Analyse expérimentale de la stabilité hydraulique de la turbine Francis à charge partielle

Click to increase image sizeClick to decrease image size

The paper begins by recalling briefly the general trends observed in the literature and the meaning of stability. Taking that the torque fluctuation on the main shaft is the most synthetic index of turbine stability into account, the paper deals, on the one hand, with the relationship between the pressure fluctuation in the draft tube cone and spiral casing, and on the other, the torque fluctuation. Fluctuations are measured on a Francis turbine model. After recording the a.m. signals on tape, automatic analysis is carried out by use of a digital-analogie analyser which delivers the following information : - frequency spectrum of each signal ; - consistency of each signal in relation to the other; - amplitude transfer between two signals ; - phase transfer between two signals. Two main parameters govern these model tests : - the first one is the degree of cavitation with a constant operating head ; - the second one is the operating head keeping as a constant either the cavitation number or the dimension of the cavitating dead water core. Main results and conclusions are described below : - When cavitating dead water core increases, pressure fluctuations in the draft tube cone are aplified for all frequencies. Pressure fluctuations in the spiral casing are largely reduced for the dominant frequency, but much less for the first harmonie. As a result, the highest torque fluctuations are found for the first harmonie and not for the dominant frequency. - It therefore appears that thedominant frequency is produced by a local flow phenomenon (eccentric rotation of cavitating dead water core) and that the first harmonie is produced by a complete flow pulse due to the complex movement of the core in the draft tube elbow. The analysis of phases confirms that statement. - When the operating head changes, keeping as aconstant either the cavitation value or the size of the vapor value, the relative values of measured fluctuations do not have any significant trend. It can be observed that the lowest operating head does not lead to the lowest relative amplitudes of fluctuations. Nevertheless, observed dimensions of cavitating dead water core are not sufficiently developed to be able to draw a firm conclusion as to the effect of the operating head. The authors are therefore preparing an extension of the measurements to visual observatons in the draft tube elvow and analysis.