Predicting the Electrical Impedance of Rolling Bearings Using Machine Learning Methods

The present paper describes a measurement setup and related prediction of the electrical impedance rolling bearings using machine learning algorithms. bearing is expected to be key in determining state health bearing, which an essential component almost all machines. In previous publications, determination has already been advanced analytical methods. Despite improvements accuracy achieved within calculations, there are still discrepancies between calculated measured impedance, leading approximately constant off-set value. This discrepancy motivates approach introduced this paper. It shown that with help data-driven methods difference reduced order up 2% across operational range analyzed so far. To introduce context research shown, first underlying physics presented. Subsequently different approaches highlighted compared each other terms their quality results part As further aspect, addition it investigated whether rotational speed at can predicted from frequency spectrum analysis independent force accuracy. background that, if sufficiently high, additional use sensors could omitted future investigations.