Fracture Geometry Models - PKN, KGD, and Radial
Overview
Classical 2D fracture models predict the geometry of a hydraulically induced fracture -- its width, length (or radius), net pressure, and volume -- as functions of injection rate, fluid properties, rock properties, and time. Three fundamental models serve as the foundation for fracture design:
| Model | Full Name | Year | Key Assumption |
|---|---|---|---|
| PKN | Perkins-Kern-Nordgren | 1961/1972 | Vertical plane strain, confined height |
| KGD | Khristianovic-Geertsma-de Klerk | 1955/1969 | Horizontal plane strain, confined height |
| Radial | Penny-shaped (Sneddon) | 1946 | No height confinement, radial symmetry |
All three models solve the coupled problem of:
- Elasticity -- rock deformation under internal fluid pressure
- Fluid mechanics -- viscous flow within the fracture
- Material balance -- injected volume = fracture volume + leakoff volume
- Fracture mechanics -- propagation at the fracture tip
Common Parameters
| Symbol | Description | Oilfield Units |
|---|---|---|
| Injection rate (one wing) | bbl/min | |
| Fracturing fluid viscosity | cP | |
| Young's modulus | psi | |
| Poisson's ratio | dimensionless | |
| Plane strain modulus, | psi | |
| Fracture height | ft | |
| Pumping time | min | |
| Carter leakoff coefficient | ft/min$^{1/2}$ | |
| Net pressure (fracture pressure minus closure stress) | psi | |
| Fracture width (average or maximum) | in | |
| Fracture half-length | ft | |
| Fracture radius (radial model) | ft |
Plane Strain Modulus
The plane strain modulus appears in all fracture width equations:
Typical values:
- Sandstone: psi,
- Shale: psi,
- Limestone: psi,
PKN Model (Perkins-Kern-Nordgren)
Physical Description
The PKN model assumes:
- Confined height: fracture height is fixed by stress barriers above and below
- Vertical plane strain: each vertical cross-section deforms independently
- Length much greater than height:
- Elliptical cross-section: width varies elliptically across the height at any position along the length
- Maximum width at wellbore: width decreases toward the fracture tip
Plan View (top down) Cross-Section at any x
┌─────────────────────┐ ┌────────────────┐
│ │ │ ◜ ◝ │ ▲
│ ○ │ │ ◜ ◝ │ │
│ well ──────────► tip │◜ w(x) ◝│ hf
│ ○ │ │ ◟ ◞ │ │
│ │ │ ◟ ◞ │ ▼
└─────────────────────┘ └────────────────┘
◄────── xf ──────────► ◄── w ──►
Width decreases toward tip Elliptical shape
Equations (No Leakoff)
Maximum width (at wellbore, ):
Where is in inches when is in bbl/min, in cP, in ft, and in psi.
Fracture half-length (no leakoff, ):
Net pressure:
Note: In the PKN model, net pressure increases with time as the fracture grows longer and friction losses increase. This is a distinguishing diagnostic feature.
Fracture volume (one wing, average width = ):
Equations (With Carter Leakoff)
When leakoff is included, the half-length is reduced from the no-leakoff case. The Nordgren (1972) solution for fracture length with Carter leakoff uses the material balance:
Where is a function of fluid efficiency that accounts for the time-dependent exposure of fracture faces, and is the average fracture width.
For practical calculation, the length with leakoff is obtained iteratively from the material balance equation coupling width, length, and leakoff volume.
PKN Parameter Ranges
| Parameter | Typical Range | Notes |
|---|---|---|
| 200 -- 2,000 ft | Limited by leakoff and pump time | |
| 0.2 -- 0.8 in | Must exceed 3x proppant diameter | |
| 100 -- 1,500 psi | Increases during pumping | |
| 30 -- 200 ft | Set by stress barriers |
KGD Model (Khristianovic-Geertsma-de Klerk)
Physical Description
The KGD model assumes:
- Confined height: fracture height is fixed
- Horizontal plane strain: deformation is uniform along the height
- Length comparable to or less than height:
- Rectangular cross-section: width is constant along the height at any position
- Maximum width at center: width is maximum at (center of length) and zero at the tips
Plan View (top down) Cross-Section at any x
┌─────────────────────┐ ┌────────────────┐
│ tip │ │ │ ▲
│ │ │ │ │ │
│ tip ── well ── tip │ │ w(x) │ hf
│ │ │ │ (uniform) │ │
│ tip │ │ │ ▼
└─────────────────────┘ └────────────────┘
◄────── xf ──────────► ◄── w ──►
Width max at center Rectangular shape
Equations (No Leakoff)
Maximum width (at center of fracture, ):
Fracture half-length (no leakoff):
Net pressure:
Note: In the KGD model, net pressure decreases with time. This is opposite to the PKN model and serves as a diagnostic indicator.
Fracture volume (one wing, average width = ):
KGD Parameter Ranges
| Parameter | Typical Range | Notes |
|---|---|---|
| 50 -- 500 ft | Shorter than PKN for same conditions | |
| 0.3 -- 1.0 in | Wider than PKN for same conditions | |
| 50 -- 500 psi | Decreases during pumping | |
| 30 -- 200 ft | Must be comparable to or greater than |
Radial (Penny-Shaped) Model
Physical Description
The radial model assumes:
- No height confinement: fracture propagates radially from the wellbore
- Circular (penny-shaped) geometry: fracture is a flat disk
- Axial symmetry: width depends only on radial distance from center
- Maximum width at center: width is maximum at the wellbore and zero at the circular tip
This model applies when stress barriers are absent or when the fracture has not yet reached confining layers (early time behavior).
Plan View (top down) Side View (cross-section)
╱────────╲
╱ ╲ ────────────────────────
╱ ╲ ◜ ◝
│ ● │ ◜ w(r) ◝
╲ well ╱ ◟ ◞
╲ ╱ ────────────────────────
╲────────╱
◄──────── R ────────► ◄────────── R ──────────►
Circular front Ellipsoidal opening
Equations (No Leakoff)
Maximum width (at wellbore center):
Fracture radius (no leakoff):
Net pressure:
Or expressed in terms of time:
Note: Like the KGD model, net pressure in the radial model decreases with time.
Fracture volume (average width = ):
Radial Model Parameter Ranges
| Parameter | Typical Range | Notes |
|---|---|---|
| 50 -- 500 ft | Limited by eventual height confinement | |
| 0.1 -- 0.6 in | Typically narrower than confined models | |
| 50 -- 500 psi | Decreases during pumping |
Carter Leakoff Model
Theory
Carter (1957) described fluid loss from a fracture face with a time-dependent velocity:
Where:
- = leakoff velocity normal to fracture face, ft/min
- = leakoff coefficient, ft/min$^{1/2}$
- = current time, min
- = time when that fracture face element was first exposed
Leakoff Coefficient
The total leakoff coefficient combines three resistances in series:
| Component | Symbol | Physical Mechanism |
|---|---|---|
| Filter-cake (wall-building) | Polymer residue on fracture face | |
| Filtrate invasion | Viscous filtrate displacing reservoir fluid | |
| Reservoir compressibility | Pressure transient into formation |
Where:
In practice, the overall is often determined directly from a mini-frac (diagnostic fracture injection test).
Typical Leakoff Coefficients
| Formation | (ft/min$^{1/2}$) | Character |
|---|---|---|
| Tight gas sand ( mD) | 0.0001 -- 0.001 | Very low leakoff |
| Moderate sand ( mD) | 0.001 -- 0.01 | Moderate leakoff |
| High-perm sand ( mD) | 0.01 -- 0.05 | High leakoff |
Cumulative Leakoff Volume
The total volume leaked off from one face of one wing over the fracture area:
For a fracture with constant-rate propagation, this simplifies using the spurt loss and the leakoff area integral.
Fluid Efficiency
The fluid efficiency at any time is:
Where (total pumped volume).
Material balance for the fracture at time :
Rearranging:
Model Comparison
Assumptions Summary
| Assumption | PKN | KGD | Radial |
|---|---|---|---|
| Height fixed | Yes | Yes | No |
| Plane strain direction | Vertical | Horizontal | Axial |
| Cross-section shape | Elliptical | Rectangular | Ellipsoidal |
| Width varies along height | Yes (elliptical) | No (uniform) | Yes (radial) |
| Width varies along length | Yes (max at well) | Yes (max at center) | Yes (max at center) |
| Length/height ratio | N/A | ||
| Tip condition | No tip singularity | Stress intensity | Stress intensity |
Scaling Laws (No Leakoff)
| Quantity | PKN | KGD | Radial |
|---|---|---|---|
| Length/Radius | |||
| Width | |||
| Net pressure | (increasing) | (decreasing) | (decreasing) |
| Volume |
When to Use Each Model
| Scenario | Recommended Model | Rationale |
|---|---|---|
| Long fracture in thin pay | PKN | , strong barriers |
| Short treatment, thick zone | KGD | , horizontal strain |
| No stress barriers | Radial | Unconstrained height growth |
| Early pumping (any case) | Radial | Before fracture reaches barriers |
| Design optimization | PKN | Most common field scenario |
| Laboratory experiments | KGD | Short fractures in blocks |
Net Pressure Diagnostic
The trend of net pressure during pumping indicates which model governs:
log(p_net)
│
│ ╱ PKN: slope = +1/5
│ ╱
│ ╱
│╱───────────────── Transition
│╲
│ ╲ KGD/Radial: slope = -1/3
│ ╲
└──────────────────→ log(t)
- Rising net pressure: PKN behavior (fracture lengthening, increasing friction)
- Falling net pressure: KGD or Radial behavior (fracture widening, decreasing resistance)
- Flat net pressure: Equilibrium or height growth (not captured by 2D models)
Related Topics
- Hydraulic Fracturing Overview -- Model selection guide and workflow
- Proppant Transport -- Settling velocity and placement design
- Well Flow Overview -- Post-frac productivity and equivalent skin
- Productivity Index -- Fractured well productivity calculations
References
Perkins, T.K. and Kern, L.R. (1961). "Widths of Hydraulic Fractures." Journal of Petroleum Technology, 13(9), pp. 937-949. SPE-89-PA.
Nordgren, R.P. (1972). "Propagation of a Vertical Hydraulic Fracture." Society of Petroleum Engineers Journal, 12(4), pp. 306-314. SPE-3009-PA.
Geertsma, J. and de Klerk, F. (1969). "A Rapid Method of Predicting Width and Extent of Hydraulically Induced Fractures." Journal of Petroleum Technology, 21(12), pp. 1571-1581. SPE-2458-PA.
Carter, R.D. (1957). "Derivation of the General Equation for Estimating the Extent of the Fractured Area." Appendix to: Howard, G.C. and Fast, C.R., "Optimum Fluid Characteristics for Fracture Extension." Drilling and Production Practice, API, pp. 261-270.
Sneddon, I.N. (1946). "The Distribution of Stress in the Neighbourhood of a Crack in an Elastic Solid." Proceedings of the Royal Society A, 187(1009), pp. 229-260.