La Houille Blanche
Number 2-3, Mars 1968
|Page(s)||123 - 132|
|Published online||23 March 2010|
Étude du comportement mécanique d'une pale de turbine kaplan en régime transitoire
A study of the mechanical operation of a Kaplan runner blade under transient conditions
Eleclricité de France, Direction des Etudes et Recherches, 6, quai Watier, 78 - Chatou.
A very comprehensive series of tests has been run on a big Kaplan turbine at the Beauchastel power plant on the Rhone with the following main objects in view : __ (i) To try out special techniques, for example sticking strain gauges in high velocity flows, fitting multi-conductor wires inside the turbine hub, among the blade control links and through the governor tube, designing a slip ring for low background noise or a radio transmission system between the turbine runner and power house, etc.; (ii) Direct turbine runner blade load and stress measurement ; (iii) Comparison of stresses measured under various normal and abnormal operating conditions. Normal (i.e. everyday) operation is considered to include no-load running, stopping, emergency shut-down, and routine running, whilst cases of abnormal operation either seldom occur (runaway) or are of a new kind (relief or sluice operation). With the results of these tests, blade behaviour can be investigated as follows : __ a) By measurement of local stresses due to hydrodynamic forces acting on the blade; b) By calculation of the total hydrodynamic load components from measured local dynamic data and static calibration results. This also enables a link to be established between these results and the model data, and the model tests to be used for types of operation liable to be dangerous on the prototype. Special instrumentation used comprised thirteen strain gauge bridges connected to five-channel frequency modulation transmitting equipment and an eight-channel slip ring providingdata straight from the blade. AlI this information was recorded for both steady and transient conditions. As a general rule, the measured hydrodynamic stresses were low. At partial runaway condition with twice the normal running speed and half the rated discharge, the highest bending stress was 8.5 kg/sq.mm at a point on the blade centre-line a quarter of a runner radius in from the blade tip. These values would certainly be very much higher at the absolute maximum runaway condition. All the extreme values recorded were associated with major variations in the hydrodynamic operating conditions, e.g. with the guide vanes near the end of their closing travel, the flow cut off by the gate, maximum speed, etc. Compression and tensile stresses recorded under transient conditions at the blade leading edge near the root of the blade were low and invariably less than those observed for steady operation. Normal stress fluctuations were up to ± 25 % at low frequency (1 to 2 cycles/second). Alternating stresses were apt to occur with comparatively high frequencies (30 to 60 cycles/second), being of a low level (± 0.3 kg/sq.mm). They were recorded by all the gauges simultaneously when the turbine cut out at low speed (0.75 n0) and acted as a pump in the draught tube with the downstream in the process of closing. The components of the total hydrodynamic load acting on the blade and the coordinates of its P.O.A. were calculated from these results for both steady and transient conditions. A check on the values found in this way showed them to be representative to within 10 % of the actual blade stress values. The inference from these calculations is that the total hydrodynamic load component normal to the blade depends on the same relationships as bending stress, confirming that the latter play a decisive part compared to torsional stresses.
© Société Hydrotechnique de France, 1968
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.