<p>This paper presents a novel fault detection approach for transmission lines using intrinsic time decomposition (ITD) and a threshold-based algorithm. ITD, a nonlinear and adaptive signal decomposition technique, effectively extracts fault features from non-stationary and transient signals by decomposing them into proper rotation components (PRCs). The proposed scheme is tested on a 3-machine sample system with faults simulated at various locations, including remote ends. The three phase input current signals measured in per unit are decomposed into PRCs using ITD and the PRC1 signal amplitude is used for fault detection. A threshold value is set based on various fault scenarios. The algorithm distinguishes between normal and faulty conditions, enabling prompt transmission of a trip signal upon fault occurrence. The proposed method is evaluated for different fault types, inception times and locations, demonstrating its effectiveness in detecting faults within millisecond. High impedance faults (HIFs), which pose significant challenges due to their unique characteristics are also addressed using a specific HIF model. The simulation results confirm the accuracy and robustness of the ITD-based fault detection approach, highlighting its potential for enhancing the reliability and safety of power transmission systems.</p>