摘要
本文提出一种基于生产实践样本数据、BP神经训练和嫁接网络对锌窑渣、水解渣共同熔炼进行预测的模型。10组数据所构建的9-16-13-8BP神经网络训练模型,是以锌窑渣、水解渣中各成分作为输入参数,冰铜、烟尘和水淬渣相应成分作为预测目标。建立输入参数和预测目标之间的模拟关系,训练误差为10-4。预测网络是采用训练网络权值、阈值,和相同设计参数,在Matlab的GUI可视化界面中输入各成分的含量,实现各元素成分的预测,误差为10-4,人机交互性强。
A kind of prediction model was studied in this paper The model was based on zinc-kiln slag,Goethite Residue of common smelting production data.The inputs of the training BP Neural Network were zinc-kiln slag,Goethite Residue parameters of each component.The outputs are ingredients of matte,zinc oxide powder and pulverized slag.Using 10 groups sample data constructed a 9-16-13-8BP neural network training model.This model can establish the relationship between input parameters and predicts.The training model showed that the training error was 10-4.Forecasting network was based on weights and threshold of the training network.It can input zinc-kiln slag and goethite residue parameters of each component to forecast constitutes of matte,dust and pulverized slag in GUI interface.Using Matlab GUI visual interface can realize friendly man-machine interactive operation.
出处
《四川冶金》
CAS
2010年第6期28-34,42,共8页
Sichuan Metallurgy
关键词
锌窑渣
水解渣
共同熔炼
BP网络
预测系统
Zinc-kiln slag
Goethite residue
Common Smelting
BP Neural Network
Forecast system
作者简介
李永祥，男，工程师，从事有色生产冶炼与管理等工作。
