Engineering and Data-Led Sustainability for Real-Time Monitoring

Abstract

The environmental impact of oil and gas wells from exploration to abandonment is a primary focus during well planning and drilling. However, there is no industry standard to quantify the impact of complex indicators and associated variables, particularly when the well is drilled in real time. Thus, a quantification method using the associated variables used for real-time well monitoring as the well is drilled and produced can be useful. A new model for the sustainability index for well design and engineering during drilling is presented and validated. This proposed method avoids some of the vagueness of well sustainability and can be used and applied in a practical manner. It is based on various metrics and weightage assigned when a well is drilled and compared to the planned metrics. Evaluating the index for well engineering is based on elements of environmental impact, well design and engineering, the impact of functionality and optimization, well and maintenance costs, health and safety, and societal impact. Each element contains subelements. This process involves individual indexing through backpropagation of neural networks combined with the bat algorithm to obtain the echo location to identify the overall index. The proposed method uses a well-engineering-based approach that defines boundaries and thresholds as the well is drilled. The Powell Method, designed for nonconstraint optimization, is also used for fast convergence for this derivative-free problem. The algorithm is fast and provides results based on an iteration method. Because engineering during the well lifecycle involves several nonlinear system and asymmetric inputs, it has been observed that identifying the effects of input uncertainties and other related calculation uncertainties (i.e., variation and errors in log data, survey data, etc.) create an effect. It was also observed that uncertainty analysis provided an optimized index without assigning preferential weightage to some of the components. The method does not reduce uncertainty, but can be used to estimate the influence of various parameters on the sustainability index. To achieve runaway uncertainty, a mutation strategy is introduced to maintain the population diversity of the indices within the defined limits.