Novel approach for modelling low enthalpy geothermal in deep sedimentary aquifers: a case study of 40 years of production data in the Dogger formation

Abstract

Geothermal energy in the Paris urban area has been exploited since the early 1970s. Deep drilling in the Paris sedimentary basin has targeted hot brine in the Dogger formations from a mid-Jurassic carbonate rocks series. More than a hundred wells have been drilled to depths ranging from 1400 to 2000 m and decades of production and reinjection have resulted in variations in the aquifer pressure and temperature in some areas. The regional numerical model discussed herein is aimed at assessing the potential interactions between doublets and for assessing the impact of the addition of more wells in the future. While the integration of a 3D refined geological model into a large-scale reservoir model needs heavy computational resources, local models simplifying the main productive areas into one or two layers surrounded by impermeable units can be used to approximately assess the reservoir characteristics. The modelling of a semi-regional area of the Dogger reservoir in the Southern part of Paris (Cachan/Orly) has been performed, using a conventional double-layer approach to simulate the natural state and production history of the area using the simulator Waiwera, a fast parallel open-source geothermal simulator from the University of Auckland in New-Zealand. Calibration of a natural state model, which describes the conditions before any geothermal operations, has been performed, including an analysis of the impact of boundary condition values on the ability of the model to match the data available in the public domain. The temperature and pressure distributions, instead of being based on geostatistical mapping methods have been obtained from a calibrated natural state model. The implementation of a regional lateral cross flow observed from past studies proved to be essential for matching the measured temperatures and pressures along with variations in the deep heat flux. A calibrated model has been obtained from matching the available production data with mismatches of no more than 1.5 °C. The modelling results confirm the dominance of the shallower productive zone in the aquifers and give insights into the extent of the cooler areas created by the long-term operations. Thus we have used the Dogger reservoir, with multiple data sets available, as a case study for calibrating a semi regional numerical model of a deep sedimentary aquifer used for geothermal direct-use. Our modelling study accounts for the conceptual understanding of the sedimentary aquifer with its heterogeneities and calibrates the numerical model against the measured historical data. Based on the calibrated reservoir model, the pressure and temperature responses in deep productive areas can be determined enabling operators or decision makers to test future strategies for sustainably operating the geothermal resource.