Considerations in the Development of 3D Network Models Using Finite Difference Reservoir Simulatio

Simulation of water flooding and thermal enhanced oil recovery processes is advantageous to optimize the amount of oil recovered from reservoirs. Steam flooding is a proven successful oil recovery enhancement method for heavy oil operations. Field implementations have been very effective. However, steam flooding operations are capital and energy intensive and target lower valued crude due to higher associated refining costs. Thus, optimizing the heating rate and well allocation is essential in maximizing profit. Decisions on optimum heat rate and allocation are made on a recurring basis. Commercial thermal simulators consume considerable time and computational resources, which limits the amount of detail and the number of what-if scenarios that can be entertained. Therefore, there is a need to develop fast and sufficiently accurate thermal proxy models. While numerous flow-network models have been proposed, they are limited to 2D and simple physics. There is interest in extending such models to 3D and to include more detailed physics. Such models would only take the input and output of wells to simulate the performance of a reservoir. The objective of this thesis is to provide sufficient benchmarks to test and validate the plausibility of 3D flow network models in comparison to Schlumberger-Eclipse for waterflooding and CMG-STARS for thermal flooding. Several cases classes were generated and run on both CMG-STARS and a prototype 3D flow network model. Also, thermal simulators must account for heat losses due to overburden and underburden. This thesis provides a foundation for inclusion of heat loss inclusion from the reservoir for the 3D proxy models.