An empirical mode decomposition-based signal process method for two-phase debris flow impact

Impact of debris flow consists of two distinctive phases due to its physical composition. One is the dynamic impact from fluid phase, and the other is collision from the solid phase. At present, there is no effective way to differentiate these two phases of impact. An empirical mode decomposition (EMD)-based signal process method was proposed in this paper to extract fluid and solid impact force of debris flow from the mixed signal. Miniaturized flume tests have been carried out with 14 work conditions, and the impact signals were captured by a digital logger. From the experiment, frequencies of fluid phase and solid phase impact signals were identified in the range of 0.05-2 Hz and 300-600 Hz, respectively. The impact signals from solid and liquid phases were reconstructed using the proposed method. In addition, the impact force of fluid phase that measured directly from the flume tests and calculated from isolated signals showed good agreement and the average difference was about 10%. However, large deviation of solid phase impact was observed especially when this method was applied to the full-scale debris flow events and the difference ranged from 26.33 to 61.47%. This proposed method provided an alternative approach to study the debris flow impact force in terms of slurry and large particles separately.