An Elliptical Envelope-Based Method for Casing Corrosion Diagnosis

Abstract

Casing corrosion leading to wall thinning or perforation poses a major threat to wellbore integrity, highlighting the urgent need for accurate corrosion diagnosis technology. To address this, this study proposes a casing corrosion diagnosis method based on the EllipticEnvelope anomaly detection algorithm. Utilizing well logging data, the method establishes a robust covariance estimation model via the EllipticEnvelope algorithm to effectively identify corrosion-related anomalies. Taking the data from Well XX in the Ordos Basin as an example, the proposed method achieved a detection accuracy of 92.3%. Compared to the Isolation Forest and threshold methods, this represents an improvement of 12.42% and 31.48%, respectively. The results demonstrate that the method can identify casing corrosion more effectively, providing a reliable technical means for downhole condition assessment.