Research on vibration source separation of the Stirling engine based on the Fast ICA algorithm
Bohdan Borowik, Bohdan Borowik, Igor Piotr Kurytnik, Barbara Borowik, Barbara Borowik · Vibroengineering PROCEDIA · 2014
In this paper we employed the Fast ICA Independent Component Analysis algorithm for decomposing vibration data in operational Stirling engine. Accurate separating vibration signals containing noises allows for the early fault detection of the Stirling engine and by this guarantees the more reliable work of this engine and its better performance. The results show that using the Fast ICA blind source separation algorithm enhances the quality of obtained fault detection information, which is more precise and makes the fault diagnosis more effective.