| Title | Application of Numerical Methods for Geothermal Pressure Transient Analysis: A Deflagration Case Study from New Zealand |
|---|---|
| Authors | Katie MCLEAN, Sadiq J. ZARROUK |
| Year | 2016 |
| Conference | Stanford Geothermal Workshop |
| Keywords | pressure transient analysis, well testing, geothermal, numerical modelling, TOUGH2, PyTOUGH, deflagration, derivative plot |
| Abstract | Numerical models to simulate geothermal pressure transients are required in place of analytical models. This is due to the high temperature and complexity of geothermal reservoirs as compared to oil and gas reservoirs for which analytical models were developed. No numerical tool for this application is currently available to the geothermal industry. However the recent development of PyTOUGH for automating TOUGH2 simulations has enabled the use of TOUGH2 for geothermal well test analysis with relative ease. A standard model design has recently been developed using PyTOUGH, specifically for pressure transient analysis. PyTOUGH is used to automate the creation of the model grid and input file, running the model, extraction of results and comparison with field data. In this study the numerical PyTOUGH/TOUGH2 standard model setup is demonstrated through a case study using field data from the Ohaaki geothermal field, New Zealand. Investigation is made into the choice of method to evaluate model layer thickness and model injectate temperature, and the impact of these choices on the results. Field data used are pressure fall-off transients from injection tests performed immediately before and after deflagration of a well. The results of pressure transient analysis are examined to determine the effect of deflagration on permeability in the skin zone and wider reservoir. These results and the overall change in injectivity index are used to draw conclusions on the efficacy of the deflagration. The contribution of thermal permeability enhancement is also considered. The results from numerical models are compared with the results obtained by applying conventional analytical models. |