Recognition of the flow regimes in the spouted bed based on fuzzy c-means clustering

Hilbert-Huang transformation has been applied to extract eigenvectors from the pressure fluctuation signals in the spouted bed. According on these eigenvectors, the flow regimes in the spouted bed could be classified into 4 clusters including ‘packed bed’, ‘stable spouting’, ‘bubbling fluidized bed’ and ‘slugging bed’ by chaos optimized fuzzy c-means clustering algorithm. The Elman neural network was used to recognize these four flow regimes, and the parameters in the Elman neural network were optimized by adaptive fuzzy particle swarm optimization algorithm. The recognition accuracies of ‘packed bed’, ‘stable spouting’, ‘bubbling fluidized bed’ and ‘slugging bed’ can reach 85%, 90%, 85% and 80% respectively.

Graphical AbstractHilbert-Huang transformation was used to extract the eigenvectors from the pressure fluctuation signals in the spouted bed. Based on the eigenvectors, chaos optimized Fuzzy c-means algorithm was used to search for the optimum clustering number of the pressure fluctuation signals, and the Elman neural network has successfully recognized the flow regimes.Figure optionsDownload as PowerPoint slide