Clustering acoustic emission data streams with sequentially appearing clusters using mixture models