The hot deformation behavior of TB17 titanium alloy was investigated with Gleeble-3500 simulator at strain rates of 0.001～10 s-1 and 800～1 000℃

obtaining the flow stress data at varied temperatures and varied strain rates as well as different true strain.Microstructure change can be observed through metallographic observation.Based on experimental data and neural network knowledge

an artificial neural network with back propagation algorithm was established

and knowledge based constitutive relations model was developed after training.The flow stress in the alloys was predicted by the BP neural network.The results are well in agreement with the experimental ones

confirming the availability of the model.