Signal Processing Of Automatic Ultrasonic Thickness Measurement For Cylinder With Complex Inner Surface

Characteristics of rough inner surface, having edge and groove and thickness range ofcylinder with complex internal surface result in large differences of echo signal and being noteasy to extract characteristics of signal and reconstruct the detection results when usingultrasonic automatically detect the thickness of cylinder. Aiming at above problems, thispaper researches feature extraction and thickness reconstruction technology of automaticultrasonic thickness.Through the analysis of echo signal and noise in the process of automatic detection, thispaper adopts median filtering and triangular pulse reconstruction technology to not onlysuppress the noise of echo signal effectively, but also preserve the peak of echo signal; inview of problem that the echo of inner and outer surface is not easy to distinguish whendetecting the thin position of cylinder, using the deconvolution filtering technique extracts thepeak of two echo as characteristic value, to calculate the thickness of cylinder; in view ofsituation that the rough inner surface of cylinder leads to location of echo signal peak orvalley value offset, this paper puts forward correlation pattern matching method thatcalculates the thickness by obtaining the time difference of two characteristic echo throughwaveform matching; through experimental verification, the correlation pattern matching isalso applicable to signal processing of thin wall position; analyzing the advantages anddisadvantages of deconvolution filtering technology and correlation pattern matching, as aresult, deconvolution filtering is better than correlation pattern matching in view of processingeffect, but correlation pattern matching is faster than deconvolution filtering for processingsignal; weak characteristic echo of edge and groove junction of cylinder makes featureinformation not easy to extract, according to the characteristics of the internal structure of cylinder, this paper puts to use the exponential interpolation function to deal with thecorresponding thickness curve.Based on the actual workpiece, this paper designs a set of automatic ultrasonic testingsystem witch above processing methods are applied in, and chooses the correlation patternmatching that has high efficiency and strong adaptability for feature extraction. The detectionresults show that: the research methods can be greatly good for detecting the thickness andthickness reconstruction for cylinder with complex inner surface, besides, detection resultsmeet the actual measurement requirements. The paper researches have reference value tomeasure the thickness of component in the engineering measurement.