Uso de Técnicas de Análisis Multivariante para la Caracterización de Fluidos de los Yacimientos de la Faja Petrolífera del Orinoco
K. Salazar and R. Meza
Journal name: Geociencias Aplicadas Latinoamericanas
Issue: Vol 2, No 1, March 2015 pp. 1 - 10
Info: Article, PDF ( 1.75Mb )
Price: € 30
The Orinoco Oil Belt (OOB) is one of the largest heavy and extra-heavy crude oil reserves in the world, with its weakly consolidated formations, low temperatures and reservoirs close to saturation conditions all representing a challenge for the characterisation of fluid properties (PVT analysis). Obtaining PVT information of an experimental nature requires high investment due to the complexity of operational procedures in the field, sampling tools and the technology needed for reliable laboratory analyses. As a consequence of the above, common practice in the Venezuelan oil industry has been based on the use of correlations or on establishing analogies with neighbouring fields to estimate PVT parameters, although these methods do have some limitations. In this regard, the purpose of this research lies in evaluating the reliability of the use of information from contiguous areas, assuming that they reproduce behaviour similar to the properties of the fluids. Through analysis of the main components (AMC), inter-associations between six of the most important variables were analysed for the characterisation of fluids, allowing the construction of a factorial design with axes describing saturation indices and the heaviness of the crude. The directions of the axes are defined by the eigenvectors associated with the eigenvalues which capture the largest portion of total variability, thereby allowing a reduction in the scalability of the study. By including in the AMC the geographic area from which the sample was taken as an illustrative categorical variable, it emerged that the nearby PVT data may behave completely differently, which means that proximity does not determine the characteristics of the fluids, and they therefore must be differentiated through proper characterisation of the oil reservoirs, allowing correlations to be established or reliable analogies to be carried out in addition to the optimisation of acquisition plans and analysis of PVT information. Another important finding refers to the relationship between the variables involved in the PVT characteristation; if easy-to-measure variables are known, such as pressure, temperature or API gravity, then any well may be located wthin the factorial designs generated, which can then be analysed according to saturation and heaviness profiles, and then compared with samples showing similar behavour.