Skew Rolling Modeling Based On Fuzzy Neural Network Research

Cross-rolling is an important production technology in the production of seamless pipe. It has become a basic technology and a key means for all kinds of seamless pipe production after 100 years'development. But it is often difficulty for people to build the exact math model by using classic mechanics approach. By using FEM, the calculating precision in cross-rolling process is higher than using other methods. But the simulation time is too long to be on-line control mathematics models. It is enforced to carry on the development and research on mathematics models of artificial neural net.This project is a study of cross-rolling process to pipe-elongating operation in the production of seamless pipe. The mapping relationship between technical parameters and quality parameters is obtained by calculating parameters in cross-rolling process, such as deformation state, mechanical state, pipe quality ect. And the fuzzy neural network and membership function of input parameters in cross-rolling process are established. Training data are gained through FEM modeling and experimental data. Therefore, the main work includes the following aspects:(1)Based on a deep knowledge of production technology of cross-rolling seamless pipe and cross-rolling theory, sample data for the need of fuzzy neural network are collected by ANSYS/LS-DYNA for modeling and Simulation.(2)Establish fuzzy neural network about cross-rolling process of seamless pipe; select membership function; and divide sample date into two groups: training date and testing date.(3)Train fuzzy neural network by using experimental data and FEM simulation data.(4)Build an experimental platform and verify the veracity and reliability of finite-element modeling of cross-rolling through studying the principle of cross-rolling by using aluminum pipe instead of steel pipe.Based on the modeling research on fuzzy neural network of cross-rolling, the relational data between technical parameters and quality parameters are established, which has directive significance for cross-rolling technology. The application of fuzzy neural network to production technology will have a very great reference value for the technology research in heavy manufacturing environment.