Abstract
The process of estimation of manoeuvring derivatives from oscillatory captive model tests is analyzed and it is demonstrated that, contrary to common belief, all important information concerning quasi-steady force components can be extracted by means of only two terms of the Fourier expansions of the output time histories of force measurements: the mean values and the first-order out-of-phase coefficient values. Additional involvement of the second-order coefficients does not alter substantially the estimates of the regression coefficients and in most cases does not make the resulting regression model more certain and accurate, as shown with an example.