MODELING AND OPTIMIZATION OF PROPPANT DISTRIBUTIONS IN MULTICLUSTER HYDRAULIC FRACTURE-NATURAL FRACTURE (HF–NF) NETWORKS

One of the foremost challenges with proppant transport in hydraulic and/or natural fractures within unconventional reservoirs is that the proppant particles do not remain in the suspension transport mode, deviating significantly from the suspension assumption shared by most existing models that allow particles to be represented by a continuous pseudo-fluid phase. The advantage of the CFD-DEM simulator used in this study is its proppant particle migration within a Lagrangian frame of reference, which enables the description of proppant transport in any transport mode. This coupled simulator can capture the complex interactions between the proppant particles, the carrier fluid, and the walls of the hydraulic and/or natural fractures. Such a knowledge-based tool can provide new insights that can be used to optimize the stimulation design, leading to better proppant placement and higher fracturing effectiveness — which ultimately leads to higher flowrates and improved hydrocarbon recovery. The first part of the study involves the development of a robust proppant-bridging criterion at the Hydraulic Fracture-Natural Fracture (HF-NF) interface. The simulation results are used with an image classification algorithm to create a reliable proppant bridging criterion obtained from a well-trained Artificial Neural Network (ANN). The ANN is subjected to a standard K-fold cross-validation and is tested against a large suite of "control" simulation results. The second part of the study focuses on optimizing the proppant distribution in a multi-cluster horizontal well system. The cumulative proppant distributions at each cluster are computed based on the number of representative particles entering each cluster at each timestep. A sensitivity analysis is performed using the key system parameters (cluster spacing, number of clusters, proppant size, and perforation width) and the results of these analyses are used to propose optimization strategies