Corrosion in the SACROC unit water flood was serious enough to warrant remedial action. A few months after installation of the flood, some injection wells were filled with 75 ft of iron hydroxides, and serious corrosion was occurring in the high-pressure pumps. The problem was studied and several methods of treating proposed, and the most promising was chosen. This method consisted of designing and constructing a tower for stripping oxygen from the injection water with countercurrent flow of natural gas.
The SACROC unit water flood in Snyder, Tex., was initiated late in 1954, employing fresh water which was Pumped from Lake J. B. Thomas, treated with lime and alum for clarification and chlorine for bacteria removal, filtered and pumped into the Canyon Reef formation. The water was received at pH of approximately 8.4, and treatment was controlled to maintain that pH. Since this water was expected to be virtually non-corrosive, bare steel was used in the injection system, consisting of 34 miles of line which ranged from 3 to 10 in. Also, the water was injected through bare casing. This information has been published in detail. Within a few months, wireline work revealed fill-up of up to 75 A in some of the injection wells, and the fill-up material proved to be iron hydroxides. Moreover, serious corrosion occurred in the high-pressure pumps. Remedial action appeared essential.
The first remedial action taken was to raise the pH of the water to 9.3 in an attempt to reduce corrosion. Corrosion rates continued high, but the elevated pH was continued because of other operational benefits. Next, oxygen tests were conducted throughout the system, including using a pressure bomb down hole. The tests indicated that at least half of the oxygen (approximately 5 ppm) was lost before the water reached the wellheads, and usually only traces reached the bottom of the wells. At the time, approximately 150,000 BWPD was being injected. At 10 ppm of oxygen in the water, this represented corrosion of 1.220 ]b of iron per day. with formation of 2,340 lb of ferric hydroxide.
The oxygen tests down hole had some side benefits which were not anticipated.1. Attempts were made to remove the sample from the sampler by using kerosene. To test the accuracy of the procedure, duplicate samples were taken, one with a sampler, the other caught and run as a standard Winkler method analysis. The sample was displaced from the sampler and analyzed in the same manner as the other sample. Consistently, the sample from the sampler was high in oxygen compared with the other. Finally. an oxygen determination was run on the kerosene, and oxygen content was found to he extremely high. Since that time, oil blankets have been discarded as an effective means of preventing ingress of oxygen.2. An attempt was made to displace the sample from the sampler with natural gas, using the same procedure as above. Oxygen content of the sample was consistently quite low, indicating that natural gas was removing oxygen from the water. (Finally, the sample was chemically "fixed" in the sampler while flowing a volume of natural gas over the pipettes and sampler opening. Consistent results were obtained.)
Removal of oxygen with chemicals was considered. The chemical most commonly used at that time was sodium sulfite (Na2SO3). Stoichiometric combination of sodium sulfite with oxygen requires 8 lb/lb. To make the reaction approach completion requires both a catalyst, such as cobaltous chloride, and an excess of sulfite, usually 10 ppm. To remove 10 ppm of oxygen requires approximately 90 ppm catalyzed sodium sulfite.
Another method which was considered was the deposition of a scale to protect the pipe and casing. It has been proposed that this can be accomplished by adding 10 ppm of complex metaphosphates to a stable water and precipitating a thin complex film. However, during the first few months of operation of the SACROC unit system. this treatment was used and no corrosion benefit was noted.
As a result of the observation that contact with natural gas reduced the oxygen content of water, a laboratory pilot was set up to test the practicality of removing the oxygen with natural gas. The system consisted of a small packed column.