| Title | VOLSUNG: INVERSE MODELLING AND UNCERTAINTY ANALYSIS USING PEST |
|---|---|
| Authors | P. Franz, J. Clearwater |
| Year | 2021 |
| Conference | New Zealand Geothermal Workshop |
| Keywords | Volsung, Geothermal, Simulation, Inverse Modelling, PEST, Cloud Computing |
| Abstract | The PEST suite of tools has long been established as a standard toolkit in groundwater and other geoscience modelling. Its inverse modelling and uncertainty analysis capabilities allow modelers to greatly shorten model development time, improve model calibration, quantify uncertainty in model parameters and determine confidence in model predictions. PEST tools have been used in the geothermal context, in particular in combination with TOUGH2; another noteworthy tool in this arena has been iTOUGH2. However the use of these tools has been greatly hampered by the complexity in setting up and running these types of simulations and the nature of TOUGH2 models. In particular, extraction of model observation data has been a really difficult task; in order to match real-world observations up to four dimensional interpolation has to be used. Since TOUGH2 does not store grid geometry this difficult task was left to the individual user. Further, parallel run management, i.e. the deployment of"workers" on different computational nodes, was also left to the individual modeler. In this work we introduce a universal coupling between the Volsung suite of reservoir and wellbore simulators and the PEST suite. It is universal in the sense that any model parameter and any model observation is accessible to the inverse modelling process. It fully supports data interpolation, i.e. the modeler can simply enter data by location/type and time. The modeler can select the adequate PEST tool for the task and use the graphical user interfaces to setup the inverse modelling problem. Run management is fully automatic and can create clone workers on cloud service platforms like Amazon Web Services (AWS). This reduces barriers to greater use of inverse modelling and uncertainty analysis in the geothermal industry. |