Research On Evaluation Of Welding Spot Quality Of Car Body Based On Ultrasonic Echo Characteristics

Resistance spot welding is the most-used connection technology of automobile body in white, so the welding spot quality directly affects the body assembly quality,body mechanics performance and vehicle quality. Because of the advantages of low cost, being simple to operate, being easy to carry, ultrasonic detection method is the most widely used one of all welding spot quality nondestructive testing methods, but the application has been maintained at the level of manual testing. To improve the testing efficiency and accuracy, and to achieve automatic detection of welding spot quality, it is necessary to do further research on the characteristics of ultrasonic testing signal of the welding spots, the signal feature extraction method, and the welding spot quality evaluation database.First of all, this paper did research on the theory of welding spot quality automatic ultrasonic testing, the ultrasonic testing signal processing method and theoretical basis of establishing evaluation database, providing theoretical supports and technical methods for the subsequent research content.Second, the paper did the study of the method for selecting the optimal evaluation characteristic subset. By establishing the parameterized model of ultrasonic testing signal, an ultrasonic echoes parameter estimation method based on EM algorithm was proposed in this paper, and the method has realized the accurate estimation of multi-echo signal parameters. Time and frequency characteristics were extracted based on the signal parameter estimation method and signal energy analysis method by analyzing characteristic of the welding spot ultrasonic testing signal in the time-frequency domain. In order to remove redundancy characteristics and select the most preferred evaluation characteristic subset, the characteristic validity supervision model was established and the weighted coefficient was defined based on J-divergence entropy, finally the optimal evaluation characteristic subset was chosen as welding spot quality assessment basis.Then, the research gathered a large number of samples of welding spot ultrasonic testing signals, by using the ultrasonic signal parameter estimation method and signal energy analysis method, the optimal characteristic subset data were extracted and the characteristic database was established. The primary weldingspot quality evaluation database based on testing data was established by analyzing characteristic database and solving different types of welding spot quality of the corresponding characteristic scope. Aiming at the deficiency of small application range of the primary evaluation database, by using the method of finite element simulation analysis, setting up different quality welding spot models, fitting ultrasonic testing signals to correct and extend the primary evaluation database, a senior welding spot quality evaluation database with wider applicable range and higher reliability was built.Finally, the paper introduced the design and development of the welding spot quality ultrasonic automation testing equipment, and the characteristic extraction method, evaluation database were integrated in the software system of the testing equipment. The actual test of the testing equipment has verified the validity of the method in this paper, the reliability of the evaluation database and the feasibility of the testing equipment used in on-line automatic detection of welding spot quality.The research results indicate that the processing method of ultrasonic testing signal of welding spot, characteristic extraction method and the evaluation database based on ultrasonic echo characteristics can be applied to the detection and evaluation of car body welding spot quality effectively. Testing examples prove that the welding spot quality testing equipment, which is developed based on the method of this paper, has high detection efficiency and the identification of welding spot quality is reliable. This testing equipment can help achieve the comprehensively automatic and intelligent detection of welding spot quality.