Signals of pressure fluctuation and local solids holdup were measured from different axial locations in a 70 mm x 70 mm, 4.424 m-high circulating fluidised bed with sand particles and white carbon black particles. The original signals were decomposed into multiscale signals based on multi-resolution analysis of wavelet transformation. A new fractal method was used to classify the different scales into three categories through the similarity of the trend on the energy ratio in different scales and the veracity of this method have been discussed. Through this method, the microscales, mesoscales and macroscales were identified and the frequencies of different categories were obtained, which was 15.6-500 hz, 0.5 (or 1) -15.6 hz, below 0.5 (or 1) hz respectively, the characteristics of different categories were further discussed and signals of pressure fluctuation and local solids holdup were analysed to comprehensively understand fluidised bed dynamics. Meanwhile, we research the relationship between statistical analysis and multiscale analysis in a circulating fluidised bed. As for pressure fluctuation signals, the average value is mainly affected by individual particles and clusters, the standard deviation is mainly affected by clusters or the background fluctuation of the entire system, which is dependent on the characteristic of fluidised particles. As for local solids holdup signals, the average value is mainly affected by clusters, the standard deviation is greatly affected by clusters at a relatively low gas velocity and influenced significantly by the individual particles when the gas velocity is relatively high. (C) 2018 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.