Data & Analytics

Integration of Real-Time Monitoring and Data Analytics To Mitigate Sand Screenouts During Fracturing Operations

A self-updating and customizable data-driven strategy for real-time monitoring and management of screenout, integrated with proppant filling index and safest fracturing pump rate, is proposed to improve operational safety and efficiency at field scale.

Fracking Diagram
Source: Filo/Getty Images

Sand screenout, the most frequent incident during hydraulic fracturing, is one of the major threats to operational safety and efficiency. Screenout occurs when advancing hydraulic fractures are blocked by injected proppant-slurry, stall, and develop fluid overpressure.

Because massive wells are still being hydraulically fractured every year, operational safety has become a critical and urgent issue that has always been overshadowed by the whether-or-not controversy. However, the suddenness and unheralded surprise of screenout make it extremely difficult to predict and handle.

Previous efforts attempt to predict screenout as discrete events by interpreting injection pressure directly. We propose and then demonstrate a self-updating (via data and experience augmentation) and customizable (numerical models and algorithms) data-driven strategy of real-time monitoring and management for screenout based on records of shale gas fracturing.

Two new indicators—proppant filling index (PFI) and safest fracturing pump rate (SFPR)—are improved and then integrated into the strategy. The PFI reveals the mismatch between injected proppant and hydraulic fractures and provides a continuous time-historical risk assessment of screenout. A pretrained ensemble learning model is applied to process the geological and hydraulic measurements in real time for the PFI evolution curve during fracturing operations. Integrated with the SFPR, a stepwise pump rate regulation strategy is deployed successfully to mitigate sand screenout for field applications.

Four field trials are elaborated, which are representative cases exhibiting the data-driven approach to monitor and manage sand screenout during hydraulic fracturing.

Footnote: This abstract is taken from paper SPE 219747 by L. Hou, Shanghai Jiao Tong University; D. Elsworth, Pennsylvania State University; P. Gong, Drilling Technology Research Institute of SINOPEC, Shengli Oilfield Service Corporation; X. Bian, SINOPEC Research Institute of Petroleum Engineering; and L. Zhang, China University of Petroleum (East China). The paper has been peer reviewed and is available as Open Access in SPE Journal on OnePetro.