Application of BP neural network in welding joint data processing

This article gathered the welding current, welding voltage and column displacement moves parameters of the U71Mn rail welding process orthogonal test, and extracted the quality characteristic quantities of luminous time ratio, energy input, weld period and cremation quantity before accelerating preceding to input to the BP neural network forecast model to research the welding process research to the U71Mn rail. The test result indicated that the welding process parameter conformed to the railroad standard stipulation, and played certain instruction role in the project practical application.