Identification of Multiple Impacts on a Plate Using the Time-Frequency Analysis and the Kalman Filter

Abstract

This article presents the method of identifying the locations of the multiple impacts on a plate where the impacts occur sequentially. The main aim of this research is to provide the basis on the human-interaction technology using variety of methods used in the area of signal processing, piezoelectric materials, and wave propagation. The work concerning the location identification of a single impact on a plate has been previously studied by means of the time-frequency analysis applied to accelerometer signals. In this article, a novel approach for the location identification of the discrete multiple impacts is presented to investigate the feasibility of applying to a more practical human-interaction system. For the identification of the series of impact locations, the major axisymmetry mode of the Lamb wave on the plate is considered and the CWT is applied to obtain the arrival time difference of the Lamb waves between sensors. Then, the Kalman filtering technique is employed to continuously track the locations of the impact load and to improve the estimation results. The results may well be applied to the real-time health monitoring of the steam generator in a nuclear power plant.