A NEW FLUID PROPERTY - INSITU FORMATION VOLUME FACTORS FROM FORMATION TESTING

摘要

Downhole fluid analysis (DFA) has been successfully employed to estimate compositions, gas/oil ratios (GOR), coloration (asphaltene content), density, viscosity and oil-based mud (OBM) filtrate contamination in real time. However, an important element – formation volume factor (FVF), defined as a ratio of the volume of a reservoir fluid at downhole conditions to the volume of stock tank oil at standard conditions – has been missing. This paper presents several real-time methods to obtain FVF from formation tester data and discusses applications of the newly derived fluid property.A novel method is developed to obtain FVF based on mass balance during a single-stage flash process performed at standard conditions. The required input parameters, including density, GOR, and composition, are all available from existing DFA measurements in real time, ensuring a robust and simple FVF calculation. The new method is compared to a pressure-volume-temperature (PVT) laboratory-derived correlation based on machine learning methods and a method based on an equation of state (EOS) that uses more than 1350 fluid samples covering all hydrocarbon reservoir fluid types. The results show that the prediction error of the new FVF algorithm has an average absolute deviation of 3.6%. A second method, which begins with the definition of FVF and utilizes optical absorbance measurements and estimated fluid composition -- CO2, C1, C2, C3-5, and C6+ -- to derive an equation for the FVF is also demonstrated. The coefficients in this equation are calibrated against an optical spectral library derived from 160 hydrocarbon samples measured at different temperatures and pressures with values up to 175 oC and 20,000 psi. The FVF prediction standard error of approximately 2.4 % for this method is estimated by comparing FVF predictions with laboratory measured FVFs on a validation dataset.The use of the new FVF algorithms in deriving several real time fluid properties is illustrated. For example, FVF is used in OBM filtrate contamination calculations during sample cleanup for focused and non-focused sampling tools and is used to convert live fluid-based OBM filtrate contamination to a stock tank liquid-based value.Best practices for selecting which FVF algorithm to use are discussed and recommendations for algorithm selection are made. Several case studies detail FVF calculations based on wireline or while-drilling formation tester data and the examples show how FVF is used in other real time DFA workflows. The results obtained from the new methods are in good agreement with results of PVT laboratory sample analysis.