(Tech Connections Column, January 2007, American Oil and Gas Reporter)

Times are good, people are extremely busy, and management wants answers now. Those providing answers must not overlook basic reservoir engineering concepts. In an annual workshop series, PTTC’s Central Gulf Region has worked cooperatively with Core Lab to present insights in areas that don’t typically make the news, but can have a profound impact on the answers on which decisions are based.

Reservoir fluid properties are as basic as it gets. Field “surface” samples often are taken, data are gathered, and recombined properties are calculated. Before moving forward with this data, it behooves an operator to check their quality–is it a reliable sample? One approach is to compare sample data with theoretical data on a plot of log “k value” versus boiling point. The theoretical will be a straight line. The shape and location of the actual data versus the theoretical reveal whether a sample is good.

Unconventional gas reservoirs challenge reservoir engineers. Reliable reserve estimates are needed early in a play’s life, yet the very nature of the reservoir (tight, multilayered completions, and long times to achieve boundary-dominated flow) complicates conventional decline curve analysis.

One method is gas production analysis (GPA). GPA, which was presented in 1993 by Palacio and Blasingame and has been further developed by Agarwal and Gardner, is a type-curve matching technique. It provides an estimate of reservoir and completion properties (kh, skin, xf) as well as a comprehensive estimate of recoverable reserves.

There is a lot of science involved (see SPE papers 78695, 84814, 84491 and 98035), but let us move to experience-based conclusions. First, accurate reserve estimates can only be obtained after a well has established a constant drainage area (boundary-dominated flow). Second, static reservoir pressures measured during short shut-ins in low-permeability reservoirs should not be used for stand-alone material balance calculations.

Time to boundary-dominated flow is extended for noncommunicating systems if the system is treated as a single layer in the analysis. Hyperbolic decline forecasts with “b” exponents greater than 1 should not be performed. The good news though, is once an operator determines that he really is in boundarydominated flow (and that cannot be determined from observing plots), reserve estimates can be reliable.

A discussion applicable to unconventional reservoirs would not be complete without a reference to shale gas. An Appalachian Region newsletter (http://karl.nrcce.wvu.edu/news/vol7no1.pdf) captures insights from a shale-oriented presentation by Randy LaFollette of BJ Services at an Independent Oil & Gas Association of West Virginia meeting. LaFollette reinforced two key points:

• The Barnett Shale is a nonuniform reservoir.

•  Operational practices matter.

Maps of the first six months’ cumulative production from Barnett Shale wells indicate a fairly well-defined core area of the best vertical wells. This core area is in the southern part of the area where the Viola Limestone is present, near the southern pinch-out. More detailed maps of cumulative production, even in the core area, indicate there actually are two highly productive areas separated by a linear, less-productive area in which a fault has been interpreted. Production sweet spots appear to be controlled by several factors, including natural fractures, fracture barriers such as the Viola Limestone, faults, Ellenberger karsting, Barnett Shale oiliness, and the water-to-gas ratio.

Horizontal wells are being extensively applied in the Barnett, especially in areas beyond the edge of the Viola Limestone. It is easier to keep horizontals from fracturing into Ellenburger water. Sweet spots encountered by these horizontal wells occur in the same geographic areas as the sweet spots encountered by the vertical wells, and also in places outside the vertical well sweet-spot area, typically to the south. When production is compared to the azimuth and length of horizontal legs, two best-well trends are apparent, roughly 180 degrees apart, but the best wells are not always the longest wells.

Natural fractures typically trend northwest-southeast in a narrow fairway. Induced fractures typically occur in a broad northeast- southwest fairway, about 90 degrees to the natural fracture trend. In the core area and to the south in Johnson County, Tx., the better horizontal wells are drilled parallel to the natural fracture trend. The induced fracture orientation is perpendicular to the well bore and to the natural fracture trend in these wells.

The results are clear: The better wells are drilled parallel to the natural fracture system, not perpendicular. The caveat is that this relationship does not necessarily hold throughout the Barnett play, and operators need to understand the interplay between natural and induced fractures in their areas of interest before drilling.

Doug Patchen, RLO director for PTTC’s Appalachian Region, summarizes several other shale-oriented presentations in the same newsletter, so take time to read it. Readers also may want to reread last month’s Tech Connections column, which focused on Barnett Shale insights presented in a Texas workshop.