7.3.5   Rotor fast fracture risk assessment


There have been major advances in the quality of monobloc rotor forgings (see Section 7.2 of this chapter) and in understanding the problems of brittle fracture. Rotor forgings can now be produced with greatly enhanced fracture resistance and the presence of significant defects can be readily detected ultrasonically.

However, large steam turbine rotor forgings cannot be produced entirely free of small, residual defects and, for the high temperature rotors under consideration, the forging ductile/brittle transition temperatures are not significantly below the temperature range experienced in operation. The possibility of a rotor brittle fracture cannot be dismissed without a careful assessment.

Having carried out a 100% volumetric NDT investigation, the defects identified must be assessed for acceptability. Acceptable defect size is defined primarily by the requirement that it shall introduce no risk of rotor brittle fracture in service. The initial defect assumed on the basis of the inspection standard will be subjected to centrifugal and thermal stress cycling. Stress cycling could cause the defect to grow by fatigue. Extensive testing of materials has permitted crack growth rate to be expressed in fracture mechanics form (see Fig 1.113).

Typicat fatigue crack growth data for various turbine steel between 100 and 550 C

The operational duty of large generating plant is such that upper bounds can be reliably established for stress cycles.

In actual service, the stresses during start-up are limited by established operational procedures (see Chapter 2) to protect the plant against high strain fatigue. The total fatigue growth can then be calculated on the basis of the upper bound stress cycles and combined centrifugal and thermal stress.

High temperature steam turbine rotors operate in the creep range and cracks within them could therefore grow under steady loading. Rates of growth under creep can be correlated with the linear elastic fracture mechanics parameter K, via an equation of the form (see Fig 1.114):

The effect of temperature on creep crack growth resistance of a turbine rotor forging

The design criteria normally adopted for high temperature turbine rotors limits the maximum rotor stress in relation to the 105 hour creep rupture stress and limits the accumulated creep strain to 0.2% in 105 hours, ensuring that the stress in critical high temperature regions is acceptably small. Calculated creep crack growth is therefore small and may be simply added to calculated fatigue crack extensions.

The region of maximum combined centrifugal and thermal stress, near the rotor centreline, coincides with the region where, for reasons of ingot cooling and heat treatment, the material fracture toughness is lowest. Valid plane strain fracture toughness specimens cannot be obtained without removing an unacceptably large diameter core and samples taken elsewhere may not provide a reliable basis from which to estimate centreline properties. Consequently it is most common to measure toughness indirectly by using Charpy FATT specimen results. Established correlations between FATT and fracture toughness (see Fig 1.115) are used to determine a KIC for the rotor material.

Established correlations between FATT and fracture toughness

In all cases, the crucial judgement to be made is whether the rotor forging can, given the longest feasible initial crack and maximum fatigue and creep growth during the specified lifetime service, survive without failure caused by instantaneous fast fracture. This will depend on whether the extended defect is longer or smaller than the critical defect size calculated for the most adverse combination of events. To further complicate the assessment, variations of rotor temperature during machine start-up produce a corresponding variation in rotor forging fracture toughness properties which influence the instantaneous critical crack size. Assessments must therefore be undertaken at each critical operating condition for complete confidence.


<<- Previous entry                  Table of contents             Next entry ->>