Geysering Discharge of a Geothermal Wellbore at Zijnil, Guatemala

Within two hours after discharge began, for a production test in 1989 of well ZCQ-4, pressure variations changed from approximately sinusoidal to a cycle of sudden and complex peak discharge pressures, to 9 bar, separated by irregular pressure declines to 5 bar. Initial cycle periods of 42 minutes evolved to 150 minutes by day 20 of continuous testing, when three of four surge peaks were well separated. Chemical signatures of fluids discharged with pressure surges were distinctive. When combined with downwell pressure measurements, assignments can be made for elevations of fluid entry points. The variety of chemical signatures indicates a scarcity of interzone connectivity. These constrained discharges are suspected to derive from altered rubble zones between layered volcanic rocks. SElTING: Geothermal exploration of the Zuni1 area began in 1973 with a regional study by lnstituto Nacional de Electrificaci6n (INDE). Reconnaissaince geology, geophysics, and geochemistry, assisted by the Japan International Cooperation Agency (JICA) led to drilling eleven temperature gradient holes. These were followed by six deep geothermal wells, drilled in 1980-81 and subsequently tested 'with assistance of Electroconsult. Four of these deep wells produce neutral sodium chloride fluids with geothermometer temperatures in the general range of 240 to 280°C. In 1987, a contract to develop and engineer a 15 Mw power plant was awarded to MK Engineers, MK Ferguson Co., and Cordon y Merida Ings. Retesting of the wells was necessary and partial results for ZCQ-3, -4, -5, and -6 are reported by Menzies, et al (1990). A geologic summary is provided by Foley, et al Lnhologies at Zunil are a basement of granodiorite overlain unconformably by lava flows and ash-flow tuffs associated with a proposed caldera (Foley, et al, 1990). Volcanic rocks range in age from Tertiary to Pleistocene and involve four volcanic sequences, each with multiple flows and/or tuffs. The six deep wells all penetrate the volcimic series and intersect the granodiorite at elevations of 1090 to 1270 m ASL. Elevation differences are suspected due partly to faults. These faults are also suspected to channel upflow of hot fluids, although current production zones are varioisly within the volcanics and the rubble zone at the top of the granodiorite. Surface elevation at ZCQ-4 is 2117 m ASL and granodiorite was encountered at 1140 m. The completion interval includes three of the four volcanic sequences. Prior to studies in 1989, ZCQ-4 was tested on five occasions between March 1981 and September 1985. The apparent power potential of ZCQ-4 is the largest among the Zunil wialls, 5.7 Mw, including ccmsiderable excess enthalpy, (Menzies, et al, 1990), but marred by severe pressure surges. Surges (1 990). were present as early as 1982, substantially in the form found by 1989 testing, but described only as a graph. A brief description of the 1989 surging is available (Menzies, et al, The surging is important because its persistence and chemical variety indicate poor vertical connectivity among producing zones, surprising in view of suspected fluid production control through the field by northwest and northeast-trending faults. Geysering also complicates evaluation of the well's power potential and depletion forecasts. The relative independence of the geysering zones makes it impractical to interpret downwell pressure measurements in the usual way. Also, production zones experience different ratios of times for dischargehecharge. Furthermore, the chemical compositions of the production zones do not appear to stabilize, complicating forecasts of scaling tendencies, fluid enthalpy, etc. Severity of the surging makes the well an impractical supply to a power plant. Selection of a best remedial method requires good understanding of the sources and causes of the surging. EVOLUTION OF PRESSURE CYCLING: Figure 1 shows the evolution of the pressure cycling during the first four days. The figure is a composite of parts of Barton chart records of flowline pressures. Trough-to-peak pressure excursions amounted to more than 50 percent of the scale range of the Barton chart, nominally 5 to 9 bars pressure in the flowline. Corresponding flowline temperatures are 160 to 182 C. Time interval between similar points on the cycles was initially about 42 minutes, but increased to 78 minutes per cycle by about the 70th hour. Peaks were labeled 'A', 'B' ... in order of their appearance in the record. The relative intensities of the peaks shifted during the first few days, but the general features of the cycling were otherwise similar. By day 16, the pressure cycling had evolved to a period of 120 minutes and the peaks that had earlier been clustered into a group became well separated and individually narrow. On day 17, a unique pressure drop with an associated pressure spike (peak D) was 'inserted' during one cycle, at a chart location just preceding the highest pressure surge (Figure 2). It occurred with increasing regularity during the next two weeks and by day 34 it was a part of all cycles. Cycles without the D-peak were 120 minutes long, those with it were 150 minutes. This cycling of identifiable surges was also observed in 1982, but with longer cycle periods; 185 versus 120 minutes without, and 208 versus 180 minutes with 'D'. The longer periods in 1982 may be related to the longer preceding time of discharge, more than three months, compared to one month for Figure 2. Absence of the 'D' peak three months into the 1982 record suggests it may have been waning. Similarity of the cycling pattern of March 1982 and October 1989 indicates that the number and relative potency of feed zones to ZCQ-4 has not changed in any substantial way. 1990).