Research On Infrared Intelligent Recognition Technology For Virtual Welding Defects Of Spacecraft Circuit Boards

Reliable circuit boards are the basis for the stable operation of electronic products,and solder joints are one of the most important components of them.False welding is a common defect in the production process of circuit boards,which will bring major hidden dangers to the debugging,use and maintenance of the circuit.Therefore,in solder joint inspection,it is important to judge the false welding defect in solder joints.For some solder joints with normal appearance and electrical connection,the existing detection technology cannot effectively detect false welding.In order to solve this problem,this paper applies pulse infrared detection technology to the false welding detection,and then researches and analyzes the identification of false welding according to this technology.The main research contents of the paper are as follows:Firstly,the false welding is introduced,and the one-dimensional heat conduction model of the solder joint is established based on the principle of heat conduction.The principles of thermal radiation and thermal imaging are introduced.The principle of false welding detection technology is analyzed,which provides a theoretical basis for subsequent experimental research.Secondly,the experimental equipment for pulse infrared detection of false welding is introduced.According to the formation mechanism of the false welding,the test pieces are prepared for the experiment.There are three levels of false soldering: 20%,40%,and 60%.The temperature change data is collected.The pulse phase method,the infrared signal trend method and the apparent effusivity coefficient method were used to extract the characteristic parameters that can reflect the false welding defect.Among them,the parameters collected by the apparent effusivity coefficient method have no obvious law,and the parameters extracted by the pulse phase method and the infrared signal trend method can reflect the false welding defects and the different degrees of false welding.These parameters are used to judge the defects and the degree of false welding.Finally,the BP neural network is used to establish a recognition algorithm,which is used to judge the false welding.The BP neural network uses a 4-8-1 topology,with the parameters as input and the degree of false welding as output.Through the training of experimental data,the recognition algorithm can automatically determine the degree of false welding of the test samples.It can be concluded that BP neural network technology can be applied to the detection and identification of false welding.