Stator Bar Vibration Sensors and Fiber-Optic Accelerometers

Vibration analysis has been conducted on large rotating machines for many decades. Proximity probes and accelerometers are traditionally used to measure bearing and shaft vibration in an effort to keep mechanical condition in check. Simple measurement of amplitude, to the more complex analysis such as Fast Fourier Transforms, are made by maintenance personnel in an effort to identify the source, severity and trend of vibration threatening proper operation of their generator fleet. Up until the last few years, vibration analysis technology has generally focused on external components, resulting in an incomplete method of identifying problems caused from internal movement. New technology is now emerging to help maintenance management take the next step – internal vibration analysis focusing on stator windings in the slots and end-winding region. The ability to conduct accurate and direct vibration analysis of stator winding systems can help operators implement a more effective condition-based and predictive maintenance approach. This paper will review field experience with two products that conduct on-line monitoring inside the generator: 1. Stator Bar Vibration (SBV) sensors used to measure in-slot displacement of stator bars. 2. Fiber-Optic Accelerometers (FOA) used to measure vibration of end-windings. Introduction Vibration analysis is the cornerstone of mechanical diagnostic monitoring on rotating machines. Well-placed proximity sensors can provide accurate and reliable vibration data on generator rotor shafts and bearings. This type of analysis is important for basic machine operation. It has been said that (for power producers), vibration sensitive generator rotors may cost hundreds of thousands of dollars in downtime [1]. Moreover, a major US generator manufacturer reports its installed fleet as “rapidly approaching the limit of the original intended life” [2]. This suggests that a large number of units are reaching an age where advanced on-line diagnostics is equal to, if not more important than traditional methods. Traditional vibration analysis does not provide a