刨花板热压位置伺服智能控制方法研究
|
摘要
随着我国森林资源的紧缺,人造板的生产已经受到板材生产企业的充分重视。刨花板以其生产和使用的种种优点,给刨花板生产企业的发展带来了前所未有的良机。刨花板生产中影响板材最终产品质量的重要步骤是热压工序。论文以刨花板热压工艺为研究对象,分别从板坯内部热质传递和热压位置伺服控制两方面进行深入研究。
由于实际生产中板坯芯层温度难以测量,本文借助BP (Back Propagation)神经网络自学习能力,通过训练已有的实验数据,来预测板坯芯层的导热参数从而推导出板坯芯层温度。并用改进的ELM (Extreme Learning Machine-ELM)算法进一步提高了预测精度。所得的导热性能预测方法可以改进实际生产中的实时热压控制工艺。
在研究了热压机相关工作原理的基础上,以刨花板热压机为控制对象,针对热压位置控制存在的非线性特性及平稳无超调控制的特殊工艺要求,提出采用模糊自适应的控制方法对板坯厚度进行控制。该方法通过模糊优化输入参数对热压控制器实施在线自整定,得以提高刨花板热压机控制器的控制能力。
通过MATLAB计算机仿真研究,验证上述理论结果。先与常规位置伺服PID控制器进行对比,验证了模糊自适应位置伺服控制器在不同组初值的情况下均能达到迅速稳定和无超调的响应,显著提升了刨花板热压机的位置伺服控制系统的稳定性及控制精度。热压控制不应单一考虑热压机的位置控制,还应充分考虑热压过程中热质扰动对板坯内部环境的影响。因此,在本控制系统中增加了抗扰动模糊模块。通过仿真结果表明,加入抗扰动模糊模块之后,位置伺服控制系统的抗干扰能力明显增强,提高了最终板材产品的质量。
With the shortage of wood resources, people are increasingly paying attention to wood-based panel production. Particleboard has brought about new opportunities and challenges to the development of particleboard industry by its own advantages. In particleboard production process, hot-pressing is one of the most important processes. It directly influences the quality and the yields of the final board products. To particleboard hot-pressing process as the research object, this paper respectively from two aspects of the slab heat and mass transfer and the position servo control have been studied.
It is difficult to measure the slab core temperature in the actual production. Through self-learning ability of Back Propagation neural network and training of existing experimental data, this sets up a model of forecast the thermal conductivity under various conditions. In order to improve the prediction precision, Extreme Learning Machine is introduced into this model. At the same time, this research has a certain significance in the future actual production.
According to some phenomena such as non-linear characteristic, the special process requirements of steady control without overshoot, and base on the principle of the pressing machine, we adopt the fuzz self-tuning strategy to control the slab thickness. The method can improve control precision and stabilization of the electro-hydraulic position servo system, which by fuzzy optimization implement parameters online self-tuning.
MATLAB computer simulation verifies the theoretical results. First of all, compared with the conventional PID controller, the fuzzy adaptive controller in the case of the initial value can be fast stable and no oyershoot response, which markedly improved the particleboard hot-pressing control accuracy and stability in the servo control system. Hot-pressing control not only should consider the location of a single control process, but also should take account of the heat and mass disturbance effect within the slab. Therefore, the system increased thermal mass anti-disturbing fuzzy module. The simulation results show that the addition of anti-disturbing fuzzy module, the position servo control system significantly enhanced anti-disturbing capability and improved the quality of the final particleboard products.
由于实际生产中板坯芯层温度难以测量,本文借助BP (Back Propagation)神经网络自学习能力,通过训练已有的实验数据,来预测板坯芯层的导热参数从而推导出板坯芯层温度。并用改进的ELM (Extreme Learning Machine-ELM)算法进一步提高了预测精度。所得的导热性能预测方法可以改进实际生产中的实时热压控制工艺。
在研究了热压机相关工作原理的基础上,以刨花板热压机为控制对象,针对热压位置控制存在的非线性特性及平稳无超调控制的特殊工艺要求,提出采用模糊自适应的控制方法对板坯厚度进行控制。该方法通过模糊优化输入参数对热压控制器实施在线自整定,得以提高刨花板热压机控制器的控制能力。
通过MATLAB计算机仿真研究,验证上述理论结果。先与常规位置伺服PID控制器进行对比,验证了模糊自适应位置伺服控制器在不同组初值的情况下均能达到迅速稳定和无超调的响应,显著提升了刨花板热压机的位置伺服控制系统的稳定性及控制精度。热压控制不应单一考虑热压机的位置控制,还应充分考虑热压过程中热质扰动对板坯内部环境的影响。因此,在本控制系统中增加了抗扰动模糊模块。通过仿真结果表明,加入抗扰动模糊模块之后,位置伺服控制系统的抗干扰能力明显增强,提高了最终板材产品的质量。
With the shortage of wood resources, people are increasingly paying attention to wood-based panel production. Particleboard has brought about new opportunities and challenges to the development of particleboard industry by its own advantages. In particleboard production process, hot-pressing is one of the most important processes. It directly influences the quality and the yields of the final board products. To particleboard hot-pressing process as the research object, this paper respectively from two aspects of the slab heat and mass transfer and the position servo control have been studied.
It is difficult to measure the slab core temperature in the actual production. Through self-learning ability of Back Propagation neural network and training of existing experimental data, this sets up a model of forecast the thermal conductivity under various conditions. In order to improve the prediction precision, Extreme Learning Machine is introduced into this model. At the same time, this research has a certain significance in the future actual production.
According to some phenomena such as non-linear characteristic, the special process requirements of steady control without overshoot, and base on the principle of the pressing machine, we adopt the fuzz self-tuning strategy to control the slab thickness. The method can improve control precision and stabilization of the electro-hydraulic position servo system, which by fuzzy optimization implement parameters online self-tuning.
MATLAB computer simulation verifies the theoretical results. First of all, compared with the conventional PID controller, the fuzzy adaptive controller in the case of the initial value can be fast stable and no oyershoot response, which markedly improved the particleboard hot-pressing control accuracy and stability in the servo control system. Hot-pressing control not only should consider the location of a single control process, but also should take account of the heat and mass disturbance effect within the slab. Therefore, the system increased thermal mass anti-disturbing fuzzy module. The simulation results show that the addition of anti-disturbing fuzzy module, the position servo control system significantly enhanced anti-disturbing capability and improved the quality of the final particleboard products.
引文
[1]Xi-ping Wang, Ming-xi Shi. Matlab for Forecasting of Electric Power Load Based on BP Neural Network. Intelligent Computing and Information Science,2011,134(1):636-642
[2]莫建虎,李落星等.采用BP神经网络和挤压形状因子预测挤压力与挤压出口温度.矿冶工程,2009,29(2)：90～94
[3]王志春,李忠飞.基于BP网络的钢坯表面温度预测模型的设计.冶金自动化,2009,23(2)：56～60
[4]崔宝侠,周硕.基于改进Elman神经网络的石膏纤维板厚度控制建模.沈阳工业大学学,2008,30(2)：98～120
[5]冯晓华,刘鹏,李静,徐美玲.一种改进的基于神经网络的板形缺陷识别方法.重型机械,2009(6)：77～81
[6]诸静.模糊控制理论与系统原理.北京：机械工业出版社,2005：37-65
[7]许建,肖维荣.液压位置伺服系统的模糊PID控制研究.自动化信息,2010,(6)：45～49
[8]孙延明,曹军,刘亚秋.模糊PID控制在中密度纤维板施胶中的应用研究.木材加工机械,2007(1)：82～89
[9]王东林.中密度纤维板热压智能控制系统的研究.北京林业大学,2008：10～47
[10]王补宣.多孔介质中的对流传热传质.西安交通大学学报,1994,28(5)：51～58
[11]杜朝刚,陈天全,常建民.刨花板热压过程中的传热传质研究现状和展望.林产工业,2004,31(5)：10～14
[12]Kayihan, F.A.Johnson. Heat and Moisture Movement in Wood Compolite Material during the Pressing Operation-A Simplified Model. Numerical Methods in Heat Transfer, 1983(2):511-531
[13]Harless, T.E.G., E.G. Wagner, etal. A Model to Predict the Density Profile of Parti-cleboard. Wood Fiber Sci,1987,19(1):81～92
[14]Humphrey, P.E, H.Thoemen. The Continuous Pressing of Wood-based Panel:An Analytical Simulation Model.5th Annual Pacific Rim Bio-Based Composites Symposium, 2000:123～132
[15]刘应安,蔡智勇.刨花板热压过程中热传导的几个问题.林业科技开发,1992(4)：43～45
[16]张帝树等.刨花板板坯含水率及其梯度对热传递速度的影响.林产工业,1993,20(4)：9-11
[17]刘正添,王洁瑛,于辉.影响刨花板热压传热过程因素的研究.北京林业大学学报,1995,17(2)：64～71
[18]黎爱纯.中密度纤维板导热系数的测定.林产工业.1997(02)：29～32
[19]蔡纪华.谈人造板热压机控制系统的进展.森林工程,2002,18(1)：9～10
[20]吴娟.大片刨花板热压过程中板坯内部环境变化研究.北京林业大学,2005：36～39
[21]刘恩永.刨花板与纤维板生产技术.中国林业出版社,2007：78～82
[22]徐咏兰,张贵麟.工艺因素对中密度纤维板(MDF)板性的影响.林业科技开发,1997(4)：42～44
[23]谢力生.人造板蒸汽喷射热压法.木材加工机械,2000,5(1)：14～17
[24]陈铭奎.中密度纤维板生产工艺之探讨.木材工业,1993,7(2)：21～30
[25]陆仁书.刨花板制造学.北京：中国林业出版社,1994：102～110
[26]Kamke.FA, M.P.Wolcot, Blacksburg. Fundamentals of flake board manufacture:wood-moisture elation ships. Wood Science and Technology,1991:132～136
[27]Bolton, A.J., P.E.HumPhery. The hot-pressing of dry-formed wood-based composite. Identifying the primary physical process and the nature of their interaction, 1988,42(6):403～406
[28]John.F.shal著.肖亦华,滕通镰,郭焰明译.木材传热传质过程.北京：中国林业出版社,1989：165～180
[29]谢力生,喻云水,曹建文等.常规热压法无胶干法纤维板热压传热研究.林产工业,2003,30(1)：26～28
[30]保昆雁.中密度纤维板热压工艺的研究.中国博士学位论文全文数据库,2008：40～44
[31]余静江,周春晖.过程控制中的软测量技术.控制理论与应用,1996,4(2)：137～142
[32]张旭东,李运泽.基于BP神经网络的纳卫星轨道温度预测.北京航空航天大学学报,2008,34(12)：1423～1427
[33]Huang G-B, Zhu Q-Y, Siew C-K. Extreme Learning Machine:a New Learning Scheme of Feedforward Neural Networks. Proceedings of the International Joint Conference on Neural Networks,2004:985～990
[34]朱大奇.人工神经网络研究现状及其展望.江南大学学报.2004,3(1)：103-109
[35]王安娜,田慧欣,姜周华等.基于信息融合算法的LF炉钢水温度预测.钢铁研究学报,2005,17(6)：71
[36]Robert Claude, Arreto Charles-Daniel, Azerad Jean, Gaudy Jean-Francois. Bibliometric Overview of the Utilization of Artificial Neural Networks in Medicine and Biology. Scientometrics,2004,59(1):117～130
[37]Huang G-B, Zhu Q-Y, Siew C-K. Real-Lime Learning Capability of Neural Networks. Singapore:School of lectrical and Electronic Engineering, Nanyang Technological University,2003.78～92
[38]Huang G-B, Zhu Q-Y, Siew C-K. Extreme learning machine:Theory and applications Neurocomputing,2006,70(1-3):489～501
[39]Q-Y Zhu, A K Qin, P N Suganthan, etal. Evolutionary Extreme Learning Machine. Pattern Recognition,2005,38(2):1759～1763
[40]高大文,王鹏,蔡臻超.人工神经网络中隐含层节点与训练次数的优化.哈尔滨工业大学学报,2003,(2)：207-209
[41]郑开镇.浅析连续平压法生产MDF热压工艺的控制.木材林业,1999,13(6)：34～36
[42]吴振顺.液压控制系统.北京：高等教育出版社,2008：57～62
[43]高翔宇,苏东海.比例变量泵控制马达系统的建模与仿真.流体传动与控制,2009,(3)：10～12
[44]张利平.液压控制系统及设计.北京：化学工业出版社,2006：133～143
[45]Jun Cao, Xiumei Zhang, Yaqiu Liu. Fuzzy Controller of Medium Density Fiberboard Hot-pressing. Conference on Industrial Electronics and Applications,
[46]2007:2045～2048
[47]李国勇.智能控制及其Matlab实现.北京：电子工业出版社,2005：252～256
[48]Harinath. Design and Tuning of Fuzzy PID Controllers for multivariable Process Systems. Memorial University of Newfoundland,2007:321～327
[49]Z.Li, D.Y.Xiao, S.Z.He. Self Adjusting PID Controller based on Fuzzy Inferences. Control Theory and Applications,2010,14(3):23～27
[50]H.Q.Zhao, Glue Dosing System Flakeboard based on Fuzzy Self-tuning PID Control. Journal of Harbin University of Commerce,2007,23(3):325-328
[51]高宪文,赵亚平.模糊自适应PID在焦炉控制中的仿真.东北大学学报,2006,27(10)：1067～1070
[52]Wang. L. X. Fuzzy Systems are Universal Approximators. Fuzzy Systems,1992: 1163～1170
[53]Cupec, Robert. Self-tuning controller based on process fuzzy model. Computational Intelligence and Applications,1999:187～192
[54]王野牧,王洁,陈先惠.液压伺服闭环控制系统的SIMULINK仿真实现.沈阳工业大学学报,2000(5)：370～372
[55]余兵,彭佑多,刘繁茂.模糊控制及其在液压伺服系统中的应用.液压与气动,2006(10)：56～64
[2]莫建虎,李落星等.采用BP神经网络和挤压形状因子预测挤压力与挤压出口温度.矿冶工程,2009,29(2)：90～94
[3]王志春,李忠飞.基于BP网络的钢坯表面温度预测模型的设计.冶金自动化,2009,23(2)：56～60
[4]崔宝侠,周硕.基于改进Elman神经网络的石膏纤维板厚度控制建模.沈阳工业大学学,2008,30(2)：98～120
[5]冯晓华,刘鹏,李静,徐美玲.一种改进的基于神经网络的板形缺陷识别方法.重型机械,2009(6)：77～81
[6]诸静.模糊控制理论与系统原理.北京：机械工业出版社,2005：37-65
[7]许建,肖维荣.液压位置伺服系统的模糊PID控制研究.自动化信息,2010,(6)：45～49
[8]孙延明,曹军,刘亚秋.模糊PID控制在中密度纤维板施胶中的应用研究.木材加工机械,2007(1)：82～89
[9]王东林.中密度纤维板热压智能控制系统的研究.北京林业大学,2008：10～47
[10]王补宣.多孔介质中的对流传热传质.西安交通大学学报,1994,28(5)：51～58
[11]杜朝刚,陈天全,常建民.刨花板热压过程中的传热传质研究现状和展望.林产工业,2004,31(5)：10～14
[12]Kayihan, F.A.Johnson. Heat and Moisture Movement in Wood Compolite Material during the Pressing Operation-A Simplified Model. Numerical Methods in Heat Transfer, 1983(2):511-531
[13]Harless, T.E.G., E.G. Wagner, etal. A Model to Predict the Density Profile of Parti-cleboard. Wood Fiber Sci,1987,19(1):81～92
[14]Humphrey, P.E, H.Thoemen. The Continuous Pressing of Wood-based Panel:An Analytical Simulation Model.5th Annual Pacific Rim Bio-Based Composites Symposium, 2000:123～132
[15]刘应安,蔡智勇.刨花板热压过程中热传导的几个问题.林业科技开发,1992(4)：43～45
[16]张帝树等.刨花板板坯含水率及其梯度对热传递速度的影响.林产工业,1993,20(4)：9-11
[17]刘正添,王洁瑛,于辉.影响刨花板热压传热过程因素的研究.北京林业大学学报,1995,17(2)：64～71
[18]黎爱纯.中密度纤维板导热系数的测定.林产工业.1997(02)：29～32
[19]蔡纪华.谈人造板热压机控制系统的进展.森林工程,2002,18(1)：9～10
[20]吴娟.大片刨花板热压过程中板坯内部环境变化研究.北京林业大学,2005：36～39
[21]刘恩永.刨花板与纤维板生产技术.中国林业出版社,2007：78～82
[22]徐咏兰,张贵麟.工艺因素对中密度纤维板(MDF)板性的影响.林业科技开发,1997(4)：42～44
[23]谢力生.人造板蒸汽喷射热压法.木材加工机械,2000,5(1)：14～17
[24]陈铭奎.中密度纤维板生产工艺之探讨.木材工业,1993,7(2)：21～30
[25]陆仁书.刨花板制造学.北京：中国林业出版社,1994：102～110
[26]Kamke.FA, M.P.Wolcot, Blacksburg. Fundamentals of flake board manufacture:wood-moisture elation ships. Wood Science and Technology,1991:132～136
[27]Bolton, A.J., P.E.HumPhery. The hot-pressing of dry-formed wood-based composite. Identifying the primary physical process and the nature of their interaction, 1988,42(6):403～406
[28]John.F.shal著.肖亦华,滕通镰,郭焰明译.木材传热传质过程.北京：中国林业出版社,1989：165～180
[29]谢力生,喻云水,曹建文等.常规热压法无胶干法纤维板热压传热研究.林产工业,2003,30(1)：26～28
[30]保昆雁.中密度纤维板热压工艺的研究.中国博士学位论文全文数据库,2008：40～44
[31]余静江,周春晖.过程控制中的软测量技术.控制理论与应用,1996,4(2)：137～142
[32]张旭东,李运泽.基于BP神经网络的纳卫星轨道温度预测.北京航空航天大学学报,2008,34(12)：1423～1427
[33]Huang G-B, Zhu Q-Y, Siew C-K. Extreme Learning Machine:a New Learning Scheme of Feedforward Neural Networks. Proceedings of the International Joint Conference on Neural Networks,2004:985～990
[34]朱大奇.人工神经网络研究现状及其展望.江南大学学报.2004,3(1)：103-109
[35]王安娜,田慧欣,姜周华等.基于信息融合算法的LF炉钢水温度预测.钢铁研究学报,2005,17(6)：71
[36]Robert Claude, Arreto Charles-Daniel, Azerad Jean, Gaudy Jean-Francois. Bibliometric Overview of the Utilization of Artificial Neural Networks in Medicine and Biology. Scientometrics,2004,59(1):117～130
[37]Huang G-B, Zhu Q-Y, Siew C-K. Real-Lime Learning Capability of Neural Networks. Singapore:School of lectrical and Electronic Engineering, Nanyang Technological University,2003.78～92
[38]Huang G-B, Zhu Q-Y, Siew C-K. Extreme learning machine:Theory and applications Neurocomputing,2006,70(1-3):489～501
[39]Q-Y Zhu, A K Qin, P N Suganthan, etal. Evolutionary Extreme Learning Machine. Pattern Recognition,2005,38(2):1759～1763
[40]高大文,王鹏,蔡臻超.人工神经网络中隐含层节点与训练次数的优化.哈尔滨工业大学学报,2003,(2)：207-209
[41]郑开镇.浅析连续平压法生产MDF热压工艺的控制.木材林业,1999,13(6)：34～36
[42]吴振顺.液压控制系统.北京：高等教育出版社,2008：57～62
[43]高翔宇,苏东海.比例变量泵控制马达系统的建模与仿真.流体传动与控制,2009,(3)：10～12
[44]张利平.液压控制系统及设计.北京：化学工业出版社,2006：133～143
[45]Jun Cao, Xiumei Zhang, Yaqiu Liu. Fuzzy Controller of Medium Density Fiberboard Hot-pressing. Conference on Industrial Electronics and Applications,
[46]2007:2045～2048
[47]李国勇.智能控制及其Matlab实现.北京：电子工业出版社,2005：252～256
[48]Harinath. Design and Tuning of Fuzzy PID Controllers for multivariable Process Systems. Memorial University of Newfoundland,2007:321～327
[49]Z.Li, D.Y.Xiao, S.Z.He. Self Adjusting PID Controller based on Fuzzy Inferences. Control Theory and Applications,2010,14(3):23～27
[50]H.Q.Zhao, Glue Dosing System Flakeboard based on Fuzzy Self-tuning PID Control. Journal of Harbin University of Commerce,2007,23(3):325-328
[51]高宪文,赵亚平.模糊自适应PID在焦炉控制中的仿真.东北大学学报,2006,27(10)：1067～1070
[52]Wang. L. X. Fuzzy Systems are Universal Approximators. Fuzzy Systems,1992: 1163～1170
[53]Cupec, Robert. Self-tuning controller based on process fuzzy model. Computational Intelligence and Applications,1999:187～192
[54]王野牧,王洁,陈先惠.液压伺服闭环控制系统的SIMULINK仿真实现.沈阳工业大学学报,2000(5)：370～372
[55]余兵,彭佑多,刘繁茂.模糊控制及其在液压伺服系统中的应用.液压与气动,2006(10)：56～64