abstract

The water-oil relative permeability of the reservoir rock is one of the essential properties in petroleum engineering studies, and the use of this data is critical in the evaluation and management of oil fields, particularly in water flooding predictions and oil recovery calculations. The experimental measurement of these quantities can not be carried out without resorting to complicated and expensive coring. Furthermore, it is very time consuming, laborious and has great difficulties associated with it. Some mathematical models have been developed to predict these important parameters from capillary pressure measurements. This study has evaluated these models, and suggests now selecting and/or modifying the most appropriate one. A new model is verified including an error analysis against representative samples. A group of homogeneous sandstone rock samples were selected based on x-ray computed tomography (CT) imaging. Adopting a medical scanner for core analysis led to poor quantitative results. It has been found that to obtain reliable images, the scanner must be recalibrated using a reference point which falls within the geological region of interest. The technique has been extended to measure non-destructively the porosity (using dual-scan method) for small sub-volumes within the tested sample. Measuring porosity variations within the rock samples with CT scanning is a very powerful tool which allows the assessment of the degree of heterogeneity quantitatively and assists in sample selection. Porosity measured using the CT scanning technique showed excellent agreement with the conventional helium and liquid porosities. A new CT single-scan method was developed for porosity measurement based on using average x-ray attenuation coefficient measurements. The newly developed CT single-scan porosity measurement technique has formed the basis of the use of CT as a means for logging porosity of the whole core. CT porosity logs were performed for 24 and 27 foot sections on sandstone whole cores in fiber glass sleeving. Subsequently, 26 and 30 plugs were removed from the scanned positions. CT porosity showed good agreement with helium porosity for core plugs which were removed from the scanned zones. Since most of the existing models rely on the wetting and the non-wetting phase determination, the wettability of the selected samples has been restored, then measured using the Amott/USBM method. As a result, water and oil have been defined as the wetting and non wetting phase respectively. Using random samples with no wettability assessment will lead to misuse of the existing models especially in intermediate and oil wet rock samples. To evaluate the existing mathematical models, capillary pressure curves have been generated using the porous plate diaphragm, centrifuge and mercury injection for all selected samples. Relative permeability curves have been generated using the proposed mathematical models respectively. It was intended to use either CT or the 1-D x-ray technique in the experimental relative permeability measurements, however, it was discovered early on that using these techniques in unsteady state experiments definitely leads to significant errors. The limitations of the existing facilities and the required technical modifications to eliminate errors are presented as part of the thesis. Therefore, an unsteady state relative permeability rig was constructed. Results obtained showed that, apart from the difficulty in interpreting the unsteady state floods, it is impossible to produce a complete relative permeability curve. In order to obtain full and reliable experimental relative permeability curves, essential for comparison with the mathematically generated relative permeability data, a steady state relative permeability rig was constructed. Steady state water and oil relative permeabilities were measured for all selected samples. Although the technique is laborious and very time consuming, it successfully generated the most reliable and complete data set. Accordingly, the experimentally measured steady state relative permeability data was selected for comparison with data derived by the proposed mathematical models. Comparison results show that the use of existing mathematical models for deriving relative permeability data from capillary pressure curves for water wet rock samples leads to significant errors. However, Burdine’s model exhibited a “numerically” reasonable agreement with the laboratory experimental data, yet, using this model in its existing form certainly produced serious errors. The model incorrectly continues generating oil relative permeability values below the residual oil saturation, (water saturation greater than or equal to 1.0 -Sor). Using these false relative permeabilities in classical reservoir engineering calculations will incorrectly predict that the reservoir will produce far more oil than is physically possible. Burdine’s model (like others) also ignores the effect of rock wettability type and is assumed to be valid for any wettability including intermediate wettability. This is considered to be a major shortcoming in the mathematical model. Burdine’s model has been newly modified to suit water wet samples. The new model showed outstanding results; it is representative, efficient, flexible and suits the practical applications for oil reservoirs. Considering the reservoir rock properties such as porosity, wettability and allowing for the residual oil saturation, minimized errors significantly. This gives the new model superiority in predicting the relative permeability data from the capillary pressure curves. Although the new model developed used air-brine capillary pressure data, results obtained show that relative permeability curves can be reliably generated from capillary pressure data obtained by the centrifuge or mercury injection methods. This makes the technique more versatile and widely applicable in practice. In addition to the work detailed above, the following new approaches are presented: Since the air-brine capillary pressure data are the most representative measurements, but very time consuming, a new experimental approach has been developed which employs thin, large diameter samples, for measuring the capillary pressure. The time required to obtain a complete capillary pressure curve is remarkably reduced. A new method for correcting permeability measurement using Darcy’s law was presented for discussion. This method allows for the inflow end effect and the depth of mud invasion in formation damage assessment.

Citation

BASHIR, Abdulaadem A. 2000. An experimental investigation of some capillary pressure-relative permeability correlations for sandstone reservoir rocks. Robert Gordon University, PhD thesis. Hosted on OpenAIR [online]. Available from: https://doi.org/10.48526/rgu-wt-2807295

• Thesis Type: Thesis
• Deposit Date: May 22, 2025
• Publicly Available Date: May 22, 2025
• DOI: https://doi.org/10.48526/rgu-wt-2807295
• Public URL: https://rgu-repository.worktribe.com/output/2807295
• Award Date: Feb 29, 2000