Loading...
Slickwater proppant transport in complex hydraulic fracture networks: experimental study and scalable correlations development
Alotaibi, Msalli A.
Alotaibi, Msalli A.
Citations
Altmetric:
Advisor
Editor
Date
Date Issued
2015
Date Submitted
Collections
Research Projects
Organizational Units
Journal Issue
Embargo Expires
2016-12-09
Abstract
Slickwater hydraulic fracturing is a game changing technology that has enabled the oil and gas industry to economically develop enormous unconventional resources over the past decade. Despite its great success, this technology carries with it challenges such as the understanding of its proppant transport mechanism in induced complex fractures and predicting the developed dune heights. Addressing these two challenges is the primary goal of this PhD research project. The project was divided into two phases to achieve this goal, an experimental study and correlations development. For the first phase, experiments were conducted using an in-house designed and developed apparatus capable of simulating proppant flow in complex fractures. The aim of these experiments was to study slickwater proppant transport and obtain lab data for the correlation development phase. Brown and white sand proppants of three sizes, 100, 30/70, and 20/40 mesh, were tested at a wide range of slurry velocities and concentrations. The experimental results show that sand dune formation undergoes four developmental stages before reaching its final shape and height. Every stage is unique in its proppant transport mechanism, dune build-up rate, dune shape, and particle size distribution. The settled sand grain size sorting is the highest for 30/70 mesh sand. A model was developed describing the anticipated 30/70 mesh sand dune size distribution. Also, the lab results show that developed proppant dunes by all sand sizes and types reached heights exceeding 90% of the equilibrium dune level (EDL) at a slurry velocity of 14.5 ft/min and concentration of 1.00 pound-mass/gallon (ppg) indicating very low slickwater proppant transportability. The settled proppant developed an increasing height curved dune shape in the fracture slot which is attributed to the fracture slot wall frictional effects on proppant settling. Increasing slurry flow rate showed an erosional effect on developed dunes following a power law trend with EDL for the 30/70 and 20/40 mesh sands and linear trend for the 100 mesh sand. The slurry velocity erosional effect is very low relative to the increase in slurry velocity. Proppant concentration showed a proportional effect on EDL following a power law trend at all tested sand sizes. Smaller size sands showed better proppant transport and lower dune heights. The brown sand displayed a better proppant transport performance than the white sand attributed to its higher surface drag coefficient resulting in higher exerted fluid drag force. More than 90% of the proppant area in the primary fracture formed in the first half length of settled proppant. Moreover, the study provides an experimentally based answer to the debatable question whether slickwater can transport proppant into tertiary fractures or not. In fact, it was found that proppant in slickwater is not only capable of “turning the corner” but also developing dunes exceeding 96% EDL of secondary and tertiary fracture slot heights. This high proppant transport indicates that fracture complexity is not the major limiting factor of slickwater proppant transport as long as enough proppant is injected. Developed dune areas in secondary fractures represent 40.8% of the total dune area in the fractures network measured at 11.02 minutes. This indicates the high proppant transport capability of slickwater to secondary fractures. However, tertiary fractures formed very low propped area of 4.5%. Five types of scalable correlations were developed for both white and brown sands that can predict dune heights at a wide range of flow rates, proppant concentrations, and median diameters. These correlations are categorized by their dependent variables. The first and second types are velocity (V) and concentration (C) based correlations, respectively, while the third and fourth types combine both V and C in their correlations, V/C based correlations. The fifth type provides a much more capable correlation than the other types as it can predict proppant dune height at any median diameter in addition to velocity and concentration within the tested values. The developed correlations were verified by comparing their predicted values to the lab data. All correlations showed very low average errors ranging from 0.17% to 0.76%. The fifth correlation was compared to the only published correlation that can predict slickwater proppant dune height at field applicable conditions by Wang et al. (2003). Both correlations showed close values at 30/70 and 20/40 mesh sands but showed large variations at the 100 mesh sand which is attributed to the lack of small size proppants in the lab data used by Wang et al. (2003).
Associated Publications
Rights
Copyright of the original work is retained by the author.