Optimized treatment design shows promise

Challenging low-permeability oil reservoirs in the Ordos basin in north-central China offer proof that in-depth reservoir understanding, advanced design and evaluation workflows, and appropriate completion and fracturing techniques result in production increases of as much as 80% to 100% over baseline. A pilot project recently conducted in the basin set out to quantify production improvements in two horizontal wells and three vertical wells placed between two horizontal wells equipped for real-time microseismic fracture monitoring.

An unconventional fracture model incorporating geologically and geophysically defined natural fracture patterns and mathematical interaction criteria for hydraulic and natural fractures was used to optimize treatment design. Real-time microseismic fracture monitoring through a dual-monitoring well setup was performed on a total 26 stages of fracturing treatments. Simultaneous treatments pumped in four stages were compared to other stages pumped in chronological sequences.

Ordos basin challenges

Although historical field development in Ordos basin tight oil reservoirs was considered economical, the average production after hydraulic fracturing was low, and most of the wells were producing only marginally. Studies indicated that insufficient connected and conductive fracture surface area induced by conventional fracture completion techniques was the major reason for low productivity.

Yet even by increasing the fracture treatment size and proppant mass pumped into the formation, no obvious incremental production gains were achieved using conventional fracture design methodologies. This pointed to the need to improve the understanding of fracture mechanical propagation behavior and the proppant distribution mechanism in the reservoir to break through the production barrier status quo.

Pilot project in Chang-7 formation

FIGURE 1. The pilot project’s horizontal laterals were drilled orthogonal to maximum horizontal stress to fracture across the laterals. (Images courtesy of Schlumberger)

To accomplish this objective, the pilot project’s two horizontal laterals were drilled orthogonal to maximum horizontal stress orientation to enable propagating transverse hydraulic fractures across the laterals as shown in Figure 1. To obtain accurate microseismic data, dual monitoring was established using two out of the three monitoring wells simultaneously.

The horizontal wells were drilled in the lower Chang-7 formation, which is very fine sandstone grading to siltstone with average permeability of 0.24 mD and an average porosity of 10.5%. Natural fractures are widely distributed across the basin; however, their nature and characteristics vary from formation to formation and from block to block. The two horizontal wells were the first in this reservoir. The vertical wells that were completed initially produced 31 b/d to 179 b/d after conventional fracturing treatment with crosslinked gel and about 50 to 120 tons of proppant placed. Production from the vertical wells declined quickly in three months to fewer than 18 b/d and could hardly produce economically afterward. From previous post-closure pressure decline analysis it was determined that effective permeability can be as low as 0.01 mD. Extremely low permeability and insufficient fracture surface area for oil production were considered the main causes for low production.

New simulation workflow

FIGURE 2. By correlating the features from several well logs, the structural variation along the laterals between the monitoring wells was identified.

To achieve the objective of maximizing the reservoir-fracture surface contact area in this basin, it was necessary to have an accurate understanding of how the fracture was penetrating into the reservoir, including complexity, effective length, fracture containment within the pay zone, and fracture conductivity for hydrocarbon production. It is well known from other tight oil plays that reservoir contact area can be significantly improved with multiple transverse fractures for horizontal wells. However, even with transverse fractures, the well potential could only be maximized with contribution from all of the pay intervals and optimized proppant distribution along the lateral.

Rather than using traditional planar fracture models based on a 1-D geomechanical model, the new Mangrove reservoir-centric stimulation design software was introduced for the pilot project. The software, which is a plug-in for the Petrel E&P software platform, uses an integrated reservoir-centric workflow to enable seismic to simulation as well as optimal completion and stimulation scenarios to enhance production from unconventional reservoirs. The “design-execution-evaluation” cycle in this workflow is distinguished from previous conventional capability by its ability to simulate complex natural fracture and hydraulic fracture interaction via new mathematical models in tight reservoirs.

For example, the 3-D geocellular earth model captures sedimentary, stratigraphic, and geomechanical details in the reservoir and bounding layers. In the past these models have been used solely for geologic study and reservoir simulation. However, in the new workflow the hydraulic fracture completion design based on the 3-D geoceullar earth model is coupled with production prediction based on the hydraulic fracture gridded onto the upscaled reservoir model. This enables the full usage of the model and a much more realistic picture of hydraulic fracturing behavior and its impact on well productivity.

FIGURE 3. The surfaces extracted from well correlation show structural variation.

For improved accuracy, model calibration plays an important role in this workflow. While 3-D earth models and complex fracture models provide details in real-world dimensions with deliberate data inputs, model accuracy is improved through fracture and production behavior calibration.

Workflow steps are integrated into a single work platform and can be implemented routinely for single well or multiple well design. Calibration cycles are performed during the fracture treatment if necessary. The workflow is tailored according to how much data are available, and the complexity of the 3-D models is chosen to enable the capturing of adequate details with acceptable computer time consumption.

In this pilot project, geostatistical well correlations to reflect the detailed formation variation along the laterals were established through the three vertical wells placed between the two horizontal target wells. By looking at the features from several well logs, a series of well tops – including the top and bottom of the target Chang-7 formation – was identified to reflect the structural variation along the laterals. After building a structural model, a 3-D zonal model was built according to the surfaces extracted from well correlation. Geologic structure and reservoir property variation were captured. Properties for each zone were determined based upon the properties of the monitoring wells along the laterals.

FIGURE 4. The 3-D zonal model with minimum horizontal stress was determined based upon properties from the monitoring wells along the laterals.

To capture the basic characteristic and quantitative modeling of the natural fracture system in this area, a discrete fracture network (DFN) approach was used based on geologic knowledge. A manual DFN model was initially created to represent the natural fractures. With the DFN model and 3-D geomechanical model, the unconventional fracture simulation result suggested basic planar fractures with some simple fracture networks by opening natural fractures and creating new fracture branches during the main fracture propagation. The DFN model can be further calibrated according to microseismic fracture monitoring results.

This unconventional fracture model can explicitly model fracture propagation in 3-D. Hydraulic fracture-natural fracture interaction, stress shadow effect between nearby fractures, proppant transportation in fracture networks, lateral stress, and property variations are considered.

New workflow results in increased production

FIGURE 5. The initial production result for both horizontal wells (wells A and B) shows a substantial increase as compared to the vertical well on the same pad.

Initial production tests on both horizontal wells showed 783 b/d and 648 b/d, respectively, significantly higher than all the wells completed in tight oil reservoirs in the Ordos basin. Production normally varied from 31 b/d to 50 b/d in vertical wells and to 201 b/d for horizontal wells, on average. The initial result showed substantial production increase compared with the production from the Chang-6 formation of the nearest offset reservoir as shown in Figure 5, which actually had slightly better reservoir quality. Compared with the best horizontal well, the initial results showed an 80% to 100% production increase.

The stabilized production rates after three months were 5.5 and 3.9 times those of the same vertical well. Results showed more than a 50% increase compared with the horizontal well completed in the same formation in the northern part of the basin.

This article was prepared from SPE article 158268, presented at the SPE Annual Technical Conference and Exhibition held in San Antonio, Texas, Oct. 8-10, 2012.