Estimation of casing material corrosion rates for supercritical geothermal development

Abstract

Supercritical geothermal development method has been proposed to increase the use of geothermal resources; the associated challenges include high temperatures and pressures, as well as an acidic environment, which increase the corrosion rate of casing materials. Here, we estimated material corrosion rates assuming that the temperature and pressure in a developmental environment were 500 °C and 60 MPa, respectively. Wells WD-1a in the Kakkonda geothermal field, Japan, and Icelandic Deep Drilling Program-1 (IDDP-1) were used as model cases for the concentrations of volcanic gasses. Geochemical calculations were conducted to obtain distributions of pH values from room temperature to 500 °C at 60 MPa. Corrosion rate distributions for the metallic materials were derived for the temperature recovery period after drilling and production and post-production periods inside the well. Consequently, the pH remained approximately constant at temperatures ≤ 300 °C and increased as the temperature exceeded 300 °C. In the supercritical temperature and pressure region, the pH increased as the pressure decreased, and it maximized at a 0.35 g/cm3 fluid density. The IDDP-1 model exhibited lower pH values than the WD-1a model; the H2S/CO2 ratio and HCl gas also increased. Furthermore, from room temperature to 300 °C, the corrosion rates were found to increase with the temperature and maximized in the subcritical regions at 300–350 °C. The corrosion rates then decreased as the temperature increased in the supercritical temperature regions. In the possible developmental processes, the corrosion rates tended to be high at temperature of 300–350 °C (~2,000 m) during the temperature recovery period after drilling, whereas the corrosion rates tended to increase near the wellhead during fluid production. The corrosion rate maximized near the wellhead immediately after the production period ceased at a well depth of 1,000–1,500 m the following week.