The Design And Study On Detection System For Flaw Of Profiled Component

The profiled component has complex structural features, it has many discontinuity interfaces such as edges and steps. When it is tested with ultrasonic instrument, the edges and steps of objects studied in this paper will reflect strong disturbing signals. It is hard to distinguish flaw signal from the noise by manual detection, so in order to improve product quality and use safely, it is very necessary to design an automatic online detection system combined with signal processing technology for profiled component.According to the structural characteristics and defect types of profiled component, a method of multi-probes synthetic judgment which based on principle of ultrasonic creeping wave and bottom echo depth is put forward to test flaws in profiled component. 8-Channel Digital ultrasonic signal processing system has been designed and front-end processing circuits of transmitting and receiving ultrasonic signals are realized,which mainly include amplification circuit, detection circuit, gate selecting circuit and peak value holding circuit. 8-Channel Data Acquisition and Control System based on ISA Bus is designed,which can set some parameters such as gate and gain according to the order of upper computer and can acquire peak voltage of echoes from the detector and store in RAM. The upper computer can control the acquisition system working or read data by ISA Bus. WinDriver is used to develop windows driver of acquisition system. Visual C++ is used to develop control software and comprehensive defection assessment platform of profiled component. It is achieved to communicate with PLC mechanical system, set parameters of detection system, display the testing results and save detection report.Through comparing interference signals( angularity wave and step wave) with flaw signals, Attenuation compensation method based on energy of signal is put forward to extract flaw signals from the ultrasonic echoes effectively. Neighbor thresholds tracking method is used to remove trend line of ultrasonic echo signal that caused by work piece assembling, and combined with threshold method flaw signals are extracted from echo of the vertical probe. Based on the data fusion theory, the flaw type is distinguished synthesizing detection results of three detection channels. 9 work pieces with artificial or natural flaws are tested. The test results verify that above methods can effectively apply in the ultrasonic automatic detection system.