Rock physics modeling

The Wolfcamp Shale Formation in the Permian Basin of West Texas is a complete reservoir system of source, seal and reservoir. A deeper understanding of rock physics is necessary for successful exploration and production. Unconventional reservoirs are typically very complex to model, and rock physics provides insight into this complexity.

On a well located in the Wolfcamp Shale Formation, CGG applied a rock physics workflow that combines kerogen maturation information and petrophysical data to model seismic velocity. The maturation-oriented rock physics model reveals a relationship between kerogen maturity in shales and compressional-wave (P-wave) velocities. Therefore, by connecting kerogen maturation stages with elastic properties of seismic data, seismic volumes can be populated with kerogen maturity information.

Conceptual model

The presence of low-density kerogen and pores filled with hydrocarbons generated during different maturity levels is the main characteristic of shale gas reservoirs. Proper consideration of kerogen content and its maturation history within lithology sequences can be very helpful in elastic property modeling.

The kerogen maturity model is based on shale samples at mature and post-mature hydrocarbon-generating stages. In the initial depositional environment, organic matter and stiff silicate minerals are deposited together. In this immature stage, kerogen, clay and clean minerals are the only constituents of the rock matrix. Mechanical compaction follows; porosity is reduced; and pyrite appears in available pore spaces, marking the transition from an immature to a mature state.

FIGURE 1. The conceptual model used for kerogen maturity evolution is based on shale samples. This is the basis for a new rock physics model. (Source: Ahmadov, 2011)

During this transition, characterized by a particular orientation of the clay particles—vertical transverse isotropy—hydrocarbon generation begins. As they are buried deeper, shales enter into a mature state where most porosity is lost due to physical and chemical compaction, and pores start to be occupied by pyrite.

At the mature stage, organic matter occurs in bed-parallel elongated lenses, is load-bearing and contributes to the rock’s stiffness along with silicate minerals. In the final post-mature state, most pores will be occupied by pyrite. Organic matter is finely scattered and is no longer load-bearing. Figure 1 shows different shale source rock maturation stages resulting from this diagenetic process.

Porosity classification makes the rock physics model more flexible in studying different maturation states. Mineralogy is divided into three groups: clean (calcite, dolomite and quartz), clay and organics (kerogen and pyrite). For modeling purposes, total porosity is divided into hydrocarbon-filled porosity, bound water-filled porosity and free water-filled porosity.

This new model is applied to the Pence #1 well in the Wolfcamp Shale to model measured velocity and decompose it into the velocities of the clean, clay and organics groups. Required model inputs such as mineralogy and porosity come from stochastic petrophysical analysis of a conventional logging suite. Clay elastic properties are taken from previous rock physics studies performed in the same area along with rock physics parameters such as aspect ratios. Resulting velocities provide insight into the kerogen maturation stages along the well path.

When the system is immature, it is thought to be fully saturated with brine, while a post-mature system will be filled with hydrocarbons. Different types of porosity considered for calculation purposes are total porosity, hydrocarbon-filled porosity, bound water-filled porosity and free water-filled porosity.

Rock physics modeling workflow

The following steps are followed to model P-sonic velocity:

1. Calculate porosities for the rock physics model (total, effective, hydrocarbon-filled, etc.).
2. Divide minerals into clean, clay and organics groups. Pyrite is in the organics group because of the occurrence of pyrite within kerogen with increasing maturity (Figure 1).
3. Use Hill average method to model clean group elastic properties.
4. Use upper Hashin-Shtrikman bound to model the elastic properties of the organics group.
5. Use differential equation medium (DEM) and Gassmann methods to calculate saturated clean, clay and organics bulk and shear moduli. Here it is considered that clean and clay groups are saturated with brine while organics are fully saturated with hydrocarbons.
6. Apply upper Hashin-Shtrikman bound to mix the elastic properties of the clay and organics groups.
7. Finally, introduce spherical shale inclusions into the sand matrix using the DEM approach. This step ensures that both sand and shale act as load-bearing components.

Classifying mineralogy and porosity distribution among sands, clays and organics is helpful. Modeled velocities are in good agreement with measured ones. The main advantage of this model is the ability to decompose measured velocity into sand, clay and organics components. Measured velocity—sonic in this case—is an effective velocity coming from different lithologies and fluids with different elastic properties. Individual velocities for each group provide greater understanding of the maturation process in unconventional reservoirs. Some variations in the measured effective velocity may be more relevant to the clean group than to the clay or organics groups. Another advantage of this new rock physics model is that it provides a qualitative tool for determining kerogen maturity by segmenting parameters in the conceptual model into immature and post-mature states.

FIGURE 2. (a) Modeled velocity using rock physics model (red curve) is compared to measured velocity (black curve). Track 1 shows velocities for brine-filled clean (blue curve) and hydrocarbon shale (green curve) groups. Track 2 shows velocity components for hydrocarbon shale, which are clay (green curve) and organics (black curve) groups. (b and c) The images show two close-ups from the bottom sections of the logs between intervals 3,213 m to 3,240 m and 3,255 m to 3,271 m where velocities are overpredicted and organics group velocity is sharply increased. These intervals can be interpreted as containing post-mature kerogen (higher concentration of pyrite) based on modeling assumptions. Note that modeling for organics group velocity is independent of calcite and dolomite fractions. (Source: CGG)

Figure 2 shows application of this improved rock physics model for the Pence #1 well. The final P-velocity track compares modeled and measured P-velocities using the new model.

Furthermore, modeled velocity is decomposed into velocities for clean, clay and organics groups. These decomposed velocities clearly give more information on kerogen and sand content that cannot be deduced from the sonic velocity itself. Modeled velocity is well matched with measured velocity except for some intervals such as 3,213 m to 3,240 m and 3,255 m to 3,271 m (10,540 ft to 10,630 ft and 10,680 ft to 10,730 ft), where the organics group velocity (velocity given at step 4) increases sharply and modeled velocity is overpredicted (Figure 2). This overprediction comes from kerogen content (organics velocity) and could be linked to assumptions made during modeling. Therefore, these can be interpreted as intervals with post-mature kerogen.

Successful development and production programs rely on an advanced rock physics relationship between petrophysics, maturation parameters and elastic properties. This rock physics approach for modeling elastic properties in unconventional reservoirs is based on a conceptual model driven by actual source rock samples. This model integrates information about the maturation process with petrophysics and can be implemented in seismic characterization studies for mapping mature kerogen intervals. Better understanding of rock properties contributes to more accurate determinations of commercially viable areas for drilling.

References available on request.