Columbia Gas Transmission operates two groups of natural- gas storage fields approximately 50 miles northeast of Columbus, Ohio. The Silurian-age Clinton sandstone formation in both fields varies significantly in thickness and reservoir quality. Although the gross Clinton interval is fairly uniform, it can include one to three separate sand members.
Reduction in storage-well performance over time dictates the need to periodically stimulate, or even re-stimulate the wells. Fracture stimulation is often required to restore well performance. The objective of the hydraulic fracturing treatments in these wells is to obtain a short, highly conductive propped fracture past the damaged area. Fracturing models are used to determine the optimum tip-screenout treatment design based on individual well characteristics.
Several fracturing fluids have been used in the attempt to achieve the optimum fracturing treatment. Previously, the best results were seen with borate-crosslinked fluids using nitrogen assist. With recent advances in fracturing-fluid technology, conventional borate-crosslinked fluids can be replaced with high-performance viscoelastic fluids, which reduce fluid damage to the proppant pack and formation. The viscoelastic fluid has excellent proppant-transport properties and predictable break times.
Three wells, two in Group I and one in Group II, were selected as candidates for fracturing with the viscoelastic fracturing fluid. Fracturing models were used to design a treatment that would result in similar fracture geometry as the borate-crosslinked fluids. This resulted in up to 40% reductions in total fluid requirements and up to 65% reductions in pad fluid volumes. All three wells were fracture-treated with the designs obtained from the fracture models. Fluid volumes recovered, flowing pressures, and choke size were all recorded during flowback and compared to values from offset wells.
This paper details the properties of the viscoelastic fracturing fluid, fracture designs, and post-fracture results of the wells stimulated with the viscoelastic fluid.
Columbia's natural-gas storage fields are located in close proximity to each other in Ashland and Richland Counties, Ohio, approximately 50 miles northeast of Columbus. These gas pools were initially discovered in 1910, and subsequently produced to economic depletion before being converted to storage. Group I was converted in 1936 and Group II in 1951.
The Silurian-age Clinton sandstone is the storage formation in both fields. Subsurface structure of the Silurian sediments is rather uniform and displays a regional dip east-southeast at a rate of approximately 50 ft per mile. The geologic setting occurred in a shore-face environment dominated by marine-tidal processes, and is represented by a series of depositional cycles of sand interrupted by periods of low energy associated with mud (shale) accumulation.
The sand varies significantly in thickness and reservoir quality throughout both fields. The depth of the Clinton sand averages 2,600 ft, and these fields are cycled at reservoir pressures between 300 and 1,200 psi. Although the gross Clinton interval is fairly uniform(Fig. 1), it can include from one to three separate sand members. The average (per well) total net pay is 12 ft. Porosity and permeability development is intergranular, and ranges from 8 to 16 %, and 25 to 125 mD, respectively.
Reduction in storage-well performance over time dictates the need to stimulate periodically, or even re-stimulate the wells in these fields. The decline in well deliverability can be caused by a number of mechanisms:
wellbore obstructions (i.e. fill and debris),
perforation tunnel restrictions (i.e. scales),
fines migration, and/or matrix permeability damage in the near-wellbore area caused by the buildup of solids and lubricants, and
various other pipeline- and compressor-maintenance fluids carried downhole during injection.
When the cleaning of wellbores and perforations yields unsatisfactory results, fracture stimulation is often required to restore well performance.
Maps of storage areas are shown in Figs. 2 and 3.