1. Stability of operation of Francis turbines Hydraulic disturbances often come with the operation of Francis turbines outside of design head and flow. In some cases, the excessive amplitude of the dynamic system response to these disturbances leads to restriction in the normal operation of the plant. Disturbances largely depend upon operating conditions and the particular turbine design. The response amplitude is under a strong influence from the feed pipe dynamics. Technical and economic consequences of a faulty stability of operation are such that the prediction of stability from model tests is one of the major challenges in present research on hydraulic machinery. Acceptance tests performed on a scale model in a hydraulic laboratory may provide the necessary elements for a full description of characteristics significant for the stability of operation of the full-size turbine in its piping system. This thesis is discussed in the report. 2. Dynamic characterization on the scale model The diagnosis for the dynamic behaviour associated with a machine design and given operating conditions is based on the observation of pressure fluctuations in various positions on the model and test circuit. Auxiliary measurements and surveys of cavitation, noise and vibrations support this data. This study presents a set of evaluation criteria for the elaboration of a diagnosis. The method is illustrated using practical examples from published works and from the numerous experiments performed by the author. A discussion of similitude shows that a diagnosis set using this method provides a good characterisation of the dynamic behaviour associated with the particular turbine design. Adapting acoustic power methods to Francis turbine tests provides a quantitative evaluation of disturbance sources. This tool makes it possible to look forward to an actual prediction of oscillation amplitudes associated with the full-size turbine operation. 3. Elements for the prediction The test laboratory provides a description of the dynamic behaviour associated with a particular turbine design in a guaranteed range of operation. This description is based on three elements of evaluation: discussion of the various dynamic phenomena according to operating conditions: relative frequency and amplitude of the part load precession, of 80 % load oscillation, relative frequency of draft tube column free oscillations, vortex-free region, organisation of the full load pulsation; survey of the draft tube cavitation compliance, lumped at the runner outlet, according to operating conditions; relative emission of the acoustic power toward the piping system, according to operating conditions. These three elements allow an adequate representation of the Francis turbine for a computation of stability of operation involving the piping system dynamics. However, this is the limit of what the test laboratory can say; the prototype piping system layout is liable to change after the turbine model acceptance tests. Once the turbine is adequately described the prediction of stability is performed using known methods and is done under responsibility of the technical consultant. 4. Conclusions The proposed method complies with technical and economic requirements of a laboratory acceptance test on a scale model. Following a simple experimental procedure, if yields a good description of the dynamic behaviour associated with a particular Francis turbine design and its guaranteed range of operation. It provides the elements for computations of stability of operation. Systematic use of the method and comparisons with field observations are the logical sequel to this research. This necessary phase of development will make the stability computations finer, for instance in handling damping in the connecting pipes. It will also provide the basis for a definition of scale effects to be used for the correction of similitudes.