电镦机中加热电流的综合集成网络预报模型的算法研究  

ALGORITHMIC RESEARCH ABOUT SYNTHETIC NEURAL NETWORK PREDICTION-MODEL OF THE HEATING CURRENT IN THE ELECTRIC UPSETTING

作　　者：田瑞利[1] 张忠夫[2] 

机构地区：[1]广州航海高等专科学校航海系,广东广州510725 [2]广东工业大学信息学院,广东广州510643

出　　处：《计算机应用与软件》2007年第5期58-59,62,共3页Computer Applications and Software

基　　金：广东省自然科学基金资助项目(990141)。

摘　　要：结合电镦机中加热电流的数学模型,详细分析了综合集成神经网络算法的实现过程,其仿真结果表明,它在“小样本”时,不仅能减少连接权值,而且能加快训练速度,提高泛化能力,实现了电镦机中的关键参数——加热电流的高精度预报。Based on the mathematic model of heating current in the electric upsetting, the paper analyses the algorithmic implementation of synthetic neural network in detail, the results of the simulink could be shown that a synthetic method of neural network can reduce the number of network weight, improve the learning speed, generalization performance of the network in the lack of learning sample, as well as realize a high precision prediction of the heating current ,which is the key parameter in the electric upsetting.

关 键 词：算法研究 综合集成神经网络 加热电流 预报 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]