基于冲击响应信号的大规模蛋壳裂纹检测仿真

Large-Scale Eggshell Crack Detection Simulation Based on Impulse Response Signal

针对当前蛋壳裂纹检测方法存在查准性和查全性差的问题,提出基于冲击响应信号的大规模蛋壳裂纹检测方法。利用激振器、鸡蛋、底座、加速度传感设备、信号采集端等构成蛋壳冲击响应信号采集装置。将得到的非稳态信号经EMD分解成多个固有的模态函数,并对信号主要分量实行Hilbert Huang变换,得到时频上的能量分布图谱。将其代入主成分分析法中,计算信号特征贡献率。由此得到蛋壳各层信号奇异性指标,并给出相应层次的奇异性指标标准值。对各层信号奇异性指标和标准值进行对比,如果奇异性指标大于标准值,则为裂纹蛋壳。反之,则为好蛋壳。通过逐层对比,实现大规模蛋壳裂纹全面检测。实验结果表明,上述方法查准率和查全率均较高,具有合理性。

Current detection method of shell crack has many defects such as poor precision and recall performance.In order to overcome the defects,based on impact response signal,a new detection method of large-scale shell crack is proposed in this article.Acquisition devise of impact response signal of shell was constituted of vibration exciter,egg,pedestal,sensing equipment of accelerated speed,and signal acquisition side.The obtained unsteady state signal was decomposed into several intrinsic mode functions via EMD and Hilbert Huang transformation was carried out for major component of signal.Energy distribution spectrum in temporal frequency was obtained.The spectrum was substituted into principal constituent analysis and rate of contribution of signal feature was calculated.Thus,signal singularity index of each layer of shell was obtained and standard value of the singularity index of corresponding layer was provided.The singularity index was compared with the standard value.If the singularity index was larger than the standard value,the shell has crack.Otherwise,the shell was unbroken.Comprehensive detection for the large-scale shell crack was achieved via layer-by-layer comparison.Simulation results show that the method has higher precision and recall performance.It has rationality.