Rapid monitoring of selected food properties using microwave dielectric spectra

In this paper, a novel microwave measurement approach is presented, which is capable of determining selected food properties from the dielectric spectrum. Such properties could be the water content, a possible amount of additionally added water, the content of conducting ions or the percentage of other constituents, having a noticeable influence on the effective complex permittivity versus frequency. The approach is based on measured dielectric data across a significant bandwidth of one or two octaves (or more), and on data analysis using the multi-variate statistical method of Principal Component Analysis (PCA) and Principal Component Regression (PCR). The prototype of a dedicated microwave instrument is described and as an experimental example the viability of the new approach is demonstrated by the exact determination of the percentage of deliberately added water in fruit juices.