Hydro-mechanical coupled discrete element modeling of geothermal reservoir stimulation and induced seismicity

Abstract

The injection of fluid underground, the use of CO2, water or waste for storage, or disposal purposes or both, results in a stress field change, the creation of fractures and the reactivation of pre-existing faults and joints. Sometimes these man-made processes are associated with seismicity, e.g. seismicity of a local magnitude 3.4 occurred in Basel Switzerland enhanced geothermal system (EGS) site. Such phenomena have lead to the development of numerical tools that are able to simulate fluid injection into underground reservoirs and predict induced seismicity. Appropriate measures for mitigating larger magnitude events (LME) can then be established after the reliability of the numerical tools is validated. In this context, this paper introduces hydro-mechanical coupled discrete element fracture network models developed for simulating hydraulic fracturing and induced seismicity. Particle Flow Code 2D is used in which the hydro-mechanical coupling routine is implemented, plus the seismicity computation routine. A fractured granitic reservoir with dimension of 2 km × 2 km is constructed using laboratory and field data from Soultz-sous-Forêts European Hot Dry Rock project. For mechanical and hydraulic data of the pre-existing fractures, measured data from Forsmark Sweden was adopted for planning the construction of the final repository for spent nuclear fuels. Hydraulic fracturing of intact reservoirs (without fractures) and fractured reservoirs is performed by means of a fluid injection at the model centre under two different scenarios: 1) a one day injection with a monotonic bottom hole pressure (BHP) increase followed by 1.5 days of shut-in where BHP decays non-linearly, 2) a one day injection with a cyclic BHP increase and the same BHP decays. Simulation results are examined in terms of: 1) fracture propagation pattern, 2) magnitude of induced events, 3) potential of LME in post shut-in, 4) influence of different injection schemes on the fracturing pattern and magnitude of induced events. The final objective of this paper is to examine if the presented modeling approach is capable of capturing the field observations and providing a good understanding of the key issues in the development of EGS, in particular for soft stimulation strategies. © Springer-Verlag Berlin Heidelberg 2013.