Evaluation of roundness error using norm theory

In order to evaluate the roundness error of a measurand, the eccentricity must be separated from the measuring signal, and the methods used include least-squares method, Kalman filtering approach and three-probe method. These methods can not be used not only to analyze the non-integral order harmonics, but to distinguish between the eccentricity and the first harmonic, and contain a lager nonlinear-error. To solve these problems, a novel method using norm theory is proposed to characterize the roundness error. All kinds of disturbing errors can be efficiently deadened by the use of the method, the computing result certainly converges to the true value of the roundness prior, and the error of parameter estimation is very small.