History Matching of Electromagnetically Heated Reservoirs Incorporating Full-Wavefield Seismic and Electromagnetic Imaging

Abstract

Electromagnetic (EM) heating is becoming a popular method for heavy-oil recovery because of its cost-efficiency and continuous technological improvements. It exploits the relationship that the viscosity of hydrocarbons decreases for increasing temperature; the heavy-oil components become more fluid-like, and hence easier to extract from the reservoir. Although several field studies have considered the effects of heating on the viscosity of the hydrocarbons, there has been very little research on the long-term effects of field production and the forecasting of the development of the reservoir. Increased flow rates within the reservoir render the moving fluids less viscous, implying fast-changing fluid-propagation patterns and increased uncertainty about the state of the oil displacement. This means, in the long term, strongly varying production projections, strong dependence on the permeability of the reservoir, and potentially undesirable fluid migration. To improve the forecasting of production in heavy-oil fields and to accurately capture the dynamics of the fluid movements, we present a history-matching framework incorporating well data and seismic and EM crosswell-imaging techniques. The incorporation of seismic and EM data into the history-matching process counteracts the changing reservoir dynamics caused by increased fluid velocity caused by heating and is shown to significantly improve reservoir matching and forecasts for a variety of different heating scenarios.

Enhanced characterization of reservoir hydrocarbon components using electromagnetic data attributes

“…These assumptions allow the EnKF to advance an approximation of the probability density functions of the state by simply advancing each member of the ensemble in time, making the EnKF computationally very attractive. Despite the Gaussian assumption and the updates being only based on means and covariances, the EnKF has shown to work remarkably well in many studies [19], [20], [59]- [61] and was therefore selected for this history matching study.…”

“…are the observed data and the ensemble mean estimate for data point i . The expression in equation (19) enables to quantify the reduction in the matching errors achieved via the incorporation of the EM data and all are measured with respect of the reference case. The results in Table 1 indicate significant improvements in matching production from the inclusion of EM data, outlining the benefits the improved depth imaging has on refining the characterization of the reservoir and enhancing reservoir history matches.…”

“…The study was preceded by a pilot test study by Marsala et al [17], [18] in which a EM borehole-to-surface survey was used to achieve better reservoir characterization. These developments outlined the growing usage of EM tomography for interwell imaging and complementing gravimetric and seismic technologies whose fluid discrimination ability is considerably less than those of EM techniques [19], [20]. Extensive research was further carried in the EM inversion for single boreholes outlining the necessity to model the salt concentration differences close to the wellbore [21]- [24].…”

Integrating gravimetric and interferometric synthetic aperture radar data for enhancing reservoir history matching of carbonate gas and volatile oil reservoirs

“…While the properties of the EnKF differ from those of the KF for small ensemble sizes, its approximation improves for increased numbers of linearly independent ensemble members. Although the EnKF updates are usually only based on means and covariances, the EnKF has been proven efficient for a variety of history matching problems even when implemented with small ensembles (Katterbauer et al 2015;Oliver and Chen, 2010;Aanonsen et al 2009;Liu and Oliver 2005). The EnKF operates in two steps: a forecast step integrating an ensemble of state vectors representing the uncertainties in the system forward in time and an analysis step to update the forecast ensemble with incoming observations.…”

“…EnKF techniques have gained momentum in the reservoir community, showing good performances for a variety of history matching problems (Katterbauer et al 2014b(Katterbauer et al , 2015Oliver and Chen 2010;Wang et al 2010;Thulin et al 2007). For reservoir applications, the joint estimation of both dynamic (e.g., pressure and saturation) and static parameters, such as permeability and porosity, poses however a challenge and may require several iterations with the EnKF to ensure sufficiently accurate estimates (Hendricks Franssen and Kinzelbach 2008;Moradkhani et al 2005;Zafari and Reynolds 2005;Wen and Chen 2007).…”