To ensure experimental repeatabilit

To ensure experimental repeatability, a series of 11 tests is con- ducted. The tests are conducted with a 20 mm nozzle aimed at the centre of the cup with an inclination angle of 20 and a head of 3.5 m. After each test the nozzle is reset to a starting position be- fore returning to the test position. Two different torque transduc- ers are used in the tests, one with a full gauge deflection of ±20 Nm (7 tests) and the other ±5 Nm (4 tests). The results are shown in Fig. 13. The maximum small sample standard deviation rn1 of the re- sults at each speed is 1.5%, and assuming a normal distribution of experimental results, this suggests that 99.8% of the measured re- sults should be within ±3rn1 (±4.5%) of the actual value. The error increases with increasing speed because the torque values de- crease, thereby increasing the relative torque measurement error. In conclusion, the test method used here provides consistent results.