Lessons from larger than expected subsidence due to production of halite and natural gas in Fryslân

Abstract

Halite solution mining and natural gas production in the west of the province of Fryslân in the Netherlands cause subsidence significantly exceeding original predictions. Subsidence at the centre of the salt subsidence bowl reached 33 cm in 2015 while only 7 cm had been predicted in the 1995 forecast. Maximum subsidence above the nearby Harlingen gas field was around 33 cm in 2015 compared to 10 cm in the 2003 forecast for the end of field life. In addition, subsidence accelerated over time. From the regulator's point of view, the development of the discrepancies was unacceptable and led to an early reassessment of the salt mining project and to suspension of the gas production by the operator in 2008. Extensive research into the causes of the exceedances was able to provide a proper explanation for both occurrences. For the salt solution mining, cavern convergence rates at the in-situ stresses and temperatures at the depth of the project are much higher than initially estimated. This leads to a different production mechanism (squeeze mining), which gives much more subsidence. After convergence and subsidence modelling were corrected for this in 1998, predicted and observed subsidence are in much better agreement. The additional and accelerating subsidence above the Harlingen gas field can only be explained by pore-collapse of the gas-bearing reservoir chalk. During the laboratory compaction tests used for the original subsidence predictions, pore collapse only occurred at stress levels beyond those possible in-situ. The explanation for the inconsistency was found to be the strong impact of loading rate on the pore-collapse stress level. The applicability of an existing rate type compaction model for the Harlingen chalk was confirmed by new variable loading rate laboratory experiments. Using the model to translate the laboratory measured pore-collapse stresses into values applicable under field conditions, a much better agreement with the observed field behaviour was obtained. The Harlingen case thus provides further evidence that the large influence of loading rate on the chalk pore collapse stress seen during laboratory experiments applies down to field loading rates and that corrections for this need to be applied in compaction and subsidence forecasts.