The Data-Driven Multivariate Process Monitoring and Diagnosis of Rides in an Amusement Park

摘要

To guarantee the safety of rides is vitally important and meaningful to the amusement park industry. A multi-scale principal component analysis (MSPCA) is engaged in the fault detection of a ride for long term, where wavelets capture correlation within a sensor and principal component analysis correlates across sensors. The data measured from a ride in an amusement park are given and fed into the MSPCA model to diagnose whether a fault would occur. For the non-stationary process of a ride on run, it is hard to clarify how healthy the ride is according to the data on once run. A strategy of long-term monitoring and diagnosis is given.