a b s t r a c tA failure investigation was conducted on a high-pressure machine with a large pump-turbine runner that was 2.9 m in diameter and had a maximum discharge rate of 32 m 3 /s.A part of the runner broke off during operation. This released a piece of the crown, which went through the machine, causing further damage. An analysis of the broken runner revealed a fatigue problem, so the dynamic/vibratory behavior of the runner during machine operation was investigated to determine the cause.First, the excitation forces acting on the runner during operation were studied. The main excitation in pump-turbines is generated by the interference between the rotating blades and the stationary vanes, known as a rotor-stator interaction (RSI), which produces large pressure pulsations. The amplitude of the pressure pulsations was measured in the prototype using pressure transducers.Second, the modal response of the runner structure was analyzed. A finite element (FEM) model of the runner was developed and the main natural frequencies and associated mode shapes were identified.A dynamic analysis was then performed to determine the runner response. A harmonic excitation simulating the pressure pulsation of the RSI was applied to the numerical model of the runner. The results showed a large stress concentration in the T-joint between the blade and crown where the crack originated. Finally, the possible causes of the damage are discussed.

No abstract

The dynamic behavior of Francis runners is of interest in the hydraulic machinery field since it is one of the most used type of hydraulic turbines for electricity generation. To evaluate the dynamic behavior of Francis runners, their natural frequencies have been studied extensively in the past. However, mode-shapes and damping ratios associated to those natural frequencies have not been studied in detail. The hydrodynamic damping, which is the damping induced by the surrounding fluid, plays an important role in the dynamic behavior of this kind of structures and it is challenging to be estimated.In this paper, an experimental investigation of a submerged Francis runner is presented. For this, a model of a medium head Francis turbine is used. The runner has been instrumented with a Piezo Electric Patch (PZTp), as well as with different accelerometers. The mode-shapes of the runner have been studied in detail in air and with the runner submerged in infinite medium of water and near a wall. Furthermore, damping ratios of every natural frequency have been obtained and compared for the different cases tested. Results have been also compared with a numerical model that includes the effects of hydrodynamic damping. The importance of considering the hydrodynamic damping and its influence on the mode-shapes of Francis turbines is shown.

The present study intends to investigate numerically the impact of: i) the fluid added mass, ii) the turbine rotational speed and iii) the variation of the runner blade bearing stiffness due to the turbine water head on the modal response of a reduced scale Kaplan turbine. The impact of each variable has been quantified by means of a series of numerical modal analyses of the turbine in vacuum and in water, and at rest and rotating. It has been found that the natural frequencies of the Kaplan turbine model are sensitive to all the variables investigated in the present paper, the added mass being the factor which has the greatest impact.

This paper presents an experimental modal test of a Kaplan turbine model and provides the corresponding analysis of the results. The modal test of the rotor including the runner, the shaft and the generator was performed using, as exciters, an impact hammer and a shaker and, as sensors, several accelerometers. Additionally, numerical models of the rotor with the runner surrounded by air (dry condition) or submerged in water (wet condition) were also built. By comparing the numerical and experimental results, the main modes of vibration of a runner blade in dry conditions and of the rotor shaft both in dry and wet conditions have been identified and discussed. The most significant deviations between experimental and numerical natural frequencies were found around 10% for the rotor shaft and around 6% for the runner blade. Moreover, it was observed that the fourth mode of vibration presents the highest added mass effect with a Frequency Reduction Ratio of about 6.7% and the second mode of vibration shows the lowest damping ratios in both dry and wet conditions with values of about 1.5 and 2.1%, respectively.