A Case History of the Hanford San Andres Miscible CO2 Project

Abstract

Summary A full-scale miscible CO2 project was implemented in the Hanford (SanAndres) Unit, Gaines County, TX. Unique aspects of the Hanford CO2 projectinclude the following:CO2 injection began only 1 year after initiation ofwater injection;the unit is run by an independent operator;the fieldhas an abundance of modern logs, core data, and pressure data available becauseit was discovered recently; anda successful technique for correctinginjection profiles was developed. The project's design, development, surveillance, and performance are discussed. Introduction The Hanford San Andres field was discovered in 1977. The field is locatedalong the northeast edge of the Central Basin platform in Gaines County, approximately 80 miles [129 km] northwest of Midland, TX (Fig. 1). Fielddevelopment used nominal 40-acre [16-ha] spacing. Primary developmentessentially was complete by 1979. Extensive high-quality data were gatheredduring development. Almost every well was logged with a full suite of openholelogs, usually consisting of a compensated-neutron/formation-density log, duallaterolog and microlaterolog, and dipmeter. Three wells were cored, andrelative permeabilities were measured with a number of core samples. Asubsurface fluid sample was caught before the reservoir pressure declined belowthe bubblepoint. Compositional and PVT analyses were made, and large amounts ofpressure buildup (PBU) data were gathered. Many of the PBU data were gatheredon an individual-zone basis. Table 1 shows reservoir and fluid properties.properties. Geology The Hanford field is located over a deepseated, nonproductive bioherm formedalong the shelf edge of the Central Basin platform. The bioherm was coveredgradually by shoaling-upward cycles of limestone. High-porosity, high-depositional-energy packstones and grainstones were deposited on theshoal. Dolomitization enhanced the original porosity of the shoals to form thehigh-quality reservoir rock in the Hanford field. The reservoir has an averagenet productive pay of 53 ft [16 m] within a gross interval of 200 ft [61 m]with approximately 200 ft [61 m] of structural closure (see Fig. 2). Theproductive interval is found at an average depth of 5,500 ft (1676 m]. Theoil/water contact (OWC), shown in Fig. 2, is tilted approximately 100 ft/mile[18.9 m/km] to the northeast. This is represented in Fig. 2 by the threecontour lines that cross the structural contour lines. The OWC tilt is believedto result from hydrodynamic effects; this is supported by data from Hubbert and Dutton and Orr. Unitization Unitization efforts began in 1980. Because miscible CO2 injection wasproposed during this time, waterflood and CO2 flood design took placeconcurrently. The unit was formed in April 1984. place concurrently. The unitwas formed in April 1984. Design Reservoir Description. The reservoir was described in great detail with logand core data, pressure buildup analysis, and geological interpretation. Fivedistinct zones of varying pay quality were identified in each well. Fig. 3shows a type log with each zone labeled. Structure, HCPV, and permeabilitydistribution maps were generated for each of the five zones. The permeabilitydistribution was based on a porosity/permeability correlation from the logdata. This log correlation was normalized with core and PBU data from Tixier'scorrelation. Fig. 4 shows a crossplot of log-derived permeability vs. coredata. Tixier's correlation was used for the initial design and gave acceptableresults. More recently, an improved correlation based on a more generalempirical relationship proposed by Wyllie and Rose was used. Fig. 5 shows thecrossplot with the improved correlation. Cross sections were built from everywell in the unit to ensure that all floodable and correlative pay wasperforated. Ultimate primary recovery was determined by decline curve analysisto be 17.9 % original oil in place (OOIP); this was supported by materialbalance. Waterflood/CO2 Flood Design. The waterflood was scaled to field proportionsafter modeling one injection pattern. Secondary proportions after modeling oneinjection pattern. Secondary performance of the entire unit was predicted withthe results of performance of the entire unit was predicted with the results ofthe model, which were calculated with the Modified Prats method, and thedetailed reservoir description of the field. Incremental secondary recovery waspredicted to be 14.2% OOIP. The CO2 flood design was part of the unitizationand water injection plan. JPT P. 924