Research On High-g Shock Acceleration Data Processing Based On Integrated Empirical Mode Decomposition

High g acceleration is one of the main characteristic parameters in strong impact test.The analysis and processing of the measured impact acceleration signal is related to the extraction of the key motion characteristics of the object,the design and optimization of the structure and so on.The traditional method is lack of self-adaptability in analyzing and processing impact acceleration signal.And under different working conditions,the process is repeated manual operation and many subjective factors are introduced.In order to overcome these shortcomings,a combined method of ensemble empirical mode decomposition and continuous mean square error is proposed to process the measured acceleration signals.The ensemble empirical mode decomposition method is based on the local characteristics of signals,and the intrinsic mode functions are obtained by empirical mode decomposition.It is suitable for nonlinear and non-stationary signal analysis with frequency varying with time.The basic principle of the method is introduced,the problems existing in the method are discussed and the corresponding improvement measures and parameters selection basis are given.On this basis,the implementation of the algoritlhm is given.Based on a dropping impact device,five groups of impact acceleration signals with different low pass cut-off frequencies of amplifier are tested.The frequency setting of obtaining rigid body acceleration signal effectively are determined by the time and frequency analysis.Based on a dropping impact device,three groups of impact acceleration signals with different transition elements of sensor are tested.The influence of transition elements on the time and frequency characteristics of test signals is analyzed.The experiments of steel projectile penetrating different layers of reinforced concrete target plate are introduced,and the components and time-frequency characteristics of typical single and double layers penetration signals are analyzed.Several groups of measured dropping impact and penetration signals are decomposed by ensemble empirical mode decomposition method.Combined with continuous mean square error theory,effective signal in components are extracted.The reconstructed signal and its integration results show that the method described in this paper can extract effective components of any measured high g impact acceleration signal adaptively.And compared with the result obtained by traditional filtering method,the result obtained by the method described in this paper is more precise.Since the entire processing is driven by the signal,without the help of other analysis tools to repeating operation artificially,and the processing results can well meet the requirements of research,it is treated as a more efficient and more universal processing method in dealing with high g impact acceleration signal.