机构地区: 哈尔滨理工大学机械动力工程学院
出 处: 《机械设计》 2006年第2期33-35,共3页
摘 要: 温度场预测是实现铣刀片槽型设计与重构的关键技术,神经网络预测模型是实现温度场预测的新途径。针对铣刀片温度场的非稳态特性,提出了一种基于BP神经网络Levenberg-Marquardt算法的三维复杂槽型铣刀片温度场预测模型,避免了传统神经网络易陷入局部极小的缺点。预测结果表明,该模型收敛速度快,预测精度高。 Prediction of temperature field is the key technique for realizing the design and reconstruction of grove pattern of milling insert, neural network prediction model is a new way for realizing the temperature field prediction. Aimed at the non stable characteristics of temperature field of milling insert a kind of temperature field prediction model for 3D complex grove patterned milling insert was put forward based on Levenberg-Marquardt algorithm of BP neural network, and thus avoided shortcomings of easy to fall into partial extreme minimum in the tra ditional neural network. The result of prediction shows that this model has a speedy convergence and a prediction of high precision.