Automatic selection of sites for drilling candidate injection wells

Background: The increasing difficulty in finding sites for drilling injection wells at the later stages of field development by NC “KazMunayGas” JSC, due to infill drilling of the grid of existing wells and uneven reserve production, is a pressing problem today. Developments in geospatial analysis and artificial intelligence have stimulated the search for new approaches to solve this problem.

Aim: The research is aimed at developing an innovative approach to automatically identifying the most promising sites for drilling injection wells, based on comprehensive analysis of large volumes of data using advanced algorithms.

Materials and methods: The work uses methods for collecting and analyzing production and geological data, uses spatial algorithms for multivariate analysis and data normalization methods, including the adjusted interquartile range to determine outliers.

Results: Results are described showing the ranking of cells by drilling potential based on comprehensive analysis, as well as the assignment of unique codes to each cell to improve decision-making accuracy.

Conclusion: Directions for further research are noted, including analysis of data inaccuracies, consideration of additional parameters, identification of effective interlayers, application of machine learning methods, and expansion of testing of the approach in other fields.