The ultimate goal of sand control design in producing hydrocarbons is to obtain solids free at lowest cost per barrel. Therefore, predictable production and its associated cost is vital to achieve the best business value. Previously, the prediction of production from sand control well is cumbersome. This paper presents the novel method which using the field data to provide the insight of sand control and deliver an optimal design for the sanding propensity well.
Gravel pack sand control designs involves many parameters and is a tedious process for an engineer to interact with several dynamics parameters. This novel method started with field data collection from previous sand control operations. The datasets are prepared into the structured form, then reservoirs are sorted based on their similarities and finally the parameters are selected based on their significance. These parameters are mapped onto the productivity index using a variety of modeling types. The prediction result show that the productivity index can be modeled with high statistic measure (i.e. R-squared). Ultimately, Net Present Value (NPV) derived from anticipated reserves are known before pumping the gravel pack job. The continuous improvement of datasets can significantly help improve the sand control design.
In this study, the novel method is presented using field dataset to optimize the sand control design. The design process can be driven with the use of data and machine-learning algorithms. This emerging technology allows greater insight into the day-to-day operations. The continuous adoption of this technology is a key enabler of futuristic industrial 4.0. The data-driven process empowers the oil & gas industry to become more durable, robust and competitive.