Use of Ranking and Clustering Analysis for Robust Well Placement and Efficient Reservoir Sweep

Abstract

Abstract With the advancement in technology and the growing availability of data sources, the size of data in the oil and gas industry is increasing at an exponential rate. While large subsurface datasets present more opportunities for field development planning, they also present some challenges in data mining and analysis. This work presents methodologies that can help unlocking the potential of reservoir models to drive insightful field development planning activities more efficiently. A simulation model of a synthetic heterogeneous reservoir is used in this work to evaluate reservoir opportunities for placing new wells. A 3D map of an opportunity index is first prepared using dynamic reservoir properties. Ten clustering analysis methods are then used to group the 3D map grid cells into different target regions. The generated regions are ranked based on their average reservoir opportunity index and hydrocarbon volumes. Ten new wells are then designed and placed in the highly ranked regions and then evaluated using numerical simulation. Results from the ten clustering methods are compared. Each clustering method resulted in a unique set of regions and hence, different locations and designs of wells. The comparison is performed by measuring the overall performance using cumulative oil production as well as by analyzing the clustering quality using silhouette score. The regional-based ranking resulted in an improved reservoir sweep. However, it may not be the optimal due to rig scheduling constraints since the selected regions are scattered. The analysis shows the impact of tuning the hyperparameters of the clustering methods and their significance on generating optimum target regions. This work highlights potential opportunities to improve the sweep efficiency in hydrocarbon reservoirs by capitalizing on data analysis techniques that complement conventional metrics and methodologies. It demonstrates an integration between clustering analysis and reservoir opportunity workflows which can lead to optimum field development planning.