Fault diagnosis of marine diesel engine based on blind source separation

This article makes use of marine diesel engine local vibration signals. These signals have properties of non-stationary, non-Gaussian and low signal-to-noise. These unknown sources about diesel vibration signals can be estimated and recurrent according to blind source separation (bss) algorithm and combining other time domain, frequency domain analysis method. The independent signal source can also be identified through the linear mixture of the mutually independent signal source. It is also necessary to use some guideline and optimization algorithm. The diesel engine can be fault diagnosed with accuracy through these separated data. The real-time on-line diagnosis instrument about diesel have be developed by comprehensively adopting embedded technology, advanced sensor technology, signal processing and communication technology.