Field development planning and operational decision making relies heavily on models predicting the future behaviour of the reservoir. Reservoir models can often be greatly improved by history matching production data but only a limited part of the reservoir is observable from the wells. TNO creates the prospect of even better models by linking data from other sources, including time lapse seismic, subsidence and gravity.
The challenges faced by the oil industry this century include rising energy demand, scarcity of easily accessible and producible oil reservoirs, increasing complexity of operations due to environmental policies, market uncertainties, safety and human factors. Such challenges require truly innovative solutions, both in exploration and production to produce more from existing reservoirs, more effectively, safely and cleanly. Huge potential for innovation is offered by recent developments in technology, including computation and data acquisition and processing, with algorithms to translate this information to improvements in reservoir models. New data types that can be used for this purpose include time-lapse seismics, subsidence observations and gravity measurements. TNO has coupled different process models to reservoir models and measurements obtained by the monitoring of these processes can be used in inversion algorithms to reduce uncertainty in the reservoir model itself.
Time lapse seismic data can be inverted in two different ways to obtain improvements in the reservoir model. Firstly, by predicting reservoir pressures and saturations to acoustic impedance models for use in synthetic seismic modelling and comparing the resulting data to real seismic data. This approach was successfully used in a CO2 injection into the Sleipner field aquifer and in the migration of injected CO2 in another project where part of the monitoring campaign involved five seismic surveys whose data were compared with synthetic seismic data based on the reservoir models. In a second approach, field observations can be inverted to reservoir pressures and saturations using the Landro method, whose susceptibility to uncertainty is solved by adding a time-shift equation induced by changes in the P-wave velocity to pore pressure and in saturation: the inferred changes can then be used in an Ensemble Kalman Filter or other data assimilation technique to improve the reservoir model.
Subsidence has always been a major threat in the Netherlands, so it is unsurprising that TNO has conducted many subsidence field studies, including prediction as a consequence of reservoir compaction. The AEsubs software and geomechanical modelling software DIANA have been developed by TNO to this end, along with inversion methods to infer reservoir compaction, fault transmissibilities and compartmentalisation from subsidence observations. Gravity data can deliver valuable information on moving gas/liquid contacts, due to the large density gradients around these contacts, and recent developments in gravimetric instruments have greatly improved the accuracy of gravity measurements, which subsequently allows the use of these observations for history matching. Gravity signals can be measured both on the surface and in a borehole. Since the signal is not attenuated by any type of material, gravity observations are suitable in determining formation bulk densities beyond casing or cement.