锻造变形均匀性的支持向量机模型及应用

Aiming at the problem of non-uniformity in the forging process of high-strength titanium alloy bars with ultra large size, a support vector machine model of forging deformation uniformity of titanium alloy bar was established by using machine learning method based on the physical thermal compression simulation tests and numerical simulation experiments. Combined with the normalization of forging process parameters, the optimization model of forging deformation uniformity was obtained. The evaluation function of the distribution uniformity for actual forging temperature and strain and the multi-objective optimization model of forging process parameters were presented. The forging process parameters based on the distribution uniformity of actual forging temperature and strain were obtained by using optimization algorithms. The above mentioned models were applied to the forging process of titanium alloy bars. Taking the forging temperature, forging speed and reduction amount as optimization variables, and the distribution uniformity of actual forging temperature and strain as optimization objectives, the combination of the seven-pass forging process parameters of the 1300 MPa titanium alloy bars with the diameter of Φ400 mm was optimized.