Full text loading...
Most of the wells drilled on the Norwegian Continental Shelf during the last two decades are equipped with permanent gauges installed both at the wellhead and downhole. The gauges enable both on-the-fly and long-term monitoring of well performance, crucial for both the short-term optimization on the well level and the long-term reservoir management. The industry of today needs automation of the gauge data analysis and interpretation. Scalable data-driven solutions are being explored, while inheriting physics of well and reservoir flows is still crucial. This paper presents a recently developed workflow combining both pure data-driven and physics-informed methods to automate analysis and interpretation of big datasets from the permanent gauges. The workflow focuses on combination of pressure and rate data, commonly used in monitoring well performance with time-lapse pressure transient analysis, although many methods may be applied to other datasets such as temperature measurements or even other industries where transient behavior of measurements is observed. We concentrate further on the value of integrating data-driven approaches with physics-informed proxy-metrics for monitoring the well and reservoir performances. The paper concludes with the current and potential application areas of the methods developed.