Investigation of Proppant Erosion in Fracturing Tree under Ultralarge Flow Rate Involving Artificial Intelligent Prediction

Summary With the rising demand for shale gas, fracturing operations are evolving toward ultrahigh flow rates (>22 m³/min), high proppant concentrations (>15%), and large fluid volumes (>2500 m³). Intense erosion can increase treatment risks. However, there is currently a lack of quantitative studies on the erosion of fracturing tree, especially under large-scale fracturing operations. To address this problem, we conducted numerical research on the fracturing tree erosion characteristics under field conditions using the Ansys® Fluent simulation software. The variations in erosion concerning different flow rates, particle sizes and densities, and particle volume fractions were analyzed thoroughly. The numerical model iwa calibrated using the erosion results from the four-way component. An artificial neural network (ANN) model was developed to predict the erosion rate of the fracturing tree. The results show that the four-way component experiences the most severe erosion within the fracturing tree. Strong turbulence is generated around the four-way, causing erosion on the side walls and forming two distinct pits near the junction. Denser large proppants, dominated by inertia, erode the pit near the junction. Lighter small particles, influenced by vortices, form pits in the lower region. Erosion can propagate downstream to the gate valve, forming many small wave-like textured pits. These pits significantly increase local stress concentration, potentially exceeding the material’s yield strength. When the flow rate exceeds 24 m³/min, the erosion rate increases significantly, especially at large particle size (≥30/50 mesh) and high sand volume fraction (≥20%). The 70/140 mesh sand increases the erosion area on the inner walls. The ANN model trained by extensive numerical samples can accurately predict the erosion rate. Engineers can use this model to optimize the pump parameters, make rational decisions, and avoid fracturing treatment risks. The innovative method can be extended to study the fracturing tree erosion under other injection methods.