03-01-2013, 12:53 PM
Failure analysis of steam turbine last stage blade tenon and shroud
1Failure analysis.pdf (Size: 1.28 MB / Downloads: 55)
Abstract
This paper presents an analysis of the cause of steam turbine blade fractures. Recently, several L-0 blades 28.5 (725 mm)
long of a steam turbine fractured 5 in. (125 mm) from the blade root platform, causing the forced outage of the turbine. A
finite-element analysis (FEA) of the blade was carried out in the beginning of the last decade to calculate the natural frequencies
and de vibratory stresses on the blade. A telemetry test was also conducted. The current investigation analyzed
the operational data during the last two years, reviewed the results of previous studies, conducted metallurgical investigations,
and identified the mechanical and metallurgical modes of the failure. The results of the investigations showed that
improper welding of the shroud to the blade was the principal cause of blade fracture.
Introduction
Steam turbines as any other rotating machine produce noise and vibrations, which may be assumed as
acceptable phenomena proper to rotation. Steam flows and spinning solid components, all of them contribute
to this noise in one or other level, either independently or through their interaction [1,2]. However, during normal
operation a strange noise was detected in a low pressure (LP) stage of a 37.5 MW steam turbine. Also, the
vibrations from bearings were noticed to increase considerably. Because of this event the machine was immediately
stopped and the casings opened. One part of the shroud corresponding to one blade group was found
out of place. It belonged to the last stage. Other loosed parts were found within the condenser where several
pipes suffered damages. Also the casing of the turbine was affected as well as much of the main body of the
turbine, which was observed through a careful inspection.
Review of the repair and operation history and visual inspection
Half a year before the accident, some repairs were conducted to the machine. The blades were modified by
adding materials looking for better protection against erosion caused by moisture, which had been formerly
detected. A stelite strip was installed to each blade while the whole blade group was welded to the shroud.
Originally, the blades were not protected with shield strips. Instead they were hardened by induction. Installation
of the strips must play some role since some of them cracked as well. Part of the investigation of the
failure in this work was focussed on proving that the natural frequency of the blade group was modified this
way. It is clear that adding weight to the shroud affects their dynamic behaviour.
Analysis
After the visual inspection some drawbacks were identified and may be attributed to the repair techniques
utilized. The cracks in the tenons of blades 3rd and 4th may indicate that the welding grade was not uniformly
applied to all the blades of the group. The excess of adding material during welding, as observed in Fig. 3 can
produce a weak union between the tenons and the shroud giving place to more vibration. Silver was used as
adding material during welding, which normally is not used for this task.
Through a careful inspection it is possible to observe beach marks on the trailing edge of the blades (see
Fig. 2b). Typical beach marks indicate that high cycle fatigue propagates the crack. Modal vibration analysis
reveals that an axial-torsion mode can contribute to generate high frequency vibration like the one responsible
for this kind of failure.