Reliability Assessment and Useful Life of Turbine Blades Using Probabilistic Methods

Steam turbines have numerous applications in various sectors of industry and it is known by experience that blade failures are the most common origin of breakdown in these machines, causing significant economic losses in turbomachinery industry. The turbines are designed to work in stable conditions of operation; nevertheless, failure in blades could appear after a short time of work. Failures are attributed to resonance of the blades to certain excitation frequencies. The vibration stresses reached by resonance conditions and the combination of other random variables could determine the useful life of the blades. In the deterministic design of turbines, the failure possibility is reduced in acceptable small levels by means of safety factors based on the good judgment. However, the possibility of failure could be reduced by using probabilistic methods. In the probabilistic approach, the variability in the properties of the material, tolerances in manufacture and uncertainties in the load are considered with statistical methods. The probabilistic method allows to evaluate the uncertainty or randomness present in some variables which translates in a high level of reliability in the results and a better operation analysis of the turbines. There are a lot of variables in the operation of the turbomachineries, in its design and construction. These variables are conceived under a certain degree of uncertainty, that is to say, these cannot be totally controlled. Generally, repeated measurements of mechanical phenomena generate a lot of input variables each one with certain instability in their magnitude that contributes to increase the probability of failure before the estimated time. In this work the reliability and useful life of blade steam turbines of 110 MW of the L-0 stage are analyzed using deterministic numerical methods and a probabilistic approach. Curves of the useful life for both cases are obtained. The probabilistic approach shows that failure occurs when a combination of variables and the presence of failure are combined.Copyright © 2011 by ASME