Study of a proportional spool valve noise by means of Functional Data Analysis: Cavitation and intensity detection

Cavitation inside hydraulic components is a well-known issue, that arises under specific operating conditions. When this phenomenon occurs in valves, it leads to several issues, such as materials damage, performance reduction, and noise. The latter of them is becoming increasingly important in the transport sector, which is moving towards new technologies (i.e., electrification) that are notoriously quiet. Therefore, cavitation detection is becoming a matter of interest for several applications. The purpose of the work presented is to develop a diagnostic technique to detect the onset of cavitation in proportional valves. A 2-way 2-position proportional spool valve has been placed inside an acoustic enclosure; noise signals have been recorded by means of a high frequency microphone, for different flow rates and valve openings. Sound pressure levels have been analysed by means of statistical techniques; a Functional Data Analysis (FDA) has been performed in the open-source R environment, in which the experimental data, collected under different non-cavitation conditions, have been firstly used to perform a Functional Principal Component Analysis (FPCA) and then to define a threshold. Then, a further experimental dataset has been evaluated through a Hotelling control chart, from which it is possible to distinguish cavitating conditions as they fall outside the limits.