Parameter estimation and multivariable model building for the non-destructive, on-line determination of eggshell strength

In the non-destructive quality assessment of agro-products using vibration analysis, the resonant frequency and the damping of the vibration are the main interest. Those parameters are usually calculated starting from the frequency spectrum, obtained after a fast Fourier transformation (FFT) of the time signal. However, this method faces several drawbacks when applied to short-time signals, as in the case of impact testing of highly damped specimen. An alternative to the FFT method is used for the high-resolution estimation of both resonant frequency and damping. Furthermore, the mass–spring model that is used in the literature for non-destructive quality assessment of various agro-products is extended with the incorporation of the damping and a shape characteristic. As a practical example, eggshell stiffness was estimated using vibration measurements. A data set consisting of 229 eggs was measured. It is shown that both the damping and the shape characteristics are of major importance to explain eggshell strength. This paper makes clear that a univariable model, as is mostly used in the literature, is not always satisfactory to describe the vibration behaviour of biological products.