Application of genetic algorithm in well placement for waterflooding

Abstract

Well placement is very important in waterflooding process. A good arrangement of well location does not only increase oil recovery but also reduces the time required to produce the oil. Either vertical or horizontal injectors can be used in this process. This thesis studies application of genetic algorithm in well placement for waterflooding process. A reservoir simulator coupled with genetic algorithm is used to find well locations for each case. The well locations depend on several parameters such as length of well, injection rate, and reservoir thickness. These parameters are varied in order to find the most suitable location. In this study, the scenario to obtain the highest recovery for each reservoir thickness is different. In case of a large reservoir thickness (more than 100 ft), the oil recovery is the highest when using a single vertical producer with two horizontal injectors at injection rate of 10,000 STB/D. This scenario provides high oil recovery and low water production. The production time required to produce oil is less than other scenarios. Although a single vertical producer with two horizontal injectors result in the highest oil recovery for a thick reservoir, this scenario is not suitable for a thin reservoir due to an early breakthrough. A large amount of produced water is turned out. For thin reservoirs, using one horizontal producer with two horizontal injectors with injection rate of 10,000 STB/D is the best choice.