A well-based field development redefinition of unconventional reservoirs through unconventionality index: Utilizing machine learning

The intriguing aspect of this study is to include illustrative and realistic well-based matrix to efficiently evaluate, characterize and develop Unconventional reservoirs (UCRs). This research targets a newly assessment in redefining UCRs, and developing a well-based tool to evaluate, characterize and predict the performance of tight UCRs. In this study, permeability and viscosity are used to develop the Unconventionality Index (UI) to reflect the combined causes of low mobility from UCRs. Machine learning is applied to synthesize a novel comprehensive understanding of UCRs modeling. A distinct pattern is developed for to distinguish between UCRs and CRs to show the Recovery Factor (RF) / UI dependency. Consequently, to establish such relationship, data from major UCRs producers were examined and utilized. In addition, UCRs classification matrix has been developed utilizing actual UCRs data from different reservoirs. Furthermore, a unique Unconventionality Index has been established to classify UCRs, determine reasons of unconventionality and ascertain efficient method/s of development. Subsequently, a correlation between different rock and fluid properties incorporating UI and recovery factor has been attained.