Ensuring reliable oil and gas transport through pipelines remains a core engineering challenge, particularly in the face of expanding infrastructure and complex operating conditions. Conventional approaches often lack the real-time insight and predictive capabilities required for timely anomaly detection and effective maintenance scheduling. In this paper, we propose a digital twin-based solution that integrates physics-driven fluid and structural modeling with an Ensemble Kalman Filter (EnKF) for real-time data assimilation. Our framework continuously updates pipeline states based on multi-sensor feedback and applies a machine learning module to classify anomalies such as leaks, blockages, and corrosion. Through this synergy of physical simulations and data-driven analytics, early faults are identified accurately, and maintenance decisions are generated to reduce operational costs and prevent catastrophic failures. Experimental evaluations on multiple pipeline scenarios demonstrate improved detection precision and robustness, indicating the significant potential of digital twin technology for proactive and intelligent pipeline management.