Reservoir Geochemistry for engineering applications

Offentlig prosjektsammendrag

The R&D project looks to combine advanced analytical methods (GPC-RI-UV) with machine learning based models calibrated to large legacy dataset to optimize oil sample characterization and to provide a method that can provide useful information on oil samples in the presence of oil-based drilling mud (OBM) contamination. The analytical method employed is called gel permeation chromatography couple with a refractive index and UV detector. The research project aims to utilize recent advances in analytical methods for petroleum sample characterization and in data analytics, particularly machine learning models, to create a powerful method to rapidly characterize an oils properties. It is current practice to collect oil samples and then send these samples to two or even three different laboratories to perform analyses to determine the oils physical and chemical characteristics. Our project aims to perform a single, rapid and cost-efficient test the rest from which will be plugged into the machine learning models to provide predictions of the properties of interest; these might include API which determines the oils value, or viscosity which determines how rapidly it may be produced from a well. Petroleum wells are commonly drilled with oil-based drilling fluids. These fluids contaminate the liquid samples collected adding complexity to the interpretation of the results obtained. The method being developed aims to provide a method to deconvolve the indigenous oil signal from the contamination, allowing us to see around the influence of the OBM which will be particularly useful in solvent extracts from rock samples taken while drilling a well. This is a much lower cost way to provide information about the characteristics of the oil in the reservoirs being produced. To achieve this OBM samples will need to be characterized to assess their characteristic signals in the GPC-RI-UV analyses sit. Once understood, methods to target the indigenous oil signature can be developed.