Prediction of Cavitation Erosion in Hydraulic Equipment Using Externally Mounted Acceleration Sensor
Please login to view abstract download link
Construction machinery employs various types of hydraulic equipment for transmitting power by fluid. Cavitation occurs in hydraulic equipment and causes noise, vibration and erosion. Especially, cavitation erosion is a serious problem because it shortens product lifetime. In order to keep construction machinery in continuous operation, it is desirable to replace hydraulic equipment before its performance deteriorates by cavitation erosion. However, since it is impossible to directly observe cavitation erosion inside hydraulic equipment, alternative methods of predicting cavitation erosion are desired. In the present study, the feasibility of the cavitation erosion prediction by using machine learning was investigated. Vibrations caused by the collapse of cavitation bubbles in a cavitation jet erosion test equipment were monitored using an externally mounted acceleration sensor, and a method to predict material damage using machine learning was investigated based on the accumulated sensor data. Acceleration sensor data and specimen surface roughness Rz for various flow conditions were used for the training and test data. As a result, predicted surface roughness was obtained by combining unsupervised and supervised learning and agreed quantitatively with measured surface roughness.