Impact of PVT Correlations on Reserve Estimation: Reliability and Qualitative Analysis

Abstract

Abstract Many correlations are currently being used to predict PVT properties in the oil and gas industry. This work thoroughly reviewed the works done by Standing (1947), Vasquez and Beggs (1980), Glaso(1980), Petrosky & Farshad (1993), Al Mahoun (1988), Kartoatomodjo & Schmidt (1994), Obomanu & Okpobiri (1987) and Ikiensikimama & Ogboja (2009). There were obvious discrepancies between these models and the experimental results mainly due to the fact that the models were built using data from the US and Middle East. Sequel to the discrepancies observed in these correlations, this works proposes a new model based on excel regression analysis to depict a more accurate value of the Oil Formation Volume Factor (OFVF) based on experimental results. The basis for comparison of the various correlations was Average Percent Relative Error, Average Absolute Percent Relative Error, Minimum Absolute Percent Relative Error, Maximum Absolute Percent Relative Error and Standard Deviation. This work utilized data from the Niger Delta and other relevant regions to modify or propose a correlation which will have a minimum error or standard deviation from the experimentally determined results. A total of 119 data sets (comprising of 833 data points) were collected and checked for accuracy. Quantitative analysis (reliability analysis) and qualitative analysis using cross plots was conducted on the seven considered literature correlations. Preliminary comparison of the performance of various correlations revealed Al-Marhoun to be the most accurate with an AAPRE of 2.669957%. However, using non-linear multiple regressions a new correlation for OFVF was which obtained significantly outperformed the other correlations. It provided a far much better Average Absolute Percent Relative Error of 2.4037%. From the volumetric approach of determining the reserve estimate, and the Stock Tank Oil Originally in Place (STOOIP) analysis, thus in conclusion, the developed correlation with minimal error in Bo had less impact on the oil originally in place and gave the most accurate STOOIP. A sensitivity analysis was carried out using @Risk™ Software, in order to quantify the impact of these variables on the overall estimate of Oil in Place.