Experience has shown that shale production and recovery rates from hydraulic fracturing vary unpredictably, not only from one well to another but also from one fracture stage to the next along a single borehole. Why?

For one thing, the material properties of shale reservoir rocks naturally exhibit high levels of vertical and lateral heterogeneity. In addition, local stress fields, pore pressures, and preexisting fractures or faults within or near the reservoir can have a significant impact on the variability of fracture stimulations.

For these reasons, operators of unconventional resource plays commonly use microseismic monitoring and imaging to help locate and characterize hydraulic fractures in existing wells, attempting to optimize both placement and completion designs in subsequent wells. Many engineers and geoscientists focus primarily on the location, direction, and length of fractures shown by microseismic events. Others, however, are beginning to investigate additional aspects of the microseismic response and to correlate microseismic data with surface seismic reservoir characterization.

Integrated analysis

Although the Montney formation of northeastern British Columbia, Canada, is generally considered a shale, it is actually a fine-grained siltstone with low-matrix permeabilities in the microdarcy range. To achieve economic gas production from a target interval at about 1,750 m (5,700 ft) true vertical depth, three wells with horizontal sections roughly 1,500 m (4,921 ft) in length were drilled and stimulated with six to eight hydraulic fracture stages.

Initial study of the location and magnitude of microseismic events recorded during all treatment stages (Figure 1) revealed three unexpected hydraulic fracture behaviors. First, a larger number of high-magnitude events occurred along one of the wells (C). Second, microseismic events extended mostly toward the southwest, while very few events were detected toward the northeast, despite being well within the detection range of the observation well. This was troubling because if hydraulic fractures grow in only one direction, productivity drops because only one side of a well drains the reservoir. Third, the geometry of microseismic events near another well (A) was quite different than that of the other two.

To better understand these observations, Progress Energy teamed up with Schlumberger to conduct a comprehensive reservoir study with three main components: enhanced seismic reservoir characterization, in-depth microseismic analysis, and engineering analysis of treatment and production data.

For the seismic reservoir characterization component, the team inverted 3-D surface amplitude-vs.-offset (AVO) data to determine Poisson’s ratio throughout the reservoir, which shows where the rock is under higher or lower stress. Fractures tend to grow more easily where low Poisson’s ratio values indicate regions of lower stress. In addition, the seismic reflection volume was run through an enhanced edge-detection algorithm known as ant tracking to assess lateral discontinuities.

To dig deeper into the microseismic responses, the team performed statistical analyses on two other valuable attributes: seismic moment density and frequency-magnitude slope. Seismic moment density, a robust measure of the strength of microseismic events, can provide insight into areas of higher or lower reservoir deformation. The slope of the relationship between frequency (number) and magnitude (size) of microseismic events – known as the b-value – can help distinguish responses associated with fault movement from smaller scale hydraulic fracturing. When fractures intersect preexisting faults, microseismic magnitudes are generally higher, while b-values are lower.

For the engineering component of the study, the team examined the instantaneous shut-in pressure at the end of each fracture treatment, which is considered a measure of the minimum principle stress. Also, daily production rates were examined to detect potential interference between wells.

Integrating these three components provided plausible explanations for the unusual behavior of hydraulic fractures in the three horizontal wells. For example, ant tracking indicated the presence of a northwest-southeast striking fault below the target reservoir, directly under the area of greatest seismic deformation. Therefore, the large number of high-magnitude microseismic events along Well C suggest that hydraulic fracture stimulation activated a vertical extension of this preexisting fault. The asymmetric propagation of fractures to the southwest appears to be controlled by lower Poisson’s ratio values in that direction, where both material rock properties and elastic stresses are somewhat different. Instantaneous shut-in pressure gradients observed in this area are consistent with lower stresses in regions of lower Poisson’s ratio values.

The relative lack of microseismic activity to the northeast of the wells appears to be the result of interaction between the hydraulic fracture and the northwest-southeast fault system, preventing further extension of fracturing in that direction. Finally, the geometry of microseismic events near Well A appear anomalous, mainly because they are distributed, as expected, orthogonal to the minimum stress direction without the fault activation observed around the other two wells. As a result, Well A has achieved the highest production of all three wells. Nevertheless, analysis of production data indicates that asymmetrical fracturing from Well B overlapped Well A’s fracture network, causing interference as both wells attempted to drain the same portion of the reservoir.

Geomechanical sweet spots

Greater understanding of the controls on hydraulic fracture generation and microseismic distribution enabled the team to accurately predict where to drill the next well in the Montney reservoir. All data indicate that geomechanical sweet spots in this area have two key characteristics. First, they exhibit low elastic stress; hence, lower injection pressures more easily fracture the rock. Second, there are no preexisting faults for injected fluids to activate, limiting normal fracture propagation.

The next drilling location, therefore, was identified by targeting a zone of very low Poisson’s ratio value north of the original horizontal wells (Figure 2) and by ant-tracking the seismic volume to verify there were no preexisting faults within or below the reservoir. The well was placed directly in the center of this predicted sweet spot. To further optimize production, the multistage stimulation design was modified to increase perforation density since microseismic events in the first three wells had shown gaps between the resulting fractures.

The reservoir in the new well fractured, as predicted, at lower injection pressures, and hydraulic fractures propagated symmetrically on both sides without any fault interference. As a result, shale gas production improved significantly. Since then, the comprehensive methodology described in this case study has been applied successfully to direct the placement of subsequent Montney wells.