Research On Laser Ultrasonic Detecting Method Of Pest-Eroded Wheat Based On Finite Element Simulation

Wheat can be easily eroded by worms to form worm-eaten wheat if it is not stored properly.Which seriously reduces wheat quality and commercial value.Therefore,research on the rapid detection of the quality of stored wheat is very important,which is of great significance for ensuring the quality of wheat and its products and reducing post-harvest grain losses.In this paper,laser ultrasonic technology is applied to the detection of worm-eaten wheat,and the process of laser irradiation of wheat is numerically simulated based on finite element analysis.The time-domain waveform characteristics of ultrasonic signals,the spectral characteristics based on wavelet packet energy entropy and the time-frequency characteristics of signals based on EMD and Hilbert transform are studied.Combined with BP neural network and ELM to identify wheat grains,the recognition accuracy of worm-eaten wheat is 97.39%and 95.65%,and the recognition accuracy of worm-eaten state is 91.59%and 87%,respectively.The detection accuracy rate is significant.The main work of this paper is listed as follows:1.The application of laser ultrasonic technology in the detection of worm-eaten wheat grains is proposed,and the physical process of the laser acting on the wheat grain model is simulated based on the finite element method.The results of numerical calculations show that the waveform of the laser ultrasonic signal carries the information of wheat parameters.This conclusion lays the foundation for the method of laser ultrasonic detection of worm-eaten wheat grains proposed in this paper.2.The statistical characteristics of the time-domain waveforms of laser ultrasonic signals of wheat models were studied.The results show that the crest factor(₁),impulse factor(,form factor(,margin factor(,and the short-term energy characteristic values(1、2 of the signal are better distinguished.And the degree of discrimination increases as the radius r increases.3.Combined with wavelet packet energy distribution and information entropy theory,the time-frequency characteristics of laser ultrasonic signals are explored.The energy entropy value(H that reflects the concentration of energy distribution in the signal frequency band is selected as the characteristic parameter.The results show that the energy distribution of the signal of the node 5 of the wavelet packet decomposition of the worm-eaten wheat grains is relatively concentrated,and the entropy value is small.4.Based on the Hilbert-Huang transform,the time-frequency distribution characteristics of laser ultrasonic signals in wheat are explored.The EMD decomposition results of the signal indicate that the energy content of the first-order IMF component is the largest,and the signal energy and the maximum amplitude of the spectrum of the IMF1 component of normal wheat are larger than those of worm-eaten wheat.The results of Hilbert spectrum analysis showed that the energy of the ultrasonic signal of wheat was mainly concentrated in the low frequency band of 0-50 Hz,and the marginal frequency of the center of gravity and the maximum amplitude of normal wheat were larger than those of worm-eaten wheat.5.Based on the extracted feature parameters,three worm-eaten wheat grain recognition models based on BP neural network and extreme learning machine(ELM were constructed.The recognition results show that the recognition accuracy is the highest when the time and frequency domain features are integrated as the input of the network.The detection result of BP network on normal grains is 100%,the detection result of insect erosion grains is 95.62%,and the overall detection accuracy rate is 97.39%;The detection result of ELM network on normal grains is 100%,the detection result of insect erosion grains is 92.31%,and the overall detection accuracy rate is 95.65%.6.Based on the correct identification of worm-eaten wheat kernels,classification models of worm-eaten state based on BP neural network and ELM were established.The detection accuracy of BP network in the early,middle,and late stages of insect erosion is 90.45%,100%,and 96.50%,and the overall detection accuracy is 91.59%;The ELM network has 66.67%,96%,and 88%detection accuracy in the early,middle,and late stages of insect erosion,respectively,and the overall detection accuracy is 87%.BP network and ELM have significant effect on the identification of worm-eaten wheat grains,which indicates that the laser ultrasonic detection method proposed in this paper can be used for the detection of worm-eaten wheat grains.The above research results provide a new idea for the establishment of a rapid identification model of worm-eaten wheat grains.