In order to solve the problems in the turbo blade investment casting die design process，such as long design time，lower utilization of expert experience and lower level of intelligence，the knowledge engineering is introduced in the turbo blade investment casting die design，and an intelligent knowledge—based investment casting die design system is developed and the key technologies of the system are researched．The practical application results indicate that this system can reuse the expert experience efficiently and improve the efficiency and quality of die design．
Fault detection and diagnosis methods for non—linear system based on wavelet neural networks are researched．Wavelet analysis and feed-forward neural networks are fused，and the algorithm is inferred．Method of wavelet neural networks is used in fault diagnosis for the lubricant system of the oA!roen-sine．Results of simulation and tests show that this method，which is good to fault diagnosis for non-linear systems，does better in convergence speed，and is not sensitive to the inputs of the network．
The properties of solid film hbri·cant including resistance of wear，high temperature，temperature alternating and oil are researched and compared with the friction wear properties of silver coating．Experiments indicate that the graphite solid film lubricant has less wear value and lower friction CO．efficient，has good properties of resistance of wear，high temperature，temperature alternating and oil，and has loss weight properties．Therefore，the lubricant possess-es excellent lubricating and reducing friction properties．
The near—isothermal forging exo periment of GH4169 alloy is carried out．The results show that the microstructure and tensile properties have hish sensitization to the temperature pattern．but low sensitization to the deformation and strain rate，which is benefit to the near—isothermal forming of GH4169 alloy．
