Please wait a minute...
浙江大学学报(工学版)
J4  2011, Vol. 45 Issue (12): 2088-2092    DOI: 10.3785/j.issn.1008-973X.2011.12.003
自动化技术     
基于选择性极限学习机集成的磨机负荷软测量
赵立杰1,2, 汤健1, 柴天佑1
1.东北大学 流程工业综合自动化国家重点实验室,辽宁 沈阳 110004;
2. 沈阳化工大学 信息工程学院, 辽宁 沈阳 110142
Soft sensor of mill load based on selective extreme
learning machine ensemble
ZHAO Li-jie1,2,TANG Jian1,CHAI Tian-you1
1. State Key Laboratory of Integrated Automation for Process Industries, Northeastern University, Shenyang 110004, China;
2.College of Information Engineering, Shenyang University of Chemical Technology, Shenyang 110042, China
 全文: PDF  HTML
摘要：

针对传统磨机负荷检测方法存在的测量精度低、性能不稳定等缺陷,建立一种基于筒体振动信号频谱特征提取的选择性极限学习机（ELM）集成方法.采用核主元分析（KPCA）提取振动信号频谱特征,避免输入信号维数过高引发维数灾难.在非线性频谱特征空间内选用学习速度快、泛化性好的ELM建立集成模型个体,有效克服了单一ELM个体模型存在的运行结果不稳定问题.基于遗传算法（GA）的子模型后续选择方法进一步排除部分劣势个体,构建泛化能力强的简约集成模型,降低计算复杂性.实验结果表明:该方法对于矿浆浓度、料球比、充填率磨机负荷参数具有较高的精度和稳定性.
Abstract:

Due to the low precision and unstable performance of the traditional measurements for the ball mill load, a selective extreme learning machine (ELM) ensemble model based on feature extraction of frequency spectrum from shell vibration signals was proposed. Kernel principal component analysis (KPCA) was used to extract the spectrum features of the shell vibration signals with high dimensions and colinearity in order to overcome the dimensional disaster. In the feature space of the frequency spectrum, ELM algorithm was inserted into the selective ensemble frame as a compoment model, since ELM runs much faster and provides better generalization performance than the other popular learning algorithm, which may overcome variations in different trials of simulation for a single ELM model. Selective ensemble based on GA algorithm futher excludes the bad ELM components from all the available ensembles. The concise resemble model has strong generalization capacity and low computation load. Experimental results show high stability and accuracy of the proposed method in terms of the mineral to ball volume ratio, pulp density and charge volume ratio in a ball mill.
出版日期: 2011-12-01
:  TP 29  
基金资助:

国家自然科学基金资助项目（61020106003，60874057）;中国博士后科学基金资助项目（20100471464）.
通讯作者: 柴天佑,男,教授,院士.     E-mail: tychai@mail.neu.edu.cn
作者简介: 赵立杰(1972－),女,博士后,研究方向为复杂工业过程建模和故障诊断.E-mail: zlj_lunlun@163.com
服务　　
把本文推荐给朋友
加入引用管理器
E-mail Alert
作者相关文章　　

引用本文:

赵立杰, 汤健, 柴天佑. 基于选择性极限学习机集成的磨机负荷软测量[J]. J4, 2011, 45(12): 2088-2092.

ZHAO Li-jie,TANG Jian,CHAI Tian-you. Soft sensor of mill load based on selective extreme
learning machine ensemble. J4, 2011, 45(12): 2088-2092.

链接本文:

https://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2011.12.003    或    https://www.zjujournals.com/eng/CN/Y2011/V45/I12/2088

［1］ TANG Jian, ZHAO Lijie, YUE Heng, et al. Experimental analysis of wet mill load based on vibration signals of laboratoryscale ball mill shell Source ［J］. Minerals Engineering, 2010, 23(9): 720-730.
［2］ 汤健,赵立杰,岳恒,等.磨机负荷检测方法研究综述［J］.控制工程,2010,17(05): 565-574.
TANG Jian, ZHAO Lijie, YUE Heng, et al. Present status and future developments of detectionmethod for mill load ［J］. Control Engineering of China, 2010, 17(05): 565-574.
［3］ ZENG Y G, FORSSBERG E. Application of vibration signal measurement for monitoring grinding parameters ［J］. Mechanical Systems and Signal Processing, 1994, 8(6): 703-713.
［4］ 汤健,赵立杰,岳恒,等.基于多源数据特征融合的球磨机负荷软测量［J］.浙江大学学报：工学版,2010,44(7): 1406-1413.
TANG Jian, ZHAO Lijie, YUE Heng,et al. Soft sensor for ball mill load based on multisource data feature fusion ［J］. Journal of Zhejiang University：Engineering Science, 2010, 44(7): 1406-1413.
［5］ SCHLKOPF B,SMOLA A J, MLLER K R. Nonlinear component analysis as a Kernel Eigenvalue Problem ［J］. Neuralcomputation,1998, 10(5): 1299-1319.
［6］ HANSEN L K, SALAMON P. Neural network ensemble ［J］. IEEE Transaction Pattern Analysis and Machine Intelligence, 1990, 12 (10): 993-1001.
［7］ ZHOU Zhihua, WU J, TANG W. Ensembling neural networks: Many could be better than all ［J］. Artificial Intelligence, 2002, 137(1/2): 239-263
［8］ HUANG Guangbin, ZHU Q Y , SIEW C K . Extreme learning machine: theory and applications ［J］. Neurocomputing, 2006, 70(1/2/3): 489-501.
［9］ YUAN Lan, YENG ChaiSoh, HUANG Guangbin. Ensemble of online sequential extreme learning machine ［J］. Neurocomputing, 2009, 72: 3391-3395.
[1] 赵立杰, 柴天佑, 袁德成, 刁晓坤. 增强操作工况识别可靠性的概率PLS-ELM方法[J]. J4, 2013, 47(10): 1747-1752.
[2] 汤健, 赵立杰, 岳恒, 柴天佑. 基于多源数据特征融合的球磨机负荷软测量[J]. J4, 2010, 44(7): 1406-1413.