摘要
采用改进的BP网络Levenberg-Marquardt优化算法对冷连轧机出口厚度进行快速预报,此网络μ参量可自适应调整,收敛速度快。冷连轧生产出口厚度预报精度大为提高,为冷连轧出口厚度预报提供了一条准确高效的新途径。
Improved BP neural network which adopted Lev-enberg-Marquardt optimized algorithm has been used to predict thickness. In this algorithm, the parameter μ can be adaptive adjusted, and network convergence speed is higher than before. Thickness predicting precision was improved. A new precise and high efficiency way for cold continuous rolling exit thickness prediction have been provided.
出处
《燕山大学学报》
CAS
2003年第1期51-53,57,共4页
Journal of Yanshan University
基金
国家自然基金资助重点项目(项目编号:50035010)
教育部科技重点资助项目(项目编号:00144)。
关键词
冷连轧机
出口厚度
人工神经网络
预报
cold continuous rolling process, exit thickness, neural network, predict.
