Research Implementation of Crack Detection Method Based on Its Acoustic Property
|Course||Detection Technology and Automation|
|Keywords||protection tube crack detection acoustic signal processing wavelet analysis parameter identification|
The cracks in the protection tubes seriously impact on the qualified rate of temperature sensors. Removing the protection tubes with cracks, could improve the qualified rate and reduce the production cost. The purpose of this thesis is to realize the recognition of protection tubes with cracks.First, the principles of acoustic detection were analyzed. Through analyzing the principle of cracks detection, acquired that the frequency and mode shape of acoustic signal, which were produced by knocking on the protection tube, would change with the cracks. Thus, whether the cracks are existed in protection tubes can be judged by the changes of related characteristics in frequency and mode shape.Second, detection system was designed and acoustic signals was conducted a pre-processing. The sound incentive and acquisition devices composed the detection system. This paper described the choice and design of incentive devices, acquisition sensors, signal conditioning circuits, data acquisition card and parameters. Through large number of experiments and data analysis, we determined the excitation signal devices, percussion strength, percussion location and so on. Finally determined the optimum experimental program by doing some pre-processing and initial studied the characteristic parameters of signal.Third, the acoustic signal was analyzed by using the traditional signal processing methods and the character parameters were extracted. Acoustic signal was analyzed in time domain and frequency domain. The character parameters like signal strength, signal power, power spectrum peak, power spectral cancroids in X axis and Y axis, and so on, were extracted. The parameters which can distinguish the cracked or crack-free protection tubes were found. Then, the protection tubes were classified by using them.Last, wavelet analysis method was introduced to distinguish the crack in cracked protection tubes. By processing the energy of sound signal in wavelet packet nodes, the rule of energy in different frequency band changes with location was found and summarized. Two indicators of ratio in wavelet packet energy square and ratio in wavelet packet were proposed and used to distinguish the crack.By using these two methods, whether the cracks were existed in the protection tubes was successfully judged. The recognition rate of crack in protection tubes reached the production requirements.