Determining Statistical Properties of Quasi-steady Measured Signals

Quasi-steady measurement methods are becoming more commonplace in MARIN's model test basins for testing the open water performance of propellers, pod units and thrusters units. During quasi-steady measurements the carriage velocity or propeller rate of revolution is varied while thrust and torque are measured continuously. Compared to steady measurements, quasi-steady measurements foremost save time. Additional information can be obtained that would otherwise be unavailable, such as the effects of boundary layer separation and reattachment on the thrust and torque of a propeller model or thruster unit.Using a quasi-steady approach where the accelerating and decelerating part of an open water test are averaged, the supposed mean value of a measurement can be found as if it were steady. However, it was requested to determine other statistical properties such as the standard deviation and the occurrence interval as well. Unfortunately, this is not straightforward. This paper describes an approach to estimate the standard deviation and hence the occurrence interval for quasi-steady measurements. The methodology outlined may be extended to find other statistical parameters, e.g. kurtosis or measurement uncertainty.