Big data modeling of ball mill based on distributed support vector machine on Hadoop platform

doi:10.3969/j.issn.1000-1565.2017.03.014

Abstract

Abstract: In the era of big data environment,a large amount of data in thermal power plant is stored in the database and cannot be fully utilized.Because of the complicated process of the double inlet and double outlet mill system,the mathematical model is difficult to build.A method of modeling based on data mining is presented.The actual operation big data which impact the coal mill material is extracted.First, the K-Means clustering is used to delete outliers,and then the principal component analysis(PCA)is used to complete attribute reduction,at last the distributed support vector machine(D_SVM)is used to build a model on the Hadoop platform in MapReduce framework for the parallel computation.The results show that modeling time is greatly reduced due to the use of the method,and the accuracy and applicability of the model are very high. Therefore the model can be used to represent the actual material properties.

Key words: double inlet and double outlet mill, Hadoop platform, D_SVM, K-Means clustering, principal component analysis

CLC Number:

TP399

GAO Xuewei,FU Zhongguang,SUN Li,ZHANG Gang. Big data modeling of ball mill based on distributed support vector machine on Hadoop platform[J]. Journal of Hebei University (Natural Science Edition), 2017, 37(3): 309-315.

References

[1] 刘福国.基于数据挖掘的钢球磨煤机运行特性建模和优化[J].煤炭学报,2010,35(5):850-854.DOI:10.13225/j.cnki.jccs.2010.05.035. LIU F G.Performance modeling and optimization of ball mill based on data mining [J].Journal of China Coal Society,2010,35(5),850-854.DOI:10.13225/j.cnki.jccs.2010.05.035.
[2] 张小桃,倪维斗,李政,等.基于现场数据的中速磨煤机动态建模研究[J].热能动力工程,2004,19(6):614-616.DOI:10.3969/j.issn.1001-2060.2004.06.016. ZHANG X T,NI W D,LI Z,et al.A study of the dynamic modeling of a medium speed pulverizer based on on-site data[J].Journal of Engineering for Thermal Energy and Power,2004,19(6):614-616.DOI:10.3969/j.issn.1001-2060.2004.06.016.
[3] 刘定平,肖蔚然.应用最小二乘支持向量机和混合遗传算法的制粉系统优化控制[J].动力工程,2007,27(5),728-731.DOI:10.3321/j.issn:1000-6761.2007.05.016. LIU D P,XIAO W R.Optimizing control of pulverizing system based on least square supported vector machine and hybrid genetic algorithms [J].Journal of Power Engineering,2007,27(5):728-731.DOI:10.3321/j.issn:1000-6761.2007.05.016.
[4] 胡吉.火电厂中速磨煤机制粉系统仿真建模[D].武汉:华中科技大学,2005. HU J.Simulation and modeling of medium speed mill of coal pulverizing system in power plant [D].Wuhan:Huazhong University of Science and Technology,2005.
[5] 冯磊华.双进双出钢球磨煤机直吹式制粉系统建模及控制[D].长沙:中南大学,2012. FENG L H.Modeling and control of direct fired pulverized coal injection system with double inlet and double outlet ball mill[D].Changsha:Central South University,2012.
[6] 冯磊华,桂卫华,杨锋.改进LS-SVM的直吹式制粉出力软测量建模[J].电机与控制学报,2011,15(11):79-82.DOI:10.3969/j.issn.1007-449X.2011.11.014. FENG L H,GUI W H,YANG F.Soft sensor modeling for mill output of direct fired system based on improved least squares support vector machines [J].Electric Machines and Control,2011,15(11):79-82.DOI:10.3969/j.issn.1007-449X.2011.11.014.
[7] 姚树建.双进双出钢球磨煤机优化技术应用研究[D].长春:吉林大学,2005. YAO S J.The Study of the optimization technique application on the BBD Mill [D].Changchun:Jilin University,2005.
[8] 常绿,杨涛,姚树建,等.基于神经网络和遗传算法的磨煤机结构和工作参数的优化[J].热能动力工程,2007,22(1):69-72.DOI:10.3969/j.issn.1001-2060.2007.01.017. CHANG L,YANG T,YAO S J,et al.Optimization of structural and operating parameters of a ball mill based on a neural network and genetic algorithm[J].Journal of Engineering for Thermal Energy and Power,2007,22(1):69-72.DOI:10.3969/j.issn.1001-2060.2007.01.017.
[9] 张继东,唐国华,姚树建,等.基于神经网络算法计算双进双出磨煤机工作参数的研究[J].热力发电,2007,5:23-25.DOI:10.3969/j.issn.1002-3364.2007.05.006. ZHANG J D,TANG G H,YAO S J,et al.Study on calculation operating parameters of coal pulverize wish double inlet and double outlet based on neural network[J].Thermal Power Generation,2007,5:23-25.DOI:10.3969/j.issn.1002-3364.2007.05.006.
[10] 曹侠.基于神经网络的双进双出磨煤机混合模型研究[D].沈阳:沈阳工业大学,2008. CAO X.Study on hybrid model of BBD ball mill based on neural network [D].Shenyang:Shenyang University of Technology,2008.
[11] 万祥,胡念苏,韩鹏飞,等.大数据挖掘技术应用于汽轮机组运行性能优化的研究[J].中国电机工程学报,2016,36(2):459-467.DOI:10.13334/j.0258-8013.pcsee.2016.02.017. WAN X,HU N S,HAN P F,et al.Research on application of big data mining technology in performance optimization of steam turbines[J].Proceedings of the CSEE,2016,36(2):459-467.DOI:10.13334/j.0258-8013.pcsee.2016.02.017.
[12] 高学伟,付忠广.基于大数据分析的双进双出钢球磨煤机建模研究[J].煤炭技术,2016,35(12):267-270.DOI:10.13301/j.cnki.ct.2016.12.102. GAO X W,FU Z G.Research on modeling of double inlet and double outlet ball mill based on big data analysis[J].Coal Technology,2016,35(12):267-270.DOI:10.13301/j.cnki.ct.2016.12.102.
[13] 张良均,杨坦,肖钢,等.MATLAB数据分析与挖掘实战[M].北京:机械工业出版社,2015.
[14] 黄宜华,苗凯翔.深入理解大数据:大数据处理与编程实践[M].北京:机械工业出版社,2014.
[15] 张亚楠.基于Hadoop云计算平台的聚类算法并行化研究[D].包头:内蒙古科技大学,2013. ZHANG Y N.Research on parallel clustering algorithm based on hadoop cloud computing platform[D].Baotou:Inner Mongolia University of Science and Technology,2013.
[16] 齐敏芳,付忠广,景源,等.基于信息熵与主成分分析的火电机组综合评价方法[J].中国电机工程学报,2013,33(2):58-64. QI M F,FU Z G,JING Y,et al.A comprehensive evaluation method of power plant units based on information entropy and principal component analysis[J].Proceedings of the CSEE,2013,33(2):58-64.
[17] 雷学智.云计算平台下分布式支持向量机在煤炭行业分类预测应用[J].煤炭技术,2013,32(11):248-250.DOI:10.3969/j.issn.1008-8725.2013.11.130. LEI X Z.Application of distributed support vector machine based on cloud platform in coal system[J].Coal Technology,2013,32(11):248-250.DOI:10.3969/j.issn.1008-8725.2013.11.130.
[18] 牛科.基于Hadoop云平台的分布式支持向量机研究[D].太原:山西师范大学,2014. NIU K.Research of distributed support vector machine(SVM)based on Hadoop cloud platform[D].Taiyuan:Shanxi Normal University,2014.
[19] 崔文斌,温孚江,牟少敏,等.基于Hadoop的局部支持向量机[J].计算机研究与发展,2014,51(S):116-121. CUI W B,WEN F J,MOU S M,et al.Local support vector machine based on handoop[J].Journal of Computer research and development,2014,51(S):116-121.
[20] 张奕武.基于Hadoop分布式平台的SVM算法优化及应用[D].中山:中山大学,2012. ZHANG Y W.Optimization and application of svm algorithm based on hadoop distributed platform [D].Zhongshan:Zhongshan University,2012.