A Simulator for the Recovery of Heavy Oil From Naturally Fractured Reservoirs Using Cyclic Steam Injection

Abstract

ABSTRACT This paper describes a simulator which has been developed for the specialised application of predicting the performance of cyclic steam injection, when used to recover heavy oil contained in naturally fractured reservoirs. The potential application of the simulator will be for evaluating fractured carbonate reservoirs, which contain a large proportion of the known heavy oil resources. The reservoir is divided into a rock matrix/fracture system with the rock matrix assumed to be described by a number of cubic blocks surrounded by interconnecting fractures. The flow of injected steam and its condensate is limited to the fractures and heat transfer to the rock matrix is only by conduction. The temperature of the matrix blocks and the heat lost to the overburden/underburden are obtained using an approximate analytical solution to the heat diffusivity equation. The simulator then predicts the rate of interchange of fluids from rock matrix to the fractures by the processes of thermal expansion and countercurrent imbibition. Production rates of oil and water from the thermally stimulated well are calculated, based on a PI which is adjusted to account for the temperature dependent fluid viscosities and the fluid heights in the fractures. During the production period, the heat loss to the overburden/underburden and the produced fluids is calculated at each time step. Results from the model show the advantage of a fracture system in conducting heat into the rock matrix, because of the large surface area exposed to heating. The simulator can be used to investigate the interaction between various reservoir and input parameters in an attempt to devise the best reservoir management strategies. Initial results show the importance of steam slug size, steam injection rate, steam temperature and applied drawdown in attaining the best economic operating conditions. The matrix block size and permeability were found to be the most important of the reservoir variables changed in the sensitivity analysis.