Applying Machine Learning NLP Algorithm for Reconciliation Geology and Petrophysics in Rock Typing

Abstract

Oil- and gas-bearing rock deposits have distinct properties that significantly influence fluid distribution in pore spaces and the rock's ability to facilitate fluid flow. Rock typing involves analyzing various subsurface data to understand property relationships, enabling predictions even in data-limited areas. Central to this is understanding porosity, permeability, and saturation, which are crucial for identifying fluid types, volumes, flow rates, and estimating fluid recovery potential. These fundamental properties form the basis for informed decision-making in hydrocarbon reservoir development. While extensive descriptions with significant information exist, the data is frozen in text format and needs integration into analytical solutions like rock typing algorithms. Applying text analysis, a crucial area in natural language processing, aims to extract meaningful insights and valuable information from unstructured textual data. With the vast amount of text generated every day, automated and efficient text analysis methods are becoming increasingly essential. Machine learning techniques have revolutionized the analysis and understanding of text data. In this paper, we present a comprehensive summary of the available methods for text analysis using machine learning, covering various stages of the process, from data preprocessing to advanced text modeling approaches. The overview explores the strengths and limitations of each method, providing researchers and practitioners with valuable insights for their text analysis endeavors. A data-driven rock typing scheme is necessary for decision-making and optimization to achieve the best ultimate recovery of hydrocarbons in the most efficient way.