Impact of Aquifer Heterogeneity on LT-ATES Performance: A Case Study from the Chalk Aquifer, London, UK

The Chalk comprises a highly heterogeneous dual porosity aquifer, characterized by intervals of high permeability formed by fracturing and/or karstification of a low-permeability matrix. Many boreholes in London show evidence of a high-permeability flow zone at the top of the Chalk. Despite this, models used to predict ATES system operation in the Chalk aquifer in London have typically assumed a homogeneous aquifer, so that simulated warm and cold plumes have a simple cylindrical geometry around the wells. In this study, we investigate the impact of aquifer heterogeneity on system operation.

We examine an operating LT-ATES installation. The system employs 4 cold wells and 4 warm wells and is sized to deliver peak heating of 1.8 MW and peak cooling of 2.75 MW. Analysis of flowrate and temperature data shows that the system has a well-balanced energy ratio of 0.09 and exhibits a low but increasing thermal recovery which is currently ca. 40% for warm storage and 25% for cold storage.

We use a Surface-Based Modelling (SBM) approach to represent geological heterogeneity, which allows us to accurately and realistically capture geometrically complex subsurface features. We develop a range of parametrised 3D models of different geological scenarios, to capture uncertainty in geological heterogeneity between the wells. Flow and heat transport during ATES operation are simulated using the Imperial College Finite Element Reservoir Simulator (IC-FERST). The models are calibrated using Nelder-Mead methods to match pressure transient data and well inflow logs obtained from the boreholes prior to commissioning. Temperature and flowrate data collected during operation are subsequently employed in thermal simulations using the calibrated models.

Our findings suggest that aquifer heterogeneity has a significant impact on the formation of the warm and cold plumes. Heterogeneity has resulted in a thermal recovery bias toward warm water, despite ambient aquifer temperature being closer to the injected cool water temperature. Two high-permeability intervals play a pivotal role in the development of pancake-shaped plumes, as opposed to simple cylindrical plumes. Greater conductive heat losses to the overlying and underlying rock is observed with pancake-shaped plumes, resulting in lower thermal recovery. Recovery is predicted to increase as the temperature of the surrounding rock gradually changes through time. Heating and cooling demand on the ATES system is generally low, so the system predominantly utilizes just a single well doublet, with the choice of operational doublets varying through time. Thermal interference between the warm and cold wells in a given operating doublet may result from the laterally extensive plume geometry.

Our results indicate the necessity of recognising and modelling subsurface heterogeneity prior to ATES operation in areas with fractured and/or karstified aquifers. Designing effective operational plans entails incorporating considerations of local heterogeneity. Our insights have broad implications for future planning and design of ATES systems in the UK and globally.