Abstract
The study of etched fracture degradation has been a subject of interest within reservoir production engineers and it is generally expected that the etched fractures created after pumping acid in carbonates at pressures exceeding fracturing pressure are subject to degradation in time. Based on past observations, the fracture production performance is expected to decrease in time, as the etched fracture is exposed to in-situ stress that tends to crush the pinch points supporting the fracture, resulting in fracture closure or partial closure therefore having direct impact on well production.
The phenomenon of fracture degradation was previously studied in the context of acid fracture and proppant fracture effectiveness comparison. This paper attempts to further develop the subject by describing a workflow to analyze the reservoir lifetime production data of a North Sea chalk reservoir to identify production decline and quantify possible fracture half-length decrease. The ongoing study, comprising several phases, also aims to define an acid frac degradation model and then analyze possible ways to address the phenomenon through fracture design optimization.
After the collection and segregation of field production data and wells intervention reports, the data was evaluated. Wells PIs were calculated and modelled using machine learning algorithms. PI decrease was further analyzed from perspective of potential fracture parameters change and the impact of the fracture parameters change was simulated in the production simulator. Numerical simulations results were compared to actual production data.
Application of machine learning algorithms resulted in an analytical model for PI decrease that show a reasonable match with actual observed filed data. Model can be used as a reference to refine post stimulation production forecast of the wells in the field which helps with field development planning. Resulted PI degradation model is subject of further improvement through refinement of the input data accuracy and addition of new data.
