Research On Pulse Compression Technology For Acoustic Level Detection

The material level detection technology is becoming more and more important in the automated production process of modern industry.The commonly used material level detection methods include laser,ultrasonic,radar,etc.However,the detection effect is not ideal in the environment of high dust in large silos,and most of them can only realize single point detection.At present,there are foreign products to realize three-dimensional distribution detection of material level based on acoustic detection method.The penetration of high concentration dust is realized by using the characteristics of large wavelength of acoustic signal,but the detection precision and accuracy of three-dimensional distribution of material level still need to be improved.This thesis focuses on the acoustic level detection signal processing method in the high dust environment of large silos.The main research contents include:(1)In this thesis,after analyzing and comparing the influence of bandpass filtering,wavelet transform,median filtering and other processing methods on the accuracy of position detection,based on the frequency characteristics of loudspeaker,an improved nonlinear frequency modulation pulse compression method is proposed.Firstly,nonlinear frequency modulation signal is used for pulse compression processing,and the pulse compression signal is obtained,and the performance of the signal is improved.Then the window function of the signal is improved,and the indicators of the signal are further improved.Finally,the simulation results show that the method in this thesis has good performance in both signal-to-noise ratio and lateral resolution.(2)In order to overcome the influence of the sidelobe on the detection accuracy after pulse compression,the windowed weighted sidelobe suppression technique is introduced to suppress the sidelobe,and the windowed weighted method is improved.Firstly,an ideal pulse compression result is obtained,and then the target window function weighted filter frequency response is obtained by inverse inference.It is also observed that most common window functions are bell normal distribution.The improved window function is obtained,and the simulation results show that the proposed method can improve the suppression ability of sidelobe.(3)Based on the quadrilateral microphone formation,an improved controllable response power Phase transform(SRP-Phat)sound source location algorithm is proposed in this thesis.Firstly,the method of random area reduction is introduced to improve the efficiency of space grid point search.Then,a new search strategy from rough to fine is adopted to optimize the performance of the algorithm.Finally,the simulation results show that the improved algorithm has higher positioning accuracy.