| Abstract |
Drilling and field development in geothermal systems can be hindered by acidic fluids. It is essential to understand the source and mechanism of acidity in the early stages of exploration in order to formulate a suitable development strategy. The objective of this study is to understand the origin of fluid acidity in an andesitic geothermal system of Mt. Labo, Philippines, by modeling with the aid of geochemical software SOLVEQ-xpt and CHIM-xpt starting from the compositions of acidic fluids discharged from exploration wells in the postulated upflow zone. The core of this software is a geochemical speciation algorithm for solving mass balance and mass action equations at various temperatures and pressures by Newton Raphson iterations discussed by Reed (1998). The analyses of gas and liquid samples from the well discharge are combined to reconstruct the reservoir fluid conditions using SOLVEQ. CHIM is used to model the reaction of the reservoir fluid with the host rock or the well casing. The analytical data of gas and liquid samples obtained from the wellhead were recombined to the reservoir conditions (276°C, 6.04MPa), and then adiabatically boiled due to depressurization to 100°C and 0.1MPa using CHIM. Results of boiling show: 1) significant concentration of aqueous sulfate, bisulfate and sulfide, inferred to be from the disproportionation of magmatic SO2 forming H2S and H2SO4; and 2) formation of significant amounts of pyrite owing to the high iron concentration (157ppm) in the well discharge. To determine whether this large iron concentration originates from water-rock interaction or is due to casing corrosion, we reacted the host andesite with an estimated primary acidic fluid and found far smaller maximum aqueous Fe concentrations (38ppm) than observed in the actual waters. In contrast, the computed reaction of the fluid with casing yielded large Fe concentrations (390 ppm). This is consistent with the substantial actual casing corrosion observed in the blockage that was tagged in the slotted liner of the well. Modeling results show that the reservoir fluid pH increases after reaction with the rock. The water-rock ratio of the actual state of the reservoir, before it reacted with the casing, was approximated by comparing the acid alterations formed in the model to the actual acid alterations identified from the drill cuttings. The modeled alteration minerals such as alunite, pyrophyllite, quartz, pyrite, anhydrite and chlorite match the observed alteration (alunite, diaspore, pyrophyllite, quartz, anhydrite, chlorite, and pyrite) from petrologic analyses of drill cuttings obtained at the bottom of the wells where the temperature is greater than 270°C. The model shows that the most acidic condition had pH |