抄録
Two approaches were investigated for evaluating unreliability functions of failure events in steam turbine components. One is the parameter normalizing approach to establish the unique relationship between normalized hours or normalized cycles and unreliability for various classes of steam turbine output capacity. Another is the parameter adjustment approach to seek modifying coefficients as the functions of steam turbine design parameters for evaluating the specific unreliability functions from the common master curve. The cumulative hazard function method is applied to evaluate the actual field data fittings of those approaches for turbine rotor bowing and nozzle erosion events respectively. The data fittings through those approaches are good enough to apply the probabilistic risk analysis in the case of small sample size or number of units in the objective type of steam turbines.
