Rotor Winding Diagnosis Using Sweep Frequncy Response Analysis and Comparison to RSO

Rotor windings of rotating machines are prone to high mechanical forces, especially in high-speed machines, and therefore often cause unexpected failures. As the asset is a vital component in energy generation as well as industrial production, this can lead to high follow-up costs. One major root cause are shortages between windings of a rotor pole. A reliable diagnosis of the rotor winding is, therefore, essential to avoid such outages.The Sweep Frequency Response Analysis (SFRA) is a highly sensitive method to perform such measurements and is widely spread, especially in the world of power transformers. It was proven in several previous papers that the method can be applied as well on windings of rotating machines to check for inter-turn faults. This work targets influence factors of the measurement in practice as well as the possibility to assess the rotor winding condition reliably.To identify measurement influence factors and to identify the best frequency ranges in the SFRA measurement, 7 rotors from different machines with a rated power starting from 1,8 MVA up to 588 MVA have been investigated. On two of them, artificial short circuits between two or more windings have been implemented to investigate the sensitivity of the SFRA. Additionally, a comparison with the SFRA to the currently test implemented to detect such winding faults, the RSO (Recurrent Surge Oscillography) test was done.Having this amount of healthy and damaged windings, limits can be proposed, using a self-developed assessment algorithm. This was compared to the relative factor and the NCEPRI algorithm currently in use for transformer windings. The comparison to the Recurrent Surge Oscillography (RSO) method shows a clear advantage in sensitivity of the SFRA. Small, implemented faults where only detectable by SFRA and not with the RSO.