Experimental performance examination of a coherence technique-based numerical differential current relay for AC machine stator windings protection.

A computational technique based on a coherence method for fault detection and classification for electrical machine stator windings is presented in this article. The coherence algorithm can identify clearly and concisely inter-turn and shunt faults situated on the 3-phase stator windings of the AC machine. Besides, it can categorize the different types of internal shunt faults. The cross-coherence algorithm performs the functional role of digital differential current to find and classify the internal faults; while, the auto-coherence algorithm acts as an overcurrent detector to define the occurrence of external, internal, or inter-turn faults. A new setup of three-phase induction machine stator windings, where each winding is re-winded to produce 20 taps per phase, is used to examine the approach. The new design is intended to build current transformers at the neutral and supply sides of the three windings, and to simplify conducting comprehensive examinations to verify the efficacy and efficiency of the advanced algorithm. The protection characteristics of the developed algorithm will be analyzed and estimated using the new setup. The test results indicate that the reliability and accuracy of the protection are above 98.7%. The coherence criterion is also useful for monitoring electrical faults, sensing inter-turn faults, distinguishing between external and internal shunt faults, classifying diverse internal shunt faults within the equipment protection zone, and estimating the tripping time when inter-turn faults occur. Furthermore, a new design of protection tripping-characteristic curves is established, and the time response of the computational technique is fast.