The Winter-Kennedy flow measuring method has been investigated on a test rig with a low head Kaplan model equipped with a semi-spiral casing. The effect of the runner and pressure taps positions on the method were the object of the investigation. The Winter-Kennedy constant was different for the old and new runner, inducing a difference as much as 2% on the estimation of the flow rate, when using the constant of the old runner. The results point also out different results function of the pressure taps position, indicating an optimal position specific for the present type of spiral casing. In summary, the Winter-Kennedy is inadequate to estimate the efficiency step-up of the following turbine after refurbishment with enough accuracy.