摘要
针对聚丙烯熔融指数软测量中预测精度不高的缺点,将基于深度置信网络-极限学习机(DBN-ELM)的软测量方法应用到熔融指数的软测量中。与传统深度置信网络(DBN)不同的是,该方法将极限学习机(ELM)算法运用到深度置信网络的训练中。首先用深度置信网络对原始数据进行数值分析来提取特征,然后将提取的特征输入到极限学习机中进行训练,得到软测量模型。实验验证表明,与支持向量机和单纯的深度置信网络模型相比,该方法具有更高的测量精度。
To solve the issue of low accuracy of the traditional soft sensor methods of polypropylene melt index, an approach based on deep belief network and extreme learning machine(DBN-ELM)was used to the melt index prediction of polypropylene. Traditional deep belief network(DBN) applied the deep learning to the learning process of the deep neural networks. Different from traditional deep belief network, this approach applied the extreme learning machine algorithm(ELM) to the learning process of DBN to improve the DBN model. Firstly, deep belief network was employed to extract effective features from vibration data by numerical analysis. Then, the effective features were put into the extreme learning machine to proceed model training to obtain the soft sensor model. The experimental validation showed that the method was more accuracy than the traditional method.
出处
《化工学报》
EI
CAS
CSCD
北大核心
2016年第12期5163-5168,共6页
CIESC Journal
基金
山东省自然科学基金项目(2013ZRE28089)~~
关键词
深度置信网络
算法
极限学习机
数值分析
特征提取
实验验证
deep belief network
algorithm
extreme learning machine
numerical analysis
feature extraction
experimental validation