The 29th Manoeuvring Committee from the International Towing Tank Conference (ITTC) received as one of their terms of reference the study of the effect of data reduction techniques on the outcome of captive model tests for manoeuvring purposes. During such tests, the forces are measured by strain gauges and an appropriate magnification is applied to obtain the time series of a certain force component. The test results are commonly affected by high frequent noise, which in most towing tanks is eliminated by applying low pass filters. The question arises then if useful information is being filtered as well and how this propagates towards the prediction of the manoeuvring derivatives, used for instance for manoeuvring simulation purposes.
The 28th ITTC conference recommended the methodology as proposed by Brouwer et al. (2015) to determine the uncertainty of a mean value from a single time series measurement (ITTC, 2017). In the present paper, the methodology is extended towards the prediction of the uncertainty of the harmonic components, found through Fourier analysis, on the time series of captive planar motion mechanism (PMM) tests, namely harmonic sway and harmonic yaw tests. A preliminary study was also presented by Woodward (2013).
The methodology is applied on the time series of captive harmonic tests, executed by Flanders Hydraulics Research with the benchmark ship KCS (scale factor 1:52.667) in the frame of SIMMAN 2014. A second research question is how filtering, applied either in pre- or in postprocessing, affects these uncertainties. To this aim tests have been carried out at Akishima Laboratories, with and without the use of analogue filtering. The present paper proposes a methodology to assess these uncertainties.