Abstract
Production allocation from petroleum geochemistry is defined here as the quantitative determination of the amount or portion of a commingled fluid to be assigned to two or more individual fluid sources (e.g., a pipeline, field, reservoir, well) at a particular moment in time, based on the fluid chemistry. It requires: i) knowledge of the original chemical compositions of each of the fluids prior to mixing (referred to here as the "end members"), and ii) that statistically valid differences in their chemistries can be identified. Petroleum geochemical-based methods for production monitoring and allocation are much lower cost than using production logging tools, as there is no additional rig time or extra personnel required at the well site. Additionally, no intervention to the production of hydrocarbons from a well is required and, hence, there is none of the risk entailed in additional operational activity. Geochemical methods are applicable to a wide range of fields, irrespective of pressure, temperature, reservoir quality and reservoir fluid type. The method has been in existence for over 30 years, during which time a number of different analytical methods, data pre-processing and treatment approaches have been applied. This paper summarises these approaches, and provides examples, but also describes a "best practice" which is not a "one size fits all" approach, as is sometimes seen in the literature. A successful production allocation study consists of the following steps: i) Selection of end member samples that contribute to the commingled production fluid; ii) Determination of the differences in chemical composition of the end members through laboratory analysis of the end members (e.g. by WO-GC), replicate analyses of samples and statistical treatment of the data (e.g. PCA); iii) If statistically significant differences exist, laboratory analysis of the end members and commingled fluids with appropriate replicate analyses of samples; iv) Data selection, pre-processing (e.g. selection of ratios or concentrations of components); v) Determination of end member contributions by solving equations (e.g. least squares best fit) and uncertainty estimation (e.g. Monte Carlo or Bootstrap methods). The differences in approach for conventional versus unconventional plays are also discussed.
MINERALOGICAL MAPPING IN THE PART OF A GOLD PROSPECT USING EO-1 HYPERION DATA
