Classification of operating conditions of machinery combined with transmissibility function method

Condition monitoring of machinery is concerned for the purpose of maintenance, productivity and avoiding potential downtime. In the trend of smart factory, the virtual representation of machining process can be built by measuring the vibration. Different working conditions and fault conditions should be identified. Different working conditions have different characteristics with the general signal processing parameters in time and frequency domain. The dimension reduction method like PCA, and the cluster method like k-means can be used for further classification. In order to accurately classify different operating conditions, the feature extracted from the measured signal should be distinct enough. In this case, multiple vibration sensors have to be implemented, and transmissibility function between different sensors of the dynamic system can be considered as the characteristic parameters. This paper combines transmissibility function and signal processing parameters for classification of different operating conditions to extract more information about operating process.