The Use Of Corrosion Prediction Models In The Establishment Of A Treatment Protocol For Coal Bed Methane Wells

Recently reported CO₂ corrosion models were used to establish a treatment protocol for the production of coal bed methane. The effects of CO₂ concentration, pH, and shear are taken into account. Laboratory testing, computer modeling and a field case history are reported.

The expanded development of coal-bed methane gas reserves in the San Juan Basin in the Four Comers region of Southwestem United States began in the late 1980s. This gas from the Fruitland Coal formation is predominately methane with the range of C0₂ generally being from 1 to 15%. The majority of this gas is produced from wells with depths less than 1200 meters (4000 ft). Wellhead pressures are normally less than 6.9 Mpa (1000 psi) and wellhead temperatures are around 38° to 49° C (100° to 120° F). Although little or no liquid hydrocarbon is produced. formation water is almost always produced with the gas. A typical water analysis is shown in Table 1.

Because of the high bicarbonate levels ([HC0₃] > 12000 mg/l ) the pH based on simple field measurements was determined to be in the range of 6 to 9. Therefore, during the initial development of this coal-bed methane, it was concluded that corrosion would not be a severe problem. The fact that this conclusion was based on the misuse of criteria presented by Bonis and Crolet¹ soon became apparent when corrosion failures began to occur. It was at this time that Petrolite got involved in this field and began an investigation with this major producer in the San Juan Basin.

These failures were occurring in wells with open completions with gas production up the 7.3 cm (27/8") tubing. The failures were occurring from the internal side throughout the tubing string. However, all of the damage was occurring within 50 cm upstream of the pin end of the upset tubing. It was apparent from these failures (Figures 1 and 2) that the cause of the corrosion was due to C0₂ corrosion accelerated by turbulence, which was later described by Nesic and Lund₂. This led us to the discovery that we were using the corrosivity criteria of Bonis and Crolet outside of the valid range of conditions that they specifically designated. The problem wells were at or near the 9.1 m/sec (30 ft/sec) optimum for this model with the turbulence produced by the upsets adding to the effective shear at the failure sites.

Also at this time, investigations concerning the effect of flow and velocity on CO₂ corrosion were in progress. This work, reported by Hauser and Stegmann³, showed the effect of shear on the removal of the partially protective iron carbonate film and demonstrated the predictability of a laboratory test protocol which they developed. Using this test method called the High Speed Autoclave test, this study was expanded to include high bicarbonate brines. In these tests, mild steel test strips are rotated in a "cage type" apparatus inside a high pressure autoclave at 1500 rpm. Therefore, corrosion under high shear can be studied.