Gas Turbine Short and Long Term Operation and Failure Mechanisms

摘要

In the last decades, gas turbine technology has evolved in pursuit of higher efficiency and increased flexibility. These objectives impose higher operating temperatures, higher pressure ratios and faster ramp rates. These tougher operating conditions impose increased stresses on the gas turbine components affecting their durability. In addition, longer operating intervals between scheduled inspections increase the risk of component deterioration and failure with potential secondary damage. Under these circumstances, detection of issues and their subsequent correction is fundamental to enhance the reliability of new or modified gas turbines. This process is part of the validation process applied to gas turbines prior to offering them to potential clients. The process can be divided into short and long term validations where different failure mechanisms and defects become apparent. Short term validation allows quick detection of unexpected issues resulting from difficult to estimate or simulate phenomena or from secondary effects overlooked during the design stage. Positive results from short term validation represent a “green light” for long term validation, where the time and temperature effects on durability can be evaluated. Low cycle fatigue, erosion, corrosion and fretting are among time/start-stop dependent phenomena while creep and aging of Thermal Barrier Coatings are among those phenomena that depend on time and temperature. These effects are very difficult to estimate based on a few hundred hours of operation and require long term validation in the order of thousands of hours. The long term effects mentioned above are commonly mentioned in typical Root Cause Analysis reports of significant component failures with associated downstream damage and must therefore be detected and corrected prior to commercial introduction. This paper discusses several deterioration mechanisms that can be detected during short and long term exposure to different operating modes and presents examples of issues that can be detected only during long term validation