Laboratory test of a superconducting gravimeter without a cryogenic refrigerator: implications for noise surveys in geothermal fields

Abstract

ABSTRACT Development of an accurate reservoir model and prediction of reservoir behaviour associated with geothermal exploitation using that model are fundamentally important for sustainable geothermal utilisation. Gravity observation for a few tens of days after a field-wide shut-in of geothermal wells potentially provides valuable constraints to improve reservoir models. We investigated the applicability of superconducting gravimeters (SGs) to short-term observations after shut-ins, as a first step towards the development of a method for reservoir model improvement using short-term gravity data. Results from a reservoir simulation of a shut-in at a geothermal field in Japan showed that gravity change at 20 days after the shut-in was 10 nm/s2. To detect such changes with SGs, locations of SG stations should be chosen to minimise noise sufficient to resolve the signal. Considering that field operations and traffic in geothermal fields can make SG data noisy, laboratory gravity measurements were carried out with a SG to investigate the influence of noise level on signal detection. Test masses, which produced a gravity change (7.7–9.0 nm/s2) comparable to the simulation result, were repeatedly placed and removed above the SG during both day and night times with different ambient noise levels. Test results showed that noise level greatly affected signal detection. The signal was only detected twice, under low-noise conditions with the standard deviations of 1.1 and 5.3 nm/s2. In addition, tests in which the SG was operated without a cryogenic refrigerator revealed that the absence of the refrigerator did not increase the noise level. Enhancing the portability of a SG by reducing its weight can make multi-location noise surveys feasible. Our method, of combining reservoir simulations and preliminary noise surveys, would enable successful SG observations to be made in relatively short periods after shut-ins, which would possibly contribute to the development of a method to obtain additional constraints to be used for reservoir model calibration.