Methods for classification and evaluation of low-permeability tight reservoirs: A case study of the lower Yanchang Formation, Ordos Basin

The classification and evaluation of low-permeability tight reservoirs remain a pressing challenge in hydrocarbon exploration and exploitation. Traditional approaches are either tailored to conventional reservoirs or oversimplify the categorization of low-permeability tight reservoirs, limiting their practical applicability. Although some recently developed evaluation methods have shown promising results, their broad adoption is hindered by difficulties in parameter acquisition and high implementation costs. In this study, we examine the general reservoir characteristics in the study area using casting thin section observations, scanning electron microscopy (SEM), and high-pressure mercury injection (HPMI). Based on these results, combined with extensive data processing and multiple clustering algorithms, we propose a K-means clustering-based classification and evaluation method for low-permeability tight reservoirs. Furthermore, we mathematically define the classification boundaries using porosity-permeability cross plots. The proposed classification and evaluation system offers several advantages. By adopting an algorithm-driven approach, it overcomes the limitations of traditional experience-based criteria, thus providing more scientifically robust classification results. Although the proposed method integrates eight key parameters that capture reservoir physical properties and pore structure characteristics (i.e., porosity, permeability, sorting coefficient, median pressure, median pore radius, displacement pressure, maximum mercury saturation, and mercury withdrawal efficiency), the final classification and evaluation results rely solely on porosity and permeability, which are both the most indicative of reservoir quality and the most accessible. Therefore, this method addresses the limitations of traditional ones, including difficulty in acquiring evaluation parameters and challenges associated with widespread application. Employing mathematically defined classification boundaries, it avoids the oversimplified "one-size-fits-all" cut-offs inherent to traditional classifications. The method has been applied to the classification and play fairway prediction of low-permeability tight reservoirs in the Chang 7-9 oil groups in the lower Yanchang Formation, Dingbian-Fuxian area, Ordos Basin, providing a reliable basis for sweet spot evaluation in this area.