Research on Filtering Algorithms for Signals Collected in Explosion Experiments

The experimental signals collected in explosion experiments are always mixed with different degrees of noise interference. In order to accurately analyze the variation laws conveyed by these signals, four sets of explosion experiments were designed with different charge amounts and vacuum environments in the vacuum explosion vessel. Then, the collected impact load data were analyzed by applying both Fourier filtering algorithm and median-averaged filtering algorithm. By comparing the P-t curves processed by the two filtering algorithms with the original ones, it is found that the Fourier filtering algorithm is a global analysis of the signal, which can extract the frequency information of the function in the whole frequency domain, while the characteristics of the signal cannot be revealed in a local time range. Although the processing speed is faster, the error for the characteristic parameters is larger, and the effect of the filtering process directly applied to test signals of the blast impact is less satisfactory. The fit degree between the signals of explosion impact processed by the median-averaged filtering algorithm and the original ones is higher, and the varying details of the impact load with time in the blast container can be clearly reflected with a smaller error and a higher reliability.