Roughness of Hydraulic Fractures: Importance of In-Situ Stress and Tip Processes
- D.B. van Dam (Delft U. of Technology) | C.J. de Pater (Delft U. of Technology)
- Document ID
- Society of Petroleum Engineers
- SPE Journal
- Publication Date
- March 2001
- Document Type
- Journal Paper
- 4 - 13
- 2001. Society of Petroleum Engineers
- 1.2.3 Rock properties, 3.2.3 Hydraulic Fracturing Design, Implementation and Optimisation, 2.5.2 Fracturing Materials (Fluids, Proppant), 4.6 Natural Gas, 3 Production and Well Operations, 7.2.2 Risk Management Systems, 1.14 Casing and Cementing, 4.1.2 Separation and Treating, 4.3.4 Scale, 4.1.5 Processing Equipment
- 3 in the last 30 days
- 623 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 12.00|
|SPE Non-Member Price:||USD 35.00|
The surface roughness of hydraulic fractures reflects the fracturing process at the tip. We conducted an experimental study of hydraulic fracture propagation and observed a characteristic roughness pattern. The roughness developed without shear or torsional loading in the plane of the penny-shaped fracture and is determined by material properties and the externally applied stresses on the sample, which represent the effective in-situ stresses. We quantified the roughness and showed that it correlated well with a measure of the plastic zone size around the fracture tip. New explanations for fracture surface roughness that develops under Mode I loading conditions are presented.
Hydraulic fracture surfaces in general are not smooth but show roughness. Similar roughness is visible on tensile fracture surfaces in rocks as well as many other materials. Fracture roughness is neglected in many hydraulic fracturing models. While this may be correct in some situations, there are cases in which the fracture surface roughness is important.
Rough fracture surfaces influence the mechanical behavior and conductivity of fractures. Fracture surface roughness, in combination with a small displacement of opposing fracture surfaces relative to each other, causes imperfect closing owing to the nonmatching surfaces. This leads to a residual width after closure, which is observed in laboratory1 and field tests.2 The residual width leads to a fracture conductivity after closure that can be great enough so that proppant volumes can be reduced significantly or even omitted.3 Another effect of roughness is that it can increase the fluid pressure gradient in the fracture when the amplitude of the roughness is of the same order of magnitude as the average fracture width.4 Fracture roughness can also cause problems with proppant transport. Furthermore, the presence of a significant residual width influences the wellbore pressure during reopening of a hydraulic fracture.5
These are some examples of the practical importance of fracture surface roughness that indicate that prediction of this roughness is valuable. The aim of this paper is find explanations for the fracture surface roughness that occurs under Mode I loading, which is the situation of interest for most hydraulic fracturing situations. The fracture roughness reflects the fracturing process, so understanding the roughness of hydraulic fractures is related to understanding the fracturing process at the tip of the hydraulic fracture. We present results of scaled hydraulic fracturing experiments where we vary the parameters that could influence fracture roughness. Furthermore, we present new explanations for the fracture surface roughness that are supported by the experimental results.
Experimental Setup and Method
We performed scaled model experiments of hydraulic fracture propagation. The scaling of these experiments is based on the equations that describe linear elastic fracture propagation6 and starts from the viscosity-dominated regime. Leakoff is scaled by the fracture efficiency. To have sufficient time to do measurements, the time scale for the propagation phase is between 102 and 104 seconds. The fracture radius is approximately 0.1 m. The combination of these practical conditions with the scaling laws requires a relatively high fluid viscosity and low fracture toughness in the experiments.
Various length scales are involved in hydraulic fracturing that may be used to scale the experimental fracture size up to field scale. The length scale of the plastic zone around the fracture tip is determined by the fracture toughness and shear strength of the rock and the externally applied stresses. Scaling of the plastic zone size, together with the tensile zone size, is expected to be important for studying fracture mechanisms and surface roughness. In our experiments, these length scales are scaled with respect to the other length scales of interest. This was achieved by scaling down the fracture toughness in accordance with the linear elastic scaling and by taking a similar shear strength in the laboratory and field.7
We loaded the rock samples in a true-triaxial-compression machine with stiff loading platens to simulate effective in-situ-stress states. Because this is an open system, pore pressure cannot be applied. We used 0.1-mm-thick Teflon sheets greased with Vaseline to reduce friction between the sample and the loading platens. We measured the friction coefficient between the sample and loading platens for this configuration and found it to be less than 1% (see Ref. 7). This low value is probably caused by the Vaseline forming a lubricating layer. The loading platens are mounted on spherical seats greased with thick grease. The compression machine consists of three uniaxial frames that move independently of each other. The cubic samples are 0.30 m in size. The created hydraulic fracture is penny-shaped and has a maximum radius of approximately 0.10 m. Fig. 1 shows a schematic view of the setup.
The hydraulic fracture is oriented transversely to the wellbore wall, which was sealed with a 0.5-mm-thick glue layer. A 3-mm-deep notch was sawed in the wellbore wall to control the fracture location. The wellbore radius is 1.15 cm. The confining stress, sc, perpendicular to the plane of the fracture is always smaller than both horizontal stresses, sh, directed parallel with the fracture plane, which have equal values. The deviation of the stress inside the block from the expected value is less than 10%.
During fracture propagation, a high-pressure pump injects fluid into the wellbore, which has a dead string mounted on its end where wellbore pressure is measured. A linear variable differential transformer (LVDT), mounted with clamps in the wellbore, measures the fracture width within a measuring error of approximately 10%. The fracturing fluid we used is silicon oil, which behaves approximately Newtonian at the shear rates of interest. We conducted a number of propagation tests, which showed a good reproducibility.7,8
We also performed ultrasonic measurements, which yielded information about the fracture radius and width.1,7,9 Fig. 1 shows the acoustic transducers. These measurements are done to obtain the geometry of the hydraulic fractures during propagation and closure but are not used in this study.
We used artificial rock samples made of cement and plaster and natural rock samples of diatomite. These rocks span a wide range of rock properties regarding permeability and modulus (Table 1). Preparation of the cement and plaster blocks is described in Ref. 8. The strength of the plaster blocks depends on the water content.10,11
|File Size||596 KB||Number of Pages||10|