Find in Library

Abstract Reservoir streamline numerical simulation, as an effective means to describe the trajectory of reservoir fluid and evaluate the flow size of different regions, has been widely used in various oil fields. To reflect the quantitative description of the streamline field in different development states, taking the different spatial positions and streamline attributes as the evaluation index, the reservoir flow field was clustered and evaluated combined with an artificial intelligence algorithm. The clustering results of different streamline clusters were evaluated using an evaluation coefficient. Finally, a quantitative evaluation method of reservoir streamline field simulation was formed. The results show that the peak density algorithm based on the evaluation coefficient can run efficiently in the reservoir streamline field and can quickly determine the optimal clustering number. The evaluation method was applied to the Nm3-4-1 layer of the Gang Dong oilfield. The current streamline field was divided into 14 grades. The average oil–water mobility ratio of the first kind of streamline cluster was much higher than that of the 14th kind of streamline cluster. After adjusting the streamline field, the optimal number of clusters was calculated by the evaluation coefficient again, and the number of streamline clusters was changed into seven groups, which indicated that the whole reservoir driving energy was more balanced, and the utilization degree of the reservoir was clearly improved.