In order to improve the accuracy of the gear fault diagnosis

a fusion of ant colony algorithm and BP neural network approach is introduced.According to the fault feature vectors of the gear

the fault diagnosis model based on neural network is established.The network weights and threshold as the independent variables

through the ant colony algorithm iteration

the error of the global minimum is searched

then secondary learning and training of the network is carried out

and ultimately

the gear fault diagnosis is realized.Example simulation results show that the method has high accuracy of fault diagnosis

and diagnostic uncertainty is reduced.