Fitness for purpose assessment after an over pressure incident at Arapuni Power Station

Abstract

In December 2000 the wicket gates of Arapuni Unit G7 closed rapidly when a control arm separated from a shaft and the wickets were free to move. The unit had been operating under normal conditions at the time. The resulting overpressure pulse caused the lower portion of the penstock to become barrel shaped, and a longitudinal riveted joint partially failed. A significant amount of cracking was visible in the cast steel stay vanes and deformation had occurred at the rivets in the scroll case. A program of work was instigated to determine the cause of the failure, and its consequences, i.e. the pressures reached and what work was required to ensure that the unit was fit for reuse. The root cause of the failure was the design and fabrication of a control arm installed when the wicket gate control system had been modified. This control arm had two segments which clamped around a shaft. The clamping bolts used had seen a high residual bending load and cyclic loading. They failed in fatigue. The coating on the inside of the penstock and scroll casing was brittle and an assessment of the degree of craze cracking in the coating on tensile test specimens allowed an accurate method of determining the plastic strain that had occurred in the penstock, and hence the stress and applied pressure that occurred during the incident. The major issue was the condition of the stay vanes and the rivet holes in the scroll casing. The cracking in them was quantified by NDT inspection. Most of the cracks seen in the stay vanes were original casting defects that had not propagated during the overpressure. An engineering critical assessment was carried out in accordance with recognized international standards to determine the effects of defects under brittle fracture and overload conditions for a range of pressure conditions. An initial assessment was performed using conservative estimated values for the applied stresses and materials properties. The initial calculation indicated that the units were probably fit for purpose. Mechanical testing on samples taken from the stay vane casting and strain gauging during a proof pressure test was carried out. This allowed the calculations to be refined and the confidence in the result to be increased. A fatigue assessment was then carried out in accordance with the standards. The stay vanes and the scroll in the unit were shown to be fit for purpose without any modification and fatigue was not predicted to be a major problem. The unit was returned to service without any modification to the cracked stay vanes. An inspection procedure and regime was also defined to ensure the unit remains fit for purpose.