基于SVM的盘磨间隙软测量技术研究
摘要
磨浆是制浆造纸过程中不可或缺的环节，起着举足轻重的作用，它直接影响着成浆的质量和产量，生产效率以及资源能耗。盘磨间隙测量与控制是保证磨浆质量指标高低的一个重要标度，其作为磨浆设备的关键技术之一，一直深受国内外研究者的关注。然而，目前国内大多采用的接触式测量方法，只适合于测量初始盘磨间隙，而磨片磨损对测量的影响却并未考虑；而采用非接触式测量，虽然解决了上述问题，但对传感器的性能和材质要求特别高，且购买成本太高，一般很难达到。这也决定了直接测量方法虽然被普遍采用，但精度有待提高。
本课题是在未央区科技计划项目（201001）“APMP制浆工艺优化及关键设备控制”资助下完成的。在了解磨浆过程工艺机理的基础上，经过分析与借鉴前人研究成果，针对盘磨间隙在线直接测量存在的精度问题，提出了直接测量与间接测量（基于支持向量机的软测量）相结合的方法，探讨盘磨间隙测量的新方向，本文主要工作及成果如下：
1）盘磨间隙软测量数学模型研究
由于磨浆过程的复杂性，使得盘磨间隙直接测量精度很难实现，文中提出直接测量与软测量相结合的思想。通过学习和分析，得出利用比较容易测量的二次变量(辅助变量)间接测出盘磨间隙，在该软测量模型中，主导变量为盘磨间隙，辅助变量为浆流量，浆浓度和电能耗。分别采用基于支持向量机和神经网络，以及基于poly核函数和RBF核函数两组对比进行仿真，探索适合盘磨间隙测量的软测量模型，仿真结果表明：基于RBF支持向量机更适合本研究盘磨间隙软测量模型的结论，进一步说明了支持向量机是基于小样本的，全局最优以及泛化能力强的学习工具。利用软测量得到的结果，对直接测量值进行修正，先算出软测量与直接测量在可行域下的差值，盘磨间隙的最终值为直接测量值加上所得差值。
2）盘磨间隙测量技术工业实现研究
本文结合宁夏美利纸业磨浆过程，设计了一种盘磨间隙测量技术在DCS控制系统开发平台上的实现方法。硬件开发选用西门子S7-400PLC作为平台，软件开发选用Step7和WINCC作为平台，上位机除了WINCC界面设计，还要实现与MATLAB的对接，将该测量技术嵌入到工业控制系统中，真正意义上实现理论实践相结合的思想目标。
3）工作总结与未来研究展望
将软测量技术应用在盘磨间隙的测量中，避免测量依靠硬件的局限性，真正实现盘磨间隙软测量技术的工业化途径，并为以后研究工作提供了新的方向，使课题更有意义地在线指导磨浆过程盘磨间隙的实时控制。
The beating process is a necessary part of paper industry, and plays animportant role during paper making process, it also influences the quality oryield of pulp, production efficiency and energy consumption. Measuring andcontrolling the disc refiner gap is an important scale in guaranteeing the qualityindicator of pulp, as one of the key technologies of refiner equipments, it hasbeen taken more attention by researchers all over the world. However, at present,the method of contact disc refiner gap measurement is adopted mostly at home,it’s only suitable for measuring the disc refiner initial gap, without consideringthe influence of the plate wear, the method of non-contact disc refiner gapmeasurement can solve the above problem, but the requirement of sensorperformance and material is special high, the price of sensor is also expensive,it’s hard to meet the requirement. It also determines that the direct measurementhas been widely used, but the accuracy will be needed to improve.
This subject is supported by the science and technology plan projects(201001) of Wei Yang area, named “APMP pulping optimization and the keyequipment control”. On the basis of introducing the refining process mechanismand analyzing other's research results, aiming at the accuracy problem ofmeasuring the disc refiner gap directly, the combination idea of directmeasurement and indirect measurement(soft measurement) is proposed in thispaper, for further study at the disc refiner gap measurement, the main contentsand achievements of the paper are summarized as follows：
1）Research on the mathematical model of soft measurement for discrefiner gap
Because the refining process is very complex, the accuracy of directmeasurement for disc refiner gap is difficult to achieve, the combination idea ofdirect measurement and soft measurement based on SVM (support vector machine) is proposed in this paper. Through studying and analyzing, it can usethe secondary variable measured easily to measure the disc refiner gap, in thesoft measurement model, the disc refiner gap is selected as leading variable, thepulp flow, pulp consistency and power consumption are selected as secondaryvariables. Two groups of contrast are been used to build the soft measurementmodel respectively, the simulation tools select between neural network andsupport vector machine, as well as between poly kernel function and RBFkernel function, the simulation results show that SVM based on RBF is betterthan NN (neural network) in modeling the soft measurement of disc refiner gap,it further illustrates that SVM is a study tool with small sample, global optimumand high generalization performance. The direct measurement values can beamended by using the soft measurement result, the differences between the softmeasurement and direct measurement value are worked out firstly in thefeasible region, the final disc refiner gap is the sum of direct measurementvalues and differences.
2) Research on the industrial design about soft measurement technology ofdisc refiner gap
Associating with the beating process of Ningxia Meili industry, animplementation method of measurement technology for disc refiner gap isdesigned, which based on the development platform of DCS. S7-400PLC isselected as hardware development platform, WINCC and Step7are selected assoftware development platforms. It needs to not only design the WINCCinterface but only achieve docking with the MATLAB software on uppercomputer, so that the measurement technology can be embed into DCS controlsystem, the goal of combining theory and practice is realized indeed.
3) Summarize the whole work and prospect the further study
The soft measurement technology will be applied in the measurementprocess, therefore, the measurement limitations of relying on hardware can beavoided, the soft measurement technology of disc refiner gap will arrive at theindustrialization way, and provide a new direction for the future research work,the real-time control of the disc refiner gap will be realized by online guide.
本课题是在未央区科技计划项目（201001）“APMP制浆工艺优化及关键设备控制”资助下完成的。在了解磨浆过程工艺机理的基础上，经过分析与借鉴前人研究成果，针对盘磨间隙在线直接测量存在的精度问题，提出了直接测量与间接测量（基于支持向量机的软测量）相结合的方法，探讨盘磨间隙测量的新方向，本文主要工作及成果如下：
1）盘磨间隙软测量数学模型研究
由于磨浆过程的复杂性，使得盘磨间隙直接测量精度很难实现，文中提出直接测量与软测量相结合的思想。通过学习和分析，得出利用比较容易测量的二次变量(辅助变量)间接测出盘磨间隙，在该软测量模型中，主导变量为盘磨间隙，辅助变量为浆流量，浆浓度和电能耗。分别采用基于支持向量机和神经网络，以及基于poly核函数和RBF核函数两组对比进行仿真，探索适合盘磨间隙测量的软测量模型，仿真结果表明：基于RBF支持向量机更适合本研究盘磨间隙软测量模型的结论，进一步说明了支持向量机是基于小样本的，全局最优以及泛化能力强的学习工具。利用软测量得到的结果，对直接测量值进行修正，先算出软测量与直接测量在可行域下的差值，盘磨间隙的最终值为直接测量值加上所得差值。
2）盘磨间隙测量技术工业实现研究
本文结合宁夏美利纸业磨浆过程，设计了一种盘磨间隙测量技术在DCS控制系统开发平台上的实现方法。硬件开发选用西门子S7-400PLC作为平台，软件开发选用Step7和WINCC作为平台，上位机除了WINCC界面设计，还要实现与MATLAB的对接，将该测量技术嵌入到工业控制系统中，真正意义上实现理论实践相结合的思想目标。
3）工作总结与未来研究展望
将软测量技术应用在盘磨间隙的测量中，避免测量依靠硬件的局限性，真正实现盘磨间隙软测量技术的工业化途径，并为以后研究工作提供了新的方向，使课题更有意义地在线指导磨浆过程盘磨间隙的实时控制。
The beating process is a necessary part of paper industry, and plays animportant role during paper making process, it also influences the quality oryield of pulp, production efficiency and energy consumption. Measuring andcontrolling the disc refiner gap is an important scale in guaranteeing the qualityindicator of pulp, as one of the key technologies of refiner equipments, it hasbeen taken more attention by researchers all over the world. However, at present,the method of contact disc refiner gap measurement is adopted mostly at home,it’s only suitable for measuring the disc refiner initial gap, without consideringthe influence of the plate wear, the method of non-contact disc refiner gapmeasurement can solve the above problem, but the requirement of sensorperformance and material is special high, the price of sensor is also expensive,it’s hard to meet the requirement. It also determines that the direct measurementhas been widely used, but the accuracy will be needed to improve.
This subject is supported by the science and technology plan projects(201001) of Wei Yang area, named “APMP pulping optimization and the keyequipment control”. On the basis of introducing the refining process mechanismand analyzing other's research results, aiming at the accuracy problem ofmeasuring the disc refiner gap directly, the combination idea of directmeasurement and indirect measurement(soft measurement) is proposed in thispaper, for further study at the disc refiner gap measurement, the main contentsand achievements of the paper are summarized as follows：
1）Research on the mathematical model of soft measurement for discrefiner gap
Because the refining process is very complex, the accuracy of directmeasurement for disc refiner gap is difficult to achieve, the combination idea ofdirect measurement and soft measurement based on SVM (support vector machine) is proposed in this paper. Through studying and analyzing, it can usethe secondary variable measured easily to measure the disc refiner gap, in thesoft measurement model, the disc refiner gap is selected as leading variable, thepulp flow, pulp consistency and power consumption are selected as secondaryvariables. Two groups of contrast are been used to build the soft measurementmodel respectively, the simulation tools select between neural network andsupport vector machine, as well as between poly kernel function and RBFkernel function, the simulation results show that SVM based on RBF is betterthan NN (neural network) in modeling the soft measurement of disc refiner gap,it further illustrates that SVM is a study tool with small sample, global optimumand high generalization performance. The direct measurement values can beamended by using the soft measurement result, the differences between the softmeasurement and direct measurement value are worked out firstly in thefeasible region, the final disc refiner gap is the sum of direct measurementvalues and differences.
2) Research on the industrial design about soft measurement technology ofdisc refiner gap
Associating with the beating process of Ningxia Meili industry, animplementation method of measurement technology for disc refiner gap isdesigned, which based on the development platform of DCS. S7-400PLC isselected as hardware development platform, WINCC and Step7are selected assoftware development platforms. It needs to not only design the WINCCinterface but only achieve docking with the MATLAB software on uppercomputer, so that the measurement technology can be embed into DCS controlsystem, the goal of combining theory and practice is realized indeed.
3) Summarize the whole work and prospect the further study
The soft measurement technology will be applied in the measurementprocess, therefore, the measurement limitations of relying on hardware can beavoided, the soft measurement technology of disc refiner gap will arrive at theindustrialization way, and provide a new direction for the future research work,the real-time control of the disc refiner gap will be realized by online guide.
引文
[1]董继先.高浓磨浆机建模、设计及APMP磨浆过程优化[D]．西安:陕西科技大学,2010．
[2]王宝金.热磨机磨盘真实间隙非接触测量的研究[J]．林业机械与木工设备.2004,6(32):12-15.
[3]王宝金.热磨机磨盘实际间隙精确测量与主轴运行状态监刚的研究[D].南京：南京林业大学,2004.
[4]王艳萍,朱芳.动、静磨片盘的间隙在热磨机上的重要性[J].林产工业,2006,3(4):41-42.
[5]张强,姚百让.热磨机“碰盘”现象分析[J].实用技术,2005(9):15-19.
[6] Darryl Dodson-Edgars. Gap Wear and Tram Measurement System and Method forGrinding Machine [P].United States:4820980,1989-04-11.
[7]张辉,李忠正.新的盘式磨浆机磨浆间隙在线测量技术[J].中国造纸学报,2008,23(1):85-89.
[8] John J Egan III, Christopher L Demler. Paper Pulp Refiner Control System and Method
[P].United States:5947394,1999-09-7.
[9]王孝红,刘文光,于宏亮.工业过程软测量研究[J].济南大学学报,2009,23(1):80-86．
[10]冯瑞,张浩然,邵惠鹤.基于SVM的软测量建模[J].信息与控制,2002,31(6):567-571.
[11]唐飞.基于支持向量机软测量技术的研究和应用[D].浙江:浙江工业大学,2007.
[12] Wei Tang, Junmei Huang, Dan Liu, Mengxiao Wang. Soft Measurement of Black LiquorBaume Degree&Its Implementation[C].International Conference on Mechanics andAutomation,Xi'an,China,2010:4-7.
[13] Wei TANG, Yan zi ZHAO. Structure of Mixture Kernel Function and Its Application inSoft Measurement about Beating Degree[C].2011International Conference ofInformation Technology, Applied Mechanics and Materials. Xi'an, China,2011:204-207.
[14]李方方,赵英凯,姜志兵.基于支持向量机的油品质量预测[J].计算机工程,2007,33(2):200-202.
[15]刘江华,程君实,陈佳品.支持向量机训练算法综述[J].信息与控制,2002,31(1):45-50.
[16]张学工.关于统计学习理论与支持向量机[J].自动化学报,2000,26(1):32-42.
[17]阎威武.支持向量机理论、方法和应用研究[D]．上海:上海交通大学,2003．
[18] Steve R. Gunn. Support Vector Machines for Classification and Regression[R]. England:University of Southampton,1998.
[19]张辉,李忠正.盘式磨浆机技术研究进展与趋势[J].中国造纸,2007,26(10):40-45.
[20]汤伟,罗斌,周红,王孟效.打浆过程控制的新进展[J].中国造纸学报,2009,24(1):122-129.
[21]梁钱华.造纸高浓度锥形磨浆机的研究与设计[D]．西安:陕西科技大学,2008．
[22]王秀云.打浆过程的优化控制[D]．西安:陕西科技大学,2005．
[23]孙优贤等．我国制浆造纸工业自动化的现状和发展基础自动化[J].1998,5(1):1-6．
[24]潘福迪.制浆造纸工艺基本原理与应用[M].大连:大连理工大学出版社,1990:88-116．
[25]造纸技术网．打浆工艺概述：打浆对纤维的作用(2000-2010)．http://www.zaozhijs.com/news/ShowArticle.asp-ArticleID=3003&Page=3.htm.
[26]何衍庆,马欣,王强.盘磨机打浆过程的控制[J].中国造纸,1997,23(2):35-37．
[27]罗斌.APMP高浓打浆过程建模与优化研究[D].西安:陕西科技大学,2008．
[28] Jixian Dong, Jangong Gu. Formulation and Analysis of Dynamic and EnergyConsumption Models in High Consistency Pulp-Refining,3rdISETPP,2006(11):384-388.
[29]叶冠军.高浓盘磨打浆和制浆[J].湖北造纸,1996(1):37-42.
[30]李涛,汪检涛,杨红霞.七孔高位浆箱在打浆中的应用[J].中华纸业,2003,24(2):55-57.
[31]梅云珍.高位浆箱在高档薄页纸打浆中的作用[J].纸和造纸,2002(4):22-23.
[32]黄海平,张艳,张铁兵.双圆盘磨浆机的打浆特性及工艺流程.黑龙江造纸,2002,30(1):26-28.
[33]梁孝斌,胡庆喜等.ZDPM450型中浓盘磨机串联打浆的研究[J].广东造纸,1998(4):15-18.
[34]张洪字.高维七孔箱在针叶木打浆中的应用[J].西南造纸,2004,33(3):56-58.
[35]刘士亮,陈中豪等.ZDPM型中浓液压盘磨机中浓打浆的生产应用[J].西北轻工业学院学报,2000,18(1):5-11.
[36]刘征怀.双圆盘磨浆机的打浆特性及工艺流程[J].中国造纸,1992(4):50-55.
[37]李世扬等. ZDPM盘磨机中浓打浆生产使用效果(一)[J].广东造纸,1998(5):18-20.
[38]王立鹏,张运展,梁富政.大麻木质芯秆碱H2O2机械浆的研究[J].中国造纸学报,2002,17(2):11-14.
[39]何洁,刘忠.大麻芯秆烧碱-蒽醌法制浆脱木质素反应历程[J].中国造纸,2011,30(6):41-46．
[40]曹忠荣,黄洛华,曲岩春.三种速生建筑材树种制造中纤板的制浆工艺和磨片选型[J].木材工业,1999,13(3):3-6.
[41]杨广富,马冬梅等.对北方针叶木化学热磨机械浆的研究[J].湖南造纸,1996(4):20-25.
[42]刘正非.用国产装置生产云南松CTMP浆[J].中国造纸,1995(2):13-16.
[43]姚俊,王平.盘磨机磨盘间隙的测量与调节方法[J].中国造纸,2012,31(1):67-71.
[44] K. Khawaja. Wheel-Tooling Gap Measurement System for Conform ExtrusionMachinery Based on a Capacitive Sensor [J].2005,(12):394-401.
[45]张家友.大型盘磨机位移检测系统改造[D].四川:四川大学,2006.
[46]马坚刚.间隙测量应用系统研究[D].南京:南京航空航天大学,2006.
[47]卢盛江.热磨机磨盘加压与间隙微调装置浅析[J].林业机械,1994(4):15-16.
[48]张辉,李忠正,范文.电涡流传感器测量旋转齿盘位移的特性研究[J].2008,28(1):44-49.
[49] Hector N．Guerrero. Magnetic Proximity Sensor for Measuring Gap Between OpposedRefiner Plates [P].United States:4950986,1990-8-21.
[50]陈克复主编．制浆造纸机械与设备(下)(高等学校教材第二版)[M]．北京:中国轻工业出版社,2003:35-67.
[51]张辉,李忠正.造纸盘式磨浆机磨浆间隙在线,精确测量技术原理与方法[D].南京:南京林业大学,2008.
[52] Johansson Ola M. Refiner Measurement System Method[P]. United States:091492A1,2002-07-11.
[53]贺毅,赵望达,刘勇求.软测量技术及其人工神经元网络建模方法探讨[J].自动化博览,2005(S2):79-82．
[54] Jian Tang, Lijie Zhao, et al. Soft Sensor Modeling of Ball Mill Load via PrincipalComponent Analysis and Support Vector Machines[J].Neural Network Research&Application, LNEE67,803–810.
[55] Bergstrom, J. et al. DESIGN OF REFINER PLATES [C]. Proceedings of the NationalElectronics Conference,1973(1):23-26.
[56]葛培琪,周以齐,刘鸣．新型锥形磨浆机的性能分析[J]．纸和造纸,2002(3):26-28．
[57] Alahautala, Taito.et al. Optical measurement of pulp quantity in a rotating disc refiner[J].Measurement Science and Technology,2004,15(11):2256-2262.
[58]刘士亮．磨片磨损对打浆生产使用效果的影响[J]．黑龙江造纸,2009(3):10-14．
[59] Jixian Dong, Jangong Gu. Formulation and Analysis of Dynamic and EnergyConsumption Models in High Consistency Pulp-Refining[C],3rdISETPP,2006(11):384-388.
[60]邓乃扬,田英杰.数据挖掘中的新方法--支持向量机[M],科学出版社,2004,10-95.
[61]王旭东,邵惠鹤.RBF神经网络理论及其在控制中的应用[J]．信息与控制,1997,26(4):272-284．
[62] Tang Faming, Chen Mianyun&Wang Zhongdong. New approach to training supportvector machines [J].Journal of Systems Engineering and Electronics,2006,17(1):200-205.
[63]李海生.支持向量机回归算法与应用研究[D].华南理工大学,2005,5.
[64] Man Gyunna, Heon Young Yang and Dong Hyuk Lim. A Soft Sensing Model for Feedwater Flow Rate Using Fuzzy Support Vector Regression [J].Nuclear Engineering andTechnology,2008,40(1):234-239.
[65] Yan-zi ZHAO, Wei TANG. Function Fitting about Internal Stress of Ceramic Paste Basedon BP-NN and SVM [C].2011International Conference of Information Technology,Computer Engineering and Management Sciences, Nan Jing, China,2011:294-296.
[66]王华忠,俞金寿.核函数方法及其模型选择[J].江南大学学报(自然科学版),2006,5(4):2-6．
[67]颜根廷,马广富等.一种混合核函数支持向量机算法[J].哈尔滨工业大学学报,2007,3(11):1705-1706.
[68]缪海峰.专家控制在APMP盘磨控制中的应用研究[D].西安:陕西科技大学,2007．
[69]苏昆哲.深入浅出西门WinCC V6[M]．北京:北京航空航天大学出版社,2004:47-50．
[2]王宝金.热磨机磨盘真实间隙非接触测量的研究[J]．林业机械与木工设备.2004,6(32):12-15.
[3]王宝金.热磨机磨盘实际间隙精确测量与主轴运行状态监刚的研究[D].南京：南京林业大学,2004.
[4]王艳萍,朱芳.动、静磨片盘的间隙在热磨机上的重要性[J].林产工业,2006,3(4):41-42.
[5]张强,姚百让.热磨机“碰盘”现象分析[J].实用技术,2005(9):15-19.
[6] Darryl Dodson-Edgars. Gap Wear and Tram Measurement System and Method forGrinding Machine [P].United States:4820980,1989-04-11.
[7]张辉,李忠正.新的盘式磨浆机磨浆间隙在线测量技术[J].中国造纸学报,2008,23(1):85-89.
[8] John J Egan III, Christopher L Demler. Paper Pulp Refiner Control System and Method
[P].United States:5947394,1999-09-7.
[9]王孝红,刘文光,于宏亮.工业过程软测量研究[J].济南大学学报,2009,23(1):80-86．
[10]冯瑞,张浩然,邵惠鹤.基于SVM的软测量建模[J].信息与控制,2002,31(6):567-571.
[11]唐飞.基于支持向量机软测量技术的研究和应用[D].浙江:浙江工业大学,2007.
[12] Wei Tang, Junmei Huang, Dan Liu, Mengxiao Wang. Soft Measurement of Black LiquorBaume Degree&Its Implementation[C].International Conference on Mechanics andAutomation,Xi'an,China,2010:4-7.
[13] Wei TANG, Yan zi ZHAO. Structure of Mixture Kernel Function and Its Application inSoft Measurement about Beating Degree[C].2011International Conference ofInformation Technology, Applied Mechanics and Materials. Xi'an, China,2011:204-207.
[14]李方方,赵英凯,姜志兵.基于支持向量机的油品质量预测[J].计算机工程,2007,33(2):200-202.
[15]刘江华,程君实,陈佳品.支持向量机训练算法综述[J].信息与控制,2002,31(1):45-50.
[16]张学工.关于统计学习理论与支持向量机[J].自动化学报,2000,26(1):32-42.
[17]阎威武.支持向量机理论、方法和应用研究[D]．上海:上海交通大学,2003．
[18] Steve R. Gunn. Support Vector Machines for Classification and Regression[R]. England:University of Southampton,1998.
[19]张辉,李忠正.盘式磨浆机技术研究进展与趋势[J].中国造纸,2007,26(10):40-45.
[20]汤伟,罗斌,周红,王孟效.打浆过程控制的新进展[J].中国造纸学报,2009,24(1):122-129.
[21]梁钱华.造纸高浓度锥形磨浆机的研究与设计[D]．西安:陕西科技大学,2008．
[22]王秀云.打浆过程的优化控制[D]．西安:陕西科技大学,2005．
[23]孙优贤等．我国制浆造纸工业自动化的现状和发展基础自动化[J].1998,5(1):1-6．
[24]潘福迪.制浆造纸工艺基本原理与应用[M].大连:大连理工大学出版社,1990:88-116．
[25]造纸技术网．打浆工艺概述：打浆对纤维的作用(2000-2010)．http://www.zaozhijs.com/news/ShowArticle.asp-ArticleID=3003&Page=3.htm.
[26]何衍庆,马欣,王强.盘磨机打浆过程的控制[J].中国造纸,1997,23(2):35-37．
[27]罗斌.APMP高浓打浆过程建模与优化研究[D].西安:陕西科技大学,2008．
[28] Jixian Dong, Jangong Gu. Formulation and Analysis of Dynamic and EnergyConsumption Models in High Consistency Pulp-Refining,3rdISETPP,2006(11):384-388.
[29]叶冠军.高浓盘磨打浆和制浆[J].湖北造纸,1996(1):37-42.
[30]李涛,汪检涛,杨红霞.七孔高位浆箱在打浆中的应用[J].中华纸业,2003,24(2):55-57.
[31]梅云珍.高位浆箱在高档薄页纸打浆中的作用[J].纸和造纸,2002(4):22-23.
[32]黄海平,张艳,张铁兵.双圆盘磨浆机的打浆特性及工艺流程.黑龙江造纸,2002,30(1):26-28.
[33]梁孝斌,胡庆喜等.ZDPM450型中浓盘磨机串联打浆的研究[J].广东造纸,1998(4):15-18.
[34]张洪字.高维七孔箱在针叶木打浆中的应用[J].西南造纸,2004,33(3):56-58.
[35]刘士亮,陈中豪等.ZDPM型中浓液压盘磨机中浓打浆的生产应用[J].西北轻工业学院学报,2000,18(1):5-11.
[36]刘征怀.双圆盘磨浆机的打浆特性及工艺流程[J].中国造纸,1992(4):50-55.
[37]李世扬等. ZDPM盘磨机中浓打浆生产使用效果(一)[J].广东造纸,1998(5):18-20.
[38]王立鹏,张运展,梁富政.大麻木质芯秆碱H2O2机械浆的研究[J].中国造纸学报,2002,17(2):11-14.
[39]何洁,刘忠.大麻芯秆烧碱-蒽醌法制浆脱木质素反应历程[J].中国造纸,2011,30(6):41-46．
[40]曹忠荣,黄洛华,曲岩春.三种速生建筑材树种制造中纤板的制浆工艺和磨片选型[J].木材工业,1999,13(3):3-6.
[41]杨广富,马冬梅等.对北方针叶木化学热磨机械浆的研究[J].湖南造纸,1996(4):20-25.
[42]刘正非.用国产装置生产云南松CTMP浆[J].中国造纸,1995(2):13-16.
[43]姚俊,王平.盘磨机磨盘间隙的测量与调节方法[J].中国造纸,2012,31(1):67-71.
[44] K. Khawaja. Wheel-Tooling Gap Measurement System for Conform ExtrusionMachinery Based on a Capacitive Sensor [J].2005,(12):394-401.
[45]张家友.大型盘磨机位移检测系统改造[D].四川:四川大学,2006.
[46]马坚刚.间隙测量应用系统研究[D].南京:南京航空航天大学,2006.
[47]卢盛江.热磨机磨盘加压与间隙微调装置浅析[J].林业机械,1994(4):15-16.
[48]张辉,李忠正,范文.电涡流传感器测量旋转齿盘位移的特性研究[J].2008,28(1):44-49.
[49] Hector N．Guerrero. Magnetic Proximity Sensor for Measuring Gap Between OpposedRefiner Plates [P].United States:4950986,1990-8-21.
[50]陈克复主编．制浆造纸机械与设备(下)(高等学校教材第二版)[M]．北京:中国轻工业出版社,2003:35-67.
[51]张辉,李忠正.造纸盘式磨浆机磨浆间隙在线,精确测量技术原理与方法[D].南京:南京林业大学,2008.
[52] Johansson Ola M. Refiner Measurement System Method[P]. United States:091492A1,2002-07-11.
[53]贺毅,赵望达,刘勇求.软测量技术及其人工神经元网络建模方法探讨[J].自动化博览,2005(S2):79-82．
[54] Jian Tang, Lijie Zhao, et al. Soft Sensor Modeling of Ball Mill Load via PrincipalComponent Analysis and Support Vector Machines[J].Neural Network Research&Application, LNEE67,803–810.
[55] Bergstrom, J. et al. DESIGN OF REFINER PLATES [C]. Proceedings of the NationalElectronics Conference,1973(1):23-26.
[56]葛培琪,周以齐,刘鸣．新型锥形磨浆机的性能分析[J]．纸和造纸,2002(3):26-28．
[57] Alahautala, Taito.et al. Optical measurement of pulp quantity in a rotating disc refiner[J].Measurement Science and Technology,2004,15(11):2256-2262.
[58]刘士亮．磨片磨损对打浆生产使用效果的影响[J]．黑龙江造纸,2009(3):10-14．
[59] Jixian Dong, Jangong Gu. Formulation and Analysis of Dynamic and EnergyConsumption Models in High Consistency Pulp-Refining[C],3rdISETPP,2006(11):384-388.
[60]邓乃扬,田英杰.数据挖掘中的新方法--支持向量机[M],科学出版社,2004,10-95.
[61]王旭东,邵惠鹤.RBF神经网络理论及其在控制中的应用[J]．信息与控制,1997,26(4):272-284．
[62] Tang Faming, Chen Mianyun&Wang Zhongdong. New approach to training supportvector machines [J].Journal of Systems Engineering and Electronics,2006,17(1):200-205.
[63]李海生.支持向量机回归算法与应用研究[D].华南理工大学,2005,5.
[64] Man Gyunna, Heon Young Yang and Dong Hyuk Lim. A Soft Sensing Model for Feedwater Flow Rate Using Fuzzy Support Vector Regression [J].Nuclear Engineering andTechnology,2008,40(1):234-239.
[65] Yan-zi ZHAO, Wei TANG. Function Fitting about Internal Stress of Ceramic Paste Basedon BP-NN and SVM [C].2011International Conference of Information Technology,Computer Engineering and Management Sciences, Nan Jing, China,2011:294-296.
[66]王华忠,俞金寿.核函数方法及其模型选择[J].江南大学学报(自然科学版),2006,5(4):2-6．
[67]颜根廷,马广富等.一种混合核函数支持向量机算法[J].哈尔滨工业大学学报,2007,3(11):1705-1706.
[68]缪海峰.专家控制在APMP盘磨控制中的应用研究[D].西安:陕西科技大学,2007．
[69]苏昆哲.深入浅出西门WinCC V6[M]．北京:北京航空航天大学出版社,2004:47-50．