| Resumen |
The paper describes an efficient multi-stage data-driven method for aquifer and water influx pattern classification. It is based on statistical multivariate analysis techniques, Principal Components Analysis and Cluster Analysis. The main advantage of the proposed method is the integration of the analysis of static (water and oil properties) and dynamic (production series) data in order to understand the water-related behavior of the oilfield. For oilfield aquifers classification, oil and water physicochemical properties are used. Water influx pattern classification consists of three steps. Firstly, single-well production data analysis techniques are applied: (1) conventional decline curve analysis; and (2) a novel water-cut curve analysis with a logistic growth equation. Secondly, the obtained parameters are employed for the statistical multivariate analysis of the field. Finally, logistic-type equation parameters are used for qualitative identification of velocity of water entrance, water amount and first date of water entrance. The proposed method is discussed and illustrated using Puerto Ceiba case study. The work is a step forward towards the introduction of data-driven approaches in the engineering practices of water control. |
