Large-Scale Experimental Study of Proppant Transport and Placement in Rough Fractures of Conglomerate Reservoirs

Abstract

Proppant placement in rough fractures is controlled by coupled effects of operating parameters, particle settling, and fracture-surface heterogeneity, yet quantitative criteria for evaluating final placement remain limited. This study reanalyzes previously obtained large-scale visualization data from a 20.3 m × 4.5 m × 10 mm rough-fracture system and develops an engineering-oriented framework for quantitative placement evaluation. Five indicators were introduced: normalized effective supported area, normalized effective supported length, near-wellbore accumulation ratio, distal coverage ratio, and ridge-height coefficient of variation. Their responses to pumping rate, fracturing-fluid viscosity, proppant size, and sand concentration were systematically compared. Results show that pumping rate primarily controls distal transport. Increasing the model pumping rate from 0.12 to 0.24 m³/min increased the normalized supported length from 0.875 to 1.000 and the distal coverage ratio from 0.026 to 0.115. Fluid viscosity and sand concentration showed stronger local sensitivities with respect to overall supported area, whereas particle size mainly affected the balance between distal transport and local deposition. Based on the combined indicators, four representative placement patterns were identified: near-wellbore accumulation, uniform spreading, distal-deficient placement, and bridging-prone high-loading placement. An exploratory four-factor power-law correlation was further evaluated using leave-one-out cross-validation. Although the in-sample fit was moderate, the LOOCV results showed limited predictive robustness, indicating that the correlation should be treated as a geometry-specific summary rather than a validated predictive model. The proposed indicators and classification framework provide a quantitative basis for comparing proppant-placement quality in large-scale rough fractures.