Geological and petrophysical 3D modelling of sedimentary basins for groundwater applications, BRGM
Résumé
INTRODUCTION Since last 50 years, the French geological survey, capitalizes a large amount of diversified data recorded during campaign for water and hydrocarbons exploration and underground storage application (methane, nuclear-wast....). The past 30 years, the sociable and environmental needs have influenced research applied to groundwater exploration and characterization. Methodological improvements have allowed investigating new targets, such deep reservoirs and aquifers dedicated for new geothermal targets, alternative water resources and potential CO2 storage. Such panel of data mainly located in sedimentary basins (Fig.1 & 2) provides an excellent playing field for developing new methodologies in deep geology 3D reconstruction. Figure 1&2: Fig 1 (left): Geographic location of France's three main sedimentary basins, i.e. the Paris Basin, Aquitaine Basin and Southeast Basin. Fig 2 (right): Location of seismic lines (red segments) and deep wells (black points) available for studies of France's sedimentary basins. GEOLOGICAL DATA MANAGEMENT AND VALORISATIONS BRGM, the French geological survey, is the unique public organisation in France which manages the large amount of data acquired in the field (geological maps) and from the major exploration programmes in the sedimentary basins (seismics, deep drilling, airborne geophysics, gravity, ...). BRGM works since decades on databases management which is essential for producing integrated synthetic geological models. For the deep sedimentary basins (Fig.1), the BRGM has in particular been managing since 2006 the 350,000 km petroleum seismic lines and the 6 000 boreholes (Fig. 2). Raw data are accessible to all via a public "front office" (guichet.H@brgm.fr) and dedicated re-processing is possible on request. Beyond data management and delivery, BRGM has launched a vast data enhancement programme. The seismic lines are gradually being homogeneously reprocessed and interpreted and then assembled for providing major regional reference transects. The borehole logs are systematically digitised to well logging (Gamma ray, density, resistivity,...) records and petrophysical logs (porosity, permeability), thus enabling stratigraphical and petrophysical calibration of the geological models (Fig. 3 & 4). The methodologies are also implemented internationally for geological syntheses and basin modelling. Taking advantage of an easy access to data and a renewal of activity in the sedimentary basins, BRGM is now in a position to update existing geological syntheses and to produce digital models dedicated to various applications. Consequently, the geological models must be in 3D and compatible with the dynamic software used downstream by the hydrogeologists and reservoir engineers. Today, one of the France's sedimentary basins most studied for this kind of applications is the Paris basin. Figure 3: Example of well logs digitalised (Dogger aquifer, Paris basin), from the left to the right: Gama Ray, density, volume of shale in carbonate system (VSH), equivalent porosity (without and with shale effect), facies rocks descriptions UNDERGROUND APPLICATIONS IN PARIS BASIN Following the 1973 and 1979 oil crises, the development of deep geothermal energy exploded in France between 1980 and 1986. The Dogger carbonate formations of the Paris Basin have been the principal targets for such exploration. Today, in anticipation of the thermal potential drop of Dogger aquifers, the underlying Triassic aquifers are being studied. The water resources of the shallow aquifers present major pollution risks. It is now necessary to identify alternative water resources, and thus to explore the deep water reservoirs. Several projects exploring major fracture and/or karst systems in Dogger formations are currently underway in France's in three main sedimentary basins. Particular attention is also been paid to the major aquifers for underground storage - methane which is in constant development since the 1960s, and CO2. Major French and European research programmes have led us to reconsider the principal deep reservoirs of each basin in order to study storage capacities on a regional scale (unlike the storage of methane, localized on anticlinal structures). GEOLOGICAL AND PETROPHYSICAL MODEL PRODUCTION: OUR INTEGRATED METHODOLOGY The models made so far for groundwater applications, postulated a structure consisting of homogeneous geological formations. Today, our challenge is to restore the reservoir heterogeneity in three dimensions to build more realistic geological models. Yet, the characteristics of geological formation in the reservoir (porosity, permeability) vary greatly from place to place in the sedimentary system. To build these models, we relied on a methodology developed by petroleum engineer, working at oilfield's scale (a few square kilometers). We have transposed same approach on a regional and entire basin system scale (several hundred square kilometers, Fig.4). We integrate in models, deep data (seismic lines and boreholes) interpreted in terms of sequential stratigraphy (ie subdivision of the sedimentary pile into layers bounded by isochronous surfaces), paleogeographic maps (environments deposits and / or facies rocks) and geological map data (limits on geological outcrops). Based on this approach we are then able to produce paleogeographic maps and 3D digital models suitable for geodynamic reconstruction of the basins. Prediction of favourable zones for the exploitation of such formation is evaluated by petrophysical modelling. For each type of rocks, data analysis of boreholes (logs and core data) is made. These petrophysical data are estimated and/or simulated in the geological model using geostatistical tools. Our large scale models allow locating basins areas potentially suitable for CO2 storage or geothermal resources. When these interesting zones are located in the aquifer, smaller models can be rebuilt by upscalling of largest preliminary models. In the example below (Fig. 4), we produced a gridded 3D geological model at the basin scale, dressed with petrophysical properties (Fig.4). After potential areas for CO2 storage are localised, a regional model is built (downscaling). Both scales of models allow to make dynamic simulations of flow and injection of CO2 and to assess its impact in the media over time. Finally, these methodological results are now transferable to other Paris Basin's aquifer and obviously to others sedimentary basins (ie. the Aquitaine basin, South East Basin or, the Upper Rhine Valley, Fig.1). Figure 4: 3D gridded geological and petrophysical models (with Petrel). On the left cross section in geological model of the Paris basin (LogIso project, Author: S.Gabalda) on the right: regional model applied to Dogger aquifer. (ANR-SHPCO2, Author: S.Gabalda, Geology division, BRGM) Figure 5: Simulation of CO2 injection (with Tough2) into the Dogger aquifer (Paris basin, ANR-SHPCO2, Author: C. Chiaberge, Water division, BRGM) ON-GOING CHALLENGES: COMPATIBILITY BETWEEN GEOMODELLING AND SIMULATION SOFTWARES Because of the diversity of geological context, we made the choice of using several softwares. Considering that no software possesses all the required facilities for 3D geological modelling and simulation, we have to work with the complementarity between commercial (ISATIS, Earth Vision, PETREL, etc.) and in-house softwares (GDM Multilayer, Geomodeller 3D, etc.). One or other is chosen in function of the complexity of geological context and the aim of the 3D modelling. For example, Petrel (©Schlumberger) software is mainly used for 3D gridded models of sedimentary basin, dedicated for simulations with Eclipse or Tough. More complex geological context, like basin in front of mountain range (non-tabular structures, folds, diapirs, etc) are modelled using Geomodeller 3D (in house software, see an example in Fig.6). Ones of our tracks in research and development is to improve compatibility between geomodelling and simulations softwares (see Fig. 7). The workflow allowing to work with software and to transfer the job to another has to be fluent as possible. Main challenges lie on model compatibility and on workflow rapidity. Models need to be as complete as possible (geometry, petrophysical information ...) and their integrity have to be preserved from one software to another. Figure 6: Example of complex geological context for a regional groundwater study (with Geomodeller, Author: G. Courrioux, Geology division, BRGM) Figure 7: links between software improvement at BRGM, Geomodeller (©Brgm-Intrepid): models at large scale with few data, Petrel (©Schlumberger): reservoir characterization, Isatis (©Geovariance) : complex geostatistics, Eclipse / Tough2 / ToughReact: to compute simulation