Study on the recursion neural network method with the deviation unit applied to fault diagnosis of armored vehicle gearbox

On the basis of introducing the BP neural network and its shortage, the paper put forward the recursion neural network method with deviation unit and applied it to the fault diagnosis of armored vehicle gearbox. By setting the fault to gear which is the basic parts of the gearbox, the validity of arithmetic applied to fault diagnosis of the armored vehicle gearbox is studied on the improved arithmetic of the BP neural network and the recursion neural network method with the deviation unit. The research enunciates the recursion neural network method with the deviation unit compares BP neural network with quick training speed, high accuracy, save time, accurate fault fix position.