Recent experiences in using TOUGH2 for geothermal modelling

Abstract

The geothermal modelling group in Engineering Science (University of Auckland) is involved with several geothermal R&D projects. On the development side we are running models of Ohaaki, Wairakei, Ngawha, Reporoa, Wayang Windu and Lihir. Our experiences in these projects have led on to several parallel research projects. Our model of Wairakei-Tauhara is so large that it also contains the Rotokawa system and the edge of the Ngatamriki system. Trying to understand the large-scale convection process at Wairakei-Tauhara has led on to studies of more generic convection studies and studies of larger areas of the TVZ. It has also led on to deeper models which require equations of state that can handle high pressures and temperatures. We have developed one for pure water but now wish to extend it to include CO2 and NaCl. With larger and larger models the computational demand increases quickly and we are now routinely using TOUGH_MP, the parallel version of TOUGH2. Also with large complex models dealing with input and output is a major task and we have developed a suite of PYTHON scripts (called pyTOUGH) for carrying out several model management tasks. One of the biggest challenges in geothermal modelling is model calibration and we have carried out studies using inverse modelling with iTOUGH and PEST and also Markov Chain Monte Carlo methods (MCMC). We are also carrying out miscellaneous studies of resewrvoir physics several of which involve fluid/rock interaction, for example the effects of cold water injection on permeability and subsidence in geothermal fields.