Rough Gas Storage Site – Redeveloping and Making it Hydrogen Ready

Abstract The Rough Field UKCS SNS started gas production in 1975 and was converted to gas storage in 1985. In 2017 it was partially decommissioned, and it entered a blowdown phase. In early 2022 the Rough Field was brought back into storage operation due to changing energy requirements. However, due to the age of the Rough facility, a redevelopment was required, and it was decided to future-proof the facility and wells by making Rough ‘Hydrogen-Ready’. A new static reservoir model was built to enable a calibrated dynamic model to optimise future well placement for both gas and hydrogen storage. The initial dynamic modelling was done using black oil formulation to speed up the calibration phase and the final optimisation for well placement was carried out using compositional modelling. The static model was expanded to a ‘whole earth’ model for thermal and geomechanical analysis to assess the risk of potential hydrogen and or natural gas leak path generation. After several iterations of the static model, testing various uncertainties, a calibrated dynamic model was achieved with small variations to the static properties porosity, permeability, and transmissibility between facies and across the faults in the reservoir. To achieve a good pressure match, a ‘ghost’ gas volume was required otherwise the amplitude of the pressure cycle during injection/withdrawal cycles would have been impossible to achieve. Possible sources of the ‘ghost’ gas volume could be a larger than expected Carboniferous volume, limited pressure support from the surrounding aquifer or an isolated gas accumulation adjacent to the field. The future well count to achieve injection and withdrawal targets is 16 wells, this is less than the historical 36 due to the introduction of horizontal well technology. The model calibration was challenged by less than perfect data gathering during the storage phase and the bottomhole pressures were estimated using tubing head pressure data. The temperature and geomechanical modelling indicated that past operations past had most likely not caused any damage to the reservoir integrity. The study highlights the requirement for comprehensive data gathering during the operation of any field. The assumption that a reservoir behaves like a tank is an oversimplification and can lead to incorrect assumptions of the reservoir mechanism (the requirement for the additional volume is not evident in material balance models). A ‘whole earth’ model would be ideal for any reservoir study and the perceived cost is justified by a much-improved reservoir understanding.