PSI - Issue 31

Sanjin Braut et al. / Procedia Structural Integrity 31 (2021) 33–37 Sanjin Braut et al. / Structural Integrity Procedia 00 (2019) 000–000

34 2

1. Introduction In axial turbomachines, the rotor blades are highly stressed by centrifugal forces and cyclic fluid forces, Cano et al. (2019), Cazin, et al. (2009). The fluid is often hot, corrosive and erosive, so the rotor blades are prone to damage. In the gas turbine, the loads are taken to extremes, so, the safety requirements are the highest, Collins (1993), Obolenski et al. (1989). To assess the effect of various loads, a simulation is often required, Vučković et al. (2018), Cazin et al. (2020) and Vukelić et al. (2020). It is important to consider the microstructure and effects of notches in the analysis of fatigue crack nucleation in components, Mlikota et al. (2017), Mlikota et al. (2018), Babić et al. (2018), Babić et al. (2019), Baragetti et al. (2019) and Baragetti et al. (2020), as well as crack propagation for remaining fatigue life assessment, Božić et al. (2010a), Božić et al. (2010b), Božić et al. (2011a), Božić et al. (2011b), Božić et al. (2014), Božić et al. (2018), Solob et al. (2020). Locati (1955) proposed a fast single-sample testing procedure for fatigue limit determination when the material S N curve is known or can be properly assumed. Zhang et al. (2003), modified the Locati method and explained how the requirement of, a priori known slope of S–N curve, can be eliminated. Gas Turbine Works Ltd. Karlovac, produces blades for aircraft turbine compressors. Testing of blades for fatigue strength was carried out to verify the production technology, Butković et al. (2016). This paper describes a procedure for accelerated fatigue testing of a gas turbine compressor blade based on the modified Locati method. The blades were made of aluminum alloy 2618 (T851). The tests were performed on a vibration shaker with the excitation frequency equal to the blade’s first natural frequency. First, S-N curves were estimated according to two approximate representation of Basquin type S-N curves, Stephens et al. (2000), Lee et al. (2005), namely one-slope and two-slope models. Then, the blades were tested at different loading (stress) levels starting at a level below foreseen fatigue strength. Each blade was subjected to an equal number of cycles, 2ˑ10 7 . At the end of the first block, the stress level is increased by 20 MPa and tested again until a new 2ˑ10 7 cycles are performed or until the blade failure is noticed. Blade failure was determined by the condition of lowering the natural frequency by more than 2%. After the failure, the damage was calculated according to Palmgren–Miner rule and considering multiple S–N curves. Finally, the fatigue strength of the compressor blade was determined as well as several basic statistical parameters. 2. Material and methods The cyclic strength of the real scale turbomachine blades is determined on one of the resonances that most often leads to the crack initiation. Here instead of a classical fatigue dynamic testing machine, electrodynamic shaker system is used. The natural frequencies of the blades on the shaker could be somewhat different from the operating conditions on the turbomachine, but this isn’t crucial when comparing fatigue strength. Determining high-cyclic fatigue strength is a time-consuming and therefore expensive process, therefore accelerated methods are used in industrial applications that give sufficiently reliable results, Lee et al. (2005), Obolenski et al. (1989) and Zhang et al. (2003). In this paper, the results of gas turbine compressor blade accelerated fatigue strength, are presented. 2.1. Material properties The third stage compressor blades of the gas turbine were made from aluminum alloy 2618 (T851). As this material doesn’t have a fatigue limit, the fail-safe life of 2·10 7 cycles is determined according to the client’s request, Recomendation OST (1979). Table 1 shows the basic mechanical properties of blade material, Basan (2011). The blades are made by precise forging, heat treatment, grinding, and finally polishing. For testing purposes, the blade root was not fully processed.

Table 1. Basic mechanical properties of Aluminium Alloy 2618 (T851) Density, kg/m 3 Young modulus, MPa Poisson’s ratio Yield strength, MPa

Ultimate strength, MPa

Elongation at break, %

2750

71 500

0,33

372

441

10