Title: Production capacity assessment of the Bacon-Manito geothermal reservoir, Philippines

Abstract

A sustainable production capacity of the Bacon-Manito geothermal system (BacMan) is assessed in this study
by three numerical models of different complexity. A base case of 150 MW electrical generation is considered
and production is constant up to end of Geothermal Service Contract for BacMan (year 2031). The numerical
models considered are based on volumetric, lumped-parameter and a full scale, 3D well-by-well methods. A
conceptual reservoir model is first proposed based on previous geoscientific research and downhole data. The
geothermal reservoir has an estimated area of 23 km², its thickness exceeds 1500 m and temperatures range
from 240 to 320°C. A volumetric model, using Monte Carlo style simulation, indicates that a
production capacity of 200 MWe can be maintained for another 25 years with 90% probability. The
lumped model predicts an annual pressure drawdown of 0.67 bars, resulting in a manageable total
drawdown of 25 bars in year 2031. For the well-by-well method, a distributed parameter numerical
model was developed using the simulator iTOUGH2. The simulator reduces poor matching between
observed and simulated response by optimizing a set of 15 model parameters. These include mass and
enthalpy of hot and deep recharge and 12 permeability values. Optimization resulted in far-field
permeability of 0.5-5 milli-Darcies while the productive wellfield ranges from 25 to 100
milli-Darcies. The deep recharge was calibrated as 98 kg/s of 1830 kJ/kg enthalpy. Sensitivity
analyses show that the model is most sensitive to pressure drawdown data followed by enthalpy of
flowing wells. The model predicts that at least five more production and three reinjection wells
are needed for future high-pressure steam requirement of the 150 MW powerplant. Stable enthalpies
of production wells are predicted for the 23 years studied, indicating that reservoir temperature
drawdown will be moderate. Reservoir boiling will, however, be intensive and change pressure
gradients from hydrostatic to vapour static in the centre field. Instead of predicting reservoir
performance for tens or hundreds of year, it was decided to stop production in year 2031 and
monitor recovery of heat and mass reserves. Lumped model predicts a year of recovery for every year
of production (optimistic scenario) while distributed parameter model predicts two
years of recovery for every year of production (pessimistic scenario).