Multisensor information integration for online wear condition monitoring of diesel engines

A diesel engine bench test was performed, and the online visual ferrograph (OLVF) and performance monitoring sensors were used to evaluate engine wear. The sliding window method was used to segment OLVF-monitoring data; features such as probability of smaller value and accumulated wear coefficient were extracted to clarify wear degree. The weighted combination multisensor information integration method was developed to calculate current engine condition factors. The results show that OLVF monitoring exhibits more sensitivity than other performance monitoring sensors. Using multisensor information provides an early warning of performance degradation ~40 h before the diesel engine experiences a catastrophic fault.