基于神经网络的平行折线卷筒智能设计与仿真研究
|
摘要
卷筒作为起升机构的关键部件被广泛地应用于各种起重机之中。本研究是根据传统螺旋绳槽卷筒不能满足大扬程、大起重量的起重机卷筒多层卷绕的实际需要，且卷筒产品的设计、开发周期长而提出的。在全面地了解国内外卷筒应用现状，掌握了智能设计研究现状的基础上，论文深入研究了平行折线卷筒的结构特点及其设计计算，对基于BP神经网络的平行折线卷筒智能设计和使用虚拟现实语言OpenGL对卷筒计算机仿真进行了深入地研究。
首先根据传统螺旋绳槽卷筒在多层卷绕时，容易乱绳和绳间磨损严重的缺陷，本文提出了一种在利巴斯卷筒基础上改进的新型卷筒——平行折线卷筒。根据起重机的不同的设计参数，用一般方法对平行折线卷筒的结构参数进行了设计计算，并将此计算结果作为平行折线卷筒结构参数智能设计的神经网络的学习训练样本和检验样本，以验证神经网络智能设计正确性和有效性。
其次，本文重点分析了BP神经网络理论，对改进BP神经网络算法进行了探讨与分析，论述了样本数据的预处理方法、神经网络各项结构参数的确定方法和原则，比较和分析了不同参数下的网络输出结果，最后确定出最佳的神经网络结构参数。首先用计算好的卷筒设计学习样本对改进BP神经网络进行学习和训练，形成良好的网络结构，然后利用该网络对卷筒设计的检验样本进行检验，达到智能化要求。
第三，本文重点研究了虚拟现实语言OpenGL及其在平行折线卷筒虚拟建模与仿真中的应用。研究了OpenGL的工作结构、流程和编程步骤以及建模方法和特殊效果、动画原理，结合计算机图形学和3DS理论，着重研究了平行折线卷筒等物体的几何建模和光照、材质等特殊效果，实现了对折线卷筒的虚拟建模与计算机仿真。
最后，本文利用C++开发出了折线卷筒智能设计系统，并用Visual C++对其进行了计算机仿真。
The reel is extensively applied among the various hoists as the critical parts that rise to rise the organization. This research is put forward because the tradition spiral rope slot reel can not satisfy many layers of real needs winding up in the big lift and heavy load hoist, and the design and development of reel product are grown to put forward period. Completely understanding the reel application present situation in home and abroad, and grasped on the foundation of intelligent design studies present situation, thesis was thoroughly studied of equal rank the structure characteristic of broken line reel and his design and was counted, and has been in progress thoroughly studying with using invented reality language OpenGL to reel computer simulation to parallel the based on BP's nerve cross-domain communication broken line reel intellect design.
When firstly many layers winding up according to the tradition spiral rope slot reel, wearing and tearing the serious defect between easily indiscriminate rope and the rope, one kind of follow-on new type reel - parallel broken line reel on the Li Basi reel foundation was put forward in this text, article, etc.. According to the design parameter of the difference of hoist, the structure parameter with the general method to the parallel broken line reel has been carried on design calculation, and counts this the result training sample and the inspection sample as the study of the nerve cross-domain communication of intelligent design of parallel broken line reel structure parameter, in order to verify nerve network intelligence design correct nature and effectiveness.
Next BP's nerve network theory was stressly analysed in this text, article, etc., and being in progress inquiring into with analysing to improving BP's nerve cross-domain communication algorithm, and discussed definite method and the principle of the pretreatment method of sample data and nerve all structure parameters of cross-domain communication, and compared and analyses the cross-domain communication under the different parameters sending out the result, and finally defining the optimum nerve network architecture parameter.
Firstly with counts the good reel design is learned the sample and is learned
and trains to improving BP's nerve cross-domain communication, and forms the good network architecture, then uses this cross-domain communication to test to the inspection sample of reel design, and achieves the intelligentization requirement.
Third, the application in invented building pattern and simulation of invented reality language OpenGL and his at the parallel broken line reel was still stressly studied in this text, article, etc.. Studied the work of OpenGL structure, distance travelled by a stream of water and compiled the range step as well as built the pattern method and special effect, moved the drawing principle, and combine computer graphics and 3DS's theory, and stresses to study how much on body such as parallel broken line reel etc building the pattern, as well as illumination and material nature etc special effect, to realize to the broken line reel invented build pattern and computer simulation.
Finally, this article uses C++ to develop out broken line reel intellect and designs the system, and has carried on computer simulation with Visual C++ to his.
首先根据传统螺旋绳槽卷筒在多层卷绕时，容易乱绳和绳间磨损严重的缺陷，本文提出了一种在利巴斯卷筒基础上改进的新型卷筒——平行折线卷筒。根据起重机的不同的设计参数，用一般方法对平行折线卷筒的结构参数进行了设计计算，并将此计算结果作为平行折线卷筒结构参数智能设计的神经网络的学习训练样本和检验样本，以验证神经网络智能设计正确性和有效性。
其次，本文重点分析了BP神经网络理论，对改进BP神经网络算法进行了探讨与分析，论述了样本数据的预处理方法、神经网络各项结构参数的确定方法和原则，比较和分析了不同参数下的网络输出结果，最后确定出最佳的神经网络结构参数。首先用计算好的卷筒设计学习样本对改进BP神经网络进行学习和训练，形成良好的网络结构，然后利用该网络对卷筒设计的检验样本进行检验，达到智能化要求。
第三，本文重点研究了虚拟现实语言OpenGL及其在平行折线卷筒虚拟建模与仿真中的应用。研究了OpenGL的工作结构、流程和编程步骤以及建模方法和特殊效果、动画原理，结合计算机图形学和3DS理论，着重研究了平行折线卷筒等物体的几何建模和光照、材质等特殊效果，实现了对折线卷筒的虚拟建模与计算机仿真。
最后，本文利用C++开发出了折线卷筒智能设计系统，并用Visual C++对其进行了计算机仿真。
The reel is extensively applied among the various hoists as the critical parts that rise to rise the organization. This research is put forward because the tradition spiral rope slot reel can not satisfy many layers of real needs winding up in the big lift and heavy load hoist, and the design and development of reel product are grown to put forward period. Completely understanding the reel application present situation in home and abroad, and grasped on the foundation of intelligent design studies present situation, thesis was thoroughly studied of equal rank the structure characteristic of broken line reel and his design and was counted, and has been in progress thoroughly studying with using invented reality language OpenGL to reel computer simulation to parallel the based on BP's nerve cross-domain communication broken line reel intellect design.
When firstly many layers winding up according to the tradition spiral rope slot reel, wearing and tearing the serious defect between easily indiscriminate rope and the rope, one kind of follow-on new type reel - parallel broken line reel on the Li Basi reel foundation was put forward in this text, article, etc.. According to the design parameter of the difference of hoist, the structure parameter with the general method to the parallel broken line reel has been carried on design calculation, and counts this the result training sample and the inspection sample as the study of the nerve cross-domain communication of intelligent design of parallel broken line reel structure parameter, in order to verify nerve network intelligence design correct nature and effectiveness.
Next BP's nerve network theory was stressly analysed in this text, article, etc., and being in progress inquiring into with analysing to improving BP's nerve cross-domain communication algorithm, and discussed definite method and the principle of the pretreatment method of sample data and nerve all structure parameters of cross-domain communication, and compared and analyses the cross-domain communication under the different parameters sending out the result, and finally defining the optimum nerve network architecture parameter.
Firstly with counts the good reel design is learned the sample and is learned
and trains to improving BP's nerve cross-domain communication, and forms the good network architecture, then uses this cross-domain communication to test to the inspection sample of reel design, and achieves the intelligentization requirement.
Third, the application in invented building pattern and simulation of invented reality language OpenGL and his at the parallel broken line reel was still stressly studied in this text, article, etc.. Studied the work of OpenGL structure, distance travelled by a stream of water and compiled the range step as well as built the pattern method and special effect, moved the drawing principle, and combine computer graphics and 3DS's theory, and stresses to study how much on body such as parallel broken line reel etc building the pattern, as well as illumination and material nature etc special effect, to realize to the broken line reel invented build pattern and computer simulation.
Finally, this article uses C++ to develop out broken line reel intellect and designs the system, and has carried on computer simulation with Visual C++ to his.
引文
[1]吴慧中，陈定方．机械设计专家系统研究与实践．北京：中国铁道出版社．1994
[2]周济，查建中．智能设计．北京：高等教育出版社．1998
[3]史济建．专家系统与实现技术等著．杭州：浙江大学出版社．1999
[4]戴葵．神经网络实现技术．长沙：国防科技大学出版社．2001
[5]田盛丰．人工智能与知识工程．北京：中国铁道出版社．1999
[6]卢丽丹．OpenGL程序设计．北京：清华大学出版社．2000
[7]蔡士杰译．计算机图形学．北京：电子工业出版社．2000
[8]王春森．程序员教程．北京：清华大学出版社．2001
[9]郝永涛．基于模式化知识和人工神经网络的智能设计系统框架研究：[硕士学位论文]．武汉：华中理工大学．1999
[10]刘增良．模糊逻辑与神经网络——理论研究与探索．北京：北京航空航天大学出版社．2001
[11]张秀山．虚拟现实技术与编程技巧．北京：国防科技大学出版社．1999
[12]蔡灿．折线式转筒的设计与制造．机械设计．2001．4(3)
[13]毕晓亮．快速成型软件中的OpenGL技术．计算机应用研究．2001．7(12)
[14]李素有．基于OpenGL的机械原理实验三维图形仿真系统．计算机应用研究．2001．2(22)
[15][英国]迈克尔·里德帕思．虚拟现实技术．北京：译林出版社．1998
[16]许少华．模糊神经网络学习样本的选取与网络扩展能力研究：[博士学位论文]．武汉：华中理工大学，2000
[17]王坚，孙宇浩．虚拟现实科学与技术．浙江：浙江科学技术出版社．2000
[18]郭维斌．双阶螺旋折线绳槽卷筒的设计．工程机械．2000．12
[19]金延赞．计算机图形学．杭州：浙江大学出版社．1988
[20]http://www.opengl.org
[21]孙家广．计算机辅助几何造型技术．北京：清华大学出版社．1990
[22]张福炎．程序员高级程序员程序设计第二版．北京：清华大学出版社．2000
[23]官宜全．Visual C++6.0编程实例详解．北京：电子工业出版社．1999
[24]Told S. Latham W. Evolution Art and Computers. Acadmic Press. 1992
[25]Bentley p J. Wakefild J. p. Generic Evolutionary Design. In Chawdhry P K Roy RPant. R. k eds. Soft Computeing in engineering Design and. Spring er Verlag. Part 6. 289
[26]De Garis H. An Artificial Brain New Computeing v12. 2Springer-Verlag. 1994
[27]Bentley P. J. Evolutionary Design By Computers. Morgan Kaufmann Pub. 1999
[28]Patrick Henry Winston. Artificial Intelligenec Addion Wesley. Publisahing Company[J]. 1999. 159-200
[29]David D Zhang. Automated Biometrics Technologies and Systems[M]. Kluwer Academic Publishers. 2000. 111-134
[30]Tanaka H. Fuzzy Data Analysis by Possibilistic Linear Mo-dels[J]。. Fuzzy Sets and Systems. 1987, 2 4
[31]Zolotw, A. A, et al. Method Ensuring Reliability of Flying Vehicles[A]. PTC R-USAST, Xian. 1994:66～69
[32]Tandka H, et al. Identification of Possibilistic linear Systems by Quadratic Membership Function of Fuzzy Parameters[J]. Fuzzy Sets and Systems 1991, 41: 145～160
[33]Liao T W. A Fuzzy Multicriteria Decision-Making Method for Material Selection[J]. J. of Manufacturing Systems. 1996, 15(1):1～12
[34]A. Bahrami, M. Lynch, and C. H. Dagli. Intelligent design re-trieval and packaging system application of neural networks indesign and manifacturing INT。 PPODRES. 1996, 33
[35]Debevec Paul E, Jaylar Camillo J, Malik Jitendra. Modeling and rendering architecture from photograpgs:A hybrid geometry-and-image-based approach. In:Computer Graphics(SIGGRAPH96). New york. 1997:11～20
[36]Stephen R Ellis. What are virtual environments. IEEEComputer Graphics &Applications. 1994, 14(1)
[37]李颖，薛海斌，朱伯立．OpenGL技术应用实例精粹．北京：国防工业出版社．2001
[38]张际先，宓霞．神经网络及其在工程中的应用．北京：机械工业出版社．1996
[39]王文成．神经网络及其在汽车工程中的应用．北京：北京理工大学出版社．1998
[40]徐晓刚．Visual C++6.0入门与提高．北京：清华大学出版社．1999
[41]孙家广。计算机辅助设计技术基础第二版．北京：清华大学出版社．1996
[42]潘志赓．Windows环境下图形图象程序设计．北京：清华大学出版社．1997
[43]李于剑．VisualC++6.0实践与提高图形图象编程篇．北京：中国铁道出版社．2001
[44]钱能．C++程序设计教程．北京：清华大学出版社．1999
[45]侯俊杰著．深入浅出MFC．武汉：华中科技大学出版社．2001
[46]陈昆．基于OpenGL的起升机构仿真研究：[硕士学位论文]．武汉：武汉科技大学，2002
[47]张质文，虞和谦．起重机设计手册．北京：中国铁道出版社．1998
[48]郭永．XTGDES齿轮设计专家系统．制造业自动化．2001．6(4)
[49]王士刚．液压系统可视化动态建模技术及其软件实现方法研究：[博士学位论文]．大连：大连理工大学
[50]黄心渊．虚拟现实技术与应用．北京：科学技术出版社．1999
[2]周济，查建中．智能设计．北京：高等教育出版社．1998
[3]史济建．专家系统与实现技术等著．杭州：浙江大学出版社．1999
[4]戴葵．神经网络实现技术．长沙：国防科技大学出版社．2001
[5]田盛丰．人工智能与知识工程．北京：中国铁道出版社．1999
[6]卢丽丹．OpenGL程序设计．北京：清华大学出版社．2000
[7]蔡士杰译．计算机图形学．北京：电子工业出版社．2000
[8]王春森．程序员教程．北京：清华大学出版社．2001
[9]郝永涛．基于模式化知识和人工神经网络的智能设计系统框架研究：[硕士学位论文]．武汉：华中理工大学．1999
[10]刘增良．模糊逻辑与神经网络——理论研究与探索．北京：北京航空航天大学出版社．2001
[11]张秀山．虚拟现实技术与编程技巧．北京：国防科技大学出版社．1999
[12]蔡灿．折线式转筒的设计与制造．机械设计．2001．4(3)
[13]毕晓亮．快速成型软件中的OpenGL技术．计算机应用研究．2001．7(12)
[14]李素有．基于OpenGL的机械原理实验三维图形仿真系统．计算机应用研究．2001．2(22)
[15][英国]迈克尔·里德帕思．虚拟现实技术．北京：译林出版社．1998
[16]许少华．模糊神经网络学习样本的选取与网络扩展能力研究：[博士学位论文]．武汉：华中理工大学，2000
[17]王坚，孙宇浩．虚拟现实科学与技术．浙江：浙江科学技术出版社．2000
[18]郭维斌．双阶螺旋折线绳槽卷筒的设计．工程机械．2000．12
[19]金延赞．计算机图形学．杭州：浙江大学出版社．1988
[20]http://www.opengl.org
[21]孙家广．计算机辅助几何造型技术．北京：清华大学出版社．1990
[22]张福炎．程序员高级程序员程序设计第二版．北京：清华大学出版社．2000
[23]官宜全．Visual C++6.0编程实例详解．北京：电子工业出版社．1999
[24]Told S. Latham W. Evolution Art and Computers. Acadmic Press. 1992
[25]Bentley p J. Wakefild J. p. Generic Evolutionary Design. In Chawdhry P K Roy RPant. R. k eds. Soft Computeing in engineering Design and. Spring er Verlag. Part 6. 289
[26]De Garis H. An Artificial Brain New Computeing v12. 2Springer-Verlag. 1994
[27]Bentley P. J. Evolutionary Design By Computers. Morgan Kaufmann Pub. 1999
[28]Patrick Henry Winston. Artificial Intelligenec Addion Wesley. Publisahing Company[J]. 1999. 159-200
[29]David D Zhang. Automated Biometrics Technologies and Systems[M]. Kluwer Academic Publishers. 2000. 111-134
[30]Tanaka H. Fuzzy Data Analysis by Possibilistic Linear Mo-dels[J]。. Fuzzy Sets and Systems. 1987, 2 4
[31]Zolotw, A. A, et al. Method Ensuring Reliability of Flying Vehicles[A]. PTC R-USAST, Xian. 1994:66～69
[32]Tandka H, et al. Identification of Possibilistic linear Systems by Quadratic Membership Function of Fuzzy Parameters[J]. Fuzzy Sets and Systems 1991, 41: 145～160
[33]Liao T W. A Fuzzy Multicriteria Decision-Making Method for Material Selection[J]. J. of Manufacturing Systems. 1996, 15(1):1～12
[34]A. Bahrami, M. Lynch, and C. H. Dagli. Intelligent design re-trieval and packaging system application of neural networks indesign and manifacturing INT。 PPODRES. 1996, 33
[35]Debevec Paul E, Jaylar Camillo J, Malik Jitendra. Modeling and rendering architecture from photograpgs:A hybrid geometry-and-image-based approach. In:Computer Graphics(SIGGRAPH96). New york. 1997:11～20
[36]Stephen R Ellis. What are virtual environments. IEEEComputer Graphics &Applications. 1994, 14(1)
[37]李颖，薛海斌，朱伯立．OpenGL技术应用实例精粹．北京：国防工业出版社．2001
[38]张际先，宓霞．神经网络及其在工程中的应用．北京：机械工业出版社．1996
[39]王文成．神经网络及其在汽车工程中的应用．北京：北京理工大学出版社．1998
[40]徐晓刚．Visual C++6.0入门与提高．北京：清华大学出版社．1999
[41]孙家广。计算机辅助设计技术基础第二版．北京：清华大学出版社．1996
[42]潘志赓．Windows环境下图形图象程序设计．北京：清华大学出版社．1997
[43]李于剑．VisualC++6.0实践与提高图形图象编程篇．北京：中国铁道出版社．2001
[44]钱能．C++程序设计教程．北京：清华大学出版社．1999
[45]侯俊杰著．深入浅出MFC．武汉：华中科技大学出版社．2001
[46]陈昆．基于OpenGL的起升机构仿真研究：[硕士学位论文]．武汉：武汉科技大学，2002
[47]张质文，虞和谦．起重机设计手册．北京：中国铁道出版社．1998
[48]郭永．XTGDES齿轮设计专家系统．制造业自动化．2001．6(4)
[49]王士刚．液压系统可视化动态建模技术及其软件实现方法研究：[博士学位论文]．大连：大连理工大学
[50]黄心渊．虚拟现实技术与应用．北京：科学技术出版社．1999
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |