Numerical geothermal reservoir modelling and infield reinjection design, constrained by tracer test data : case study for the Olkaria geothermal field in Kenya

Abstract

Numerical simulation of geothermal reservoirs is a very useful instrument for developing strategies of field exploitation, production well location and reinjection scheme design. Reinjection of spent geothermal fluids serves to maintain reservoir pressure and enhance energy extraction efficiency over the life of the resource. Tracer testing is used as tool for tracing flow within a geothermal system for the purpose of characterizing the system and to gain good understanding on inherent heterogeneity. Tracer transport is orders of magnitude faster than cold-front advancement around reinjection boreholes and can be used as a cooling prediction tool. This study involved interpretation of tracer test data collected during cold injection into well OW-12 in the Olkaria geothermal field in Kenya. A single fracture model showed well OW-15 to be the most affected by the cold reinjection and OW-19 to be the least affected. A pessimistic version of the model predicts well OW-15 to cool by more than 20°C and wells OW-18 and OW-19 to cool by about 6°C for a forecast period of 15 years. An optimistic model version predicts well OW-15 to cool by 16°C, OW-18 to cool by 2°C and OW-19 to cool by 4°C for the same period.

A numerical reservoir model was developed for the Olkaria East and Southeast fields, covering an area of 27 km2 . The natural state model matched well available temperature and pressure data as well as being validated by production history data. Thermal front advance compares well for both models, it takes about a year to be felt in production well. Onset of cooling is immediate for single fracture model but in numerical model there is temperature rise followed by decline. This temperature increase before decline in the complex model is attributed to steam cap collapse. Well OW-12 in Olkaria can be used as a cold reinjection well, but it has to be used intermittently according to the results of the study, injection for one year followed by a period of recovery. The current injection depth in well OW-12 is shallow, but model calculations show that if reinjection depth is considerably greater (-2600 m a.s.l.), longer injection periods are possible without collapse of the steam cap involved.