Measuring the Currently Unmeasured in a Resource Play: Model of Rulison Nuclear Re-Entry Well Shows Higher Than Expected Gas-in-Place

Brian Richter¹, Dan Simpson², and Stan Kleinsteiber^2
1Consultant, Denver, CO
²MHA Petroleum Consultants, Lakewood, CO

A reexamination of produced-gas volumes from the 1971 Rulison Nuclear test suggests that currently accepted Williams Fork Formation resource calculations are substantially pessimistic.

The authors constructed a 3-D dynamic flow model to test if Project Rulison site gas could migrate to future wells drilled within 1000 feet of the Project Rulison site. The model was constructed with currently accepted rock parameters and tested against known well performance. The results of the model fit well within the range of the current performance of area wells. However, a comparison of the results of the model with the results of the nuclear well production indicate that substantial volumes of gas are not being accounted in the Williams Fork through current production or modeling methods. This gas could be present in shales and coals or chemically bound to organic matter that are “beneath the cut-off” used in gas-in-place calculations.

Six wells had been drilled and completed in the model area including the Rulison nuclear emplacement well and the Rulison reentry well. To ensure a “real-world” stratigraphic match, the model was divided into six zones from top of the Williams Fork Formation to the top of Iles Formation. Each zone was assigned a dominant sand-body type based on field studies published by Cumella and Cole. Sandstone facies were assigned based on gamma-ray and density porosity cutoffs of 70 API and 6% porosity. The model was populated with sand bodies based on the three typical dimensions from Cole and Cumella. To ensure an acceptable match of the model to reality, sand histograms of each zone was examined and matched to actual-drilled wells.

Model rock-properties were calibrated from core data and from the Department of Energy Multi-Well Experiment wells to the north of the study area. Natural fracture permeability in the west-east direction was simulated by multiplying permeability in the north-south direction times 25. Hydraulic fracturing of wells was simulated by using a west-east permeability multiplier of 20 for 250 feet away from wells. A simulated cavity and fracture chimney, created by the Rulison nuclear detonation, was added to the model (measuring 100 ft x 100 ft x 350 ft, permeability = 1 md, porosity = 37%) and a zone of enhanced fractures (permeability of model times 5) measuring 2.8 times the radii of the chimney. These parameters were necessary to match the volume of gas produced by the nuclear re-entry well during testing in 1971.

The Rulison nuclear well produced substantially more gas from approximately one-third of the Williams Fork section, far more than any of the surrounding wells have produced from full Williams Fork completions over a similar producing period. The simulation model suggests the nuclear well liberated substantially more gas per volume of rock than was contained within the modeled rock volume. To explain the large amount of liberated gas produced at nuclear detonation site several theories have been proposed: extensive, very well connected fracture system that penetrates sandbody discontinuity related to clay drape and siltstone interbeds, extreme reduction in bound water saturation due to vaporization during nuclear detonation, pulverization of the chimney rock allowing full liberation of all gas-in-place.

The model results coupled with the nuclear well performance suggests that current gas-in-place resource estimates could be much lower than actual. The challenge will be to capture this resource.

AAPG Search and Discovery Article #90092©2009 AAPG Rocky Mountain Section, July 9-11, 2008, Denver, Colorado