尾砂分形级配与胶结强度的知识库研究
|
摘要
采用分形理论研究了尾砂材料的级配。分析了国内外大量矿山的尾砂材料分形级配与强度试验数据,用神经网络建立了尾砂胶结强度与水泥含量、浓度、孔隙分形维数及分形维数相关率的知识库模型。考虑神经网络在训练大规模样品时易陷入局部极小,用梯度下降法与混沌优化方法相结合,使神经网络实现快速训练的同时,避免陷入局部极小。研究结果表明:尾砂孔隙分形维数减小,尾砂胶结强度增高;分维形数相关率越好,尾砂胶结强度越高。孔隙分形维数和分形维数相关率反映了尾砂粒度分布的整体信息,可用来判断尾砂级配的合理性。应用知识库模型可以根据尾砂的级配特性,预测不同水泥含量、不同浓度下的尾砂胶结充填体强度,指导矿山充填设计。
Characteristics of tailings gradation are studied by using fractal theory. Experimental data about the fractal gradations of tailings material and their cementing strength from a lot of mines are analyzed. Neural network is used to establish the model of knowledge bank which embodies the relations between strengths of cemented tailings and content of cement,consistence,fractal dimension of porosity and correlation coefficient of fractal dimension. Combining grading method with chaotic optimization,the neural network model achieves rapid training and avoids local minimum when there are a lot of samples to be trained. Research results show that the strengths of cemented tailings increase with the decrease of fractal dimension of porosity and with the increase of the correlation coefficients of fractal dimension. Because fractal dimension of porosity and its correlation coefficient embody the global distribution information of tailing granule,they can be used as a standard of rationality of gradation. According to gradation of tailings,the knowledge bank model can predict the strengths of cemented tailings in different contents of cement or consistence,and guide filling design in mine.
Characteristics of tailings gradation are studied by using fractal theory. Experimental data about the fractal gradations of tailings material and their cementing strength from a lot of mines are analyzed. Neural network is used to establish the model of knowledge bank which embodies the relations between strengths of cemented tailings and content of cement,consistence,fractal dimension of porosity and correlation coefficient of fractal dimension. Combining grading method with chaotic optimization,the neural network model achieves rapid training and avoids local minimum when there are a lot of samples to be trained. Research results show that the strengths of cemented tailings increase with the decrease of fractal dimension of porosity and with the increase of the correlation coefficients of fractal dimension. Because fractal dimension of porosity and its correlation coefficient embody the global distribution information of tailing granule,they can be used as a standard of rationality of gradation. According to gradation of tailings,the knowledge bank model can predict the strengths of cemented tailings in different contents of cement or consistence,and guide filling design in mine.
引文
[1]Ellis A V.A review of some environment issues affecting marine mining[J].Marine Georesources and Geotechnology,2001,19(1):51–63.
[2]卢平.制约胶结充填采矿法发展的若干充填体力学问题[J].黄金,1994,15(7):18–22.(Lu Ping.A few points about the mechanics of back filling which restrain the development of cementing back-filling method[J].Gold,1994,15(7):18–22.(in Chinese))
[3]Jung S J,Biswas K.Review of current high density paste fill and its technology[J].Mineral Resources Engineering,2002,11(2):165–182.
[4]刘同有.充填采矿技术与应用[M].北京:冶金工业出版社,2001.(Liu Tongyou.Technology of Backfill Mining and Its Application[M].Beijing:Metallurgical Industry Press,2001.(in Chinese))
[5]刘亚俊,叶邦彦,夏伟.分形理论及其在材料非线性力学行为研究中的应用[J].材料科学与工程,2001,19(3):104–107.(Liu Yajun,Ye Bangyan,Xia Wei.Fractal theory and its application in the study of mechanics behavior of materials[J].Materials Science and Engineering,2001,19(3):104–107.(in Chinese))
[6]特科特D L.分形与混沌[M].陈颙,郑捷,季颖译.北京:地震出版社,1993.(Turcotte D L.Fractal and Chaos[M].Translated by Chen Yong,Zheng Jie,Ji Ying.Beijing:Earthquake Press,1993.(in Chinese))
[7]徐永富,孙婉莹,吴正根.我国膨胀土的分形结构研究[J].河海大学学报(自然科学版),1997,25(1):17–23.(Xu Yongfu,Sun Wanyin,Wu Zhenggen.On fractal structure of expansive soils in China[J].Journal of Hohai University(Natural Science),1997,25(1):17–23.(in Chinese))
[8]Kobayashi M,Hattori M,Yamazaki H.Multidirectional associative memory with a hidden layer[J].Systems and Computers in Japan,2002,33(6):1–9.
[9]何国光,周坚强.基于前向神经网络的知识获取[J].吉首大学学报(自然科学版),2002,23(6):62–65.(He Guoguang,Zhou Jianqiang,Study of the knowledge acquirement based on feed-forward neural network[J].Journal of Jishou University(Natural Science Edition),2002,23(6):62–65.(in Chinese))
[10]桂卫华,胡志坤,彭小奇.前馈网络的混沌梯度搜索耦合学习算法及应用[J].中南工业大学学报(自然科学版),2002,33(6):629–631.(Gui Weihua,Hu Zhikun,Peng Xiaoqi.A coupled exercise algorithm of forward neural network combined with gradient search and chaotic optimization search[J].Journal of Central South University of Technology(Natural Science),2002,33(6):629–631.(in Chinese))
[11]李兵,蒋慰孙.混沌优化方法及其应用[J].控制理论与应用,1997,14(4):613–615.(Li Bing,Jiang Weisun.Chaos optimization method and its application[J].Control Theory and Applications,1997,14(4):613–615.(in Chinese))
[2]卢平.制约胶结充填采矿法发展的若干充填体力学问题[J].黄金,1994,15(7):18–22.(Lu Ping.A few points about the mechanics of back filling which restrain the development of cementing back-filling method[J].Gold,1994,15(7):18–22.(in Chinese))
[3]Jung S J,Biswas K.Review of current high density paste fill and its technology[J].Mineral Resources Engineering,2002,11(2):165–182.
[4]刘同有.充填采矿技术与应用[M].北京:冶金工业出版社,2001.(Liu Tongyou.Technology of Backfill Mining and Its Application[M].Beijing:Metallurgical Industry Press,2001.(in Chinese))
[5]刘亚俊,叶邦彦,夏伟.分形理论及其在材料非线性力学行为研究中的应用[J].材料科学与工程,2001,19(3):104–107.(Liu Yajun,Ye Bangyan,Xia Wei.Fractal theory and its application in the study of mechanics behavior of materials[J].Materials Science and Engineering,2001,19(3):104–107.(in Chinese))
[6]特科特D L.分形与混沌[M].陈颙,郑捷,季颖译.北京:地震出版社,1993.(Turcotte D L.Fractal and Chaos[M].Translated by Chen Yong,Zheng Jie,Ji Ying.Beijing:Earthquake Press,1993.(in Chinese))
[7]徐永富,孙婉莹,吴正根.我国膨胀土的分形结构研究[J].河海大学学报(自然科学版),1997,25(1):17–23.(Xu Yongfu,Sun Wanyin,Wu Zhenggen.On fractal structure of expansive soils in China[J].Journal of Hohai University(Natural Science),1997,25(1):17–23.(in Chinese))
[8]Kobayashi M,Hattori M,Yamazaki H.Multidirectional associative memory with a hidden layer[J].Systems and Computers in Japan,2002,33(6):1–9.
[9]何国光,周坚强.基于前向神经网络的知识获取[J].吉首大学学报(自然科学版),2002,23(6):62–65.(He Guoguang,Zhou Jianqiang,Study of the knowledge acquirement based on feed-forward neural network[J].Journal of Jishou University(Natural Science Edition),2002,23(6):62–65.(in Chinese))
[10]桂卫华,胡志坤,彭小奇.前馈网络的混沌梯度搜索耦合学习算法及应用[J].中南工业大学学报(自然科学版),2002,33(6):629–631.(Gui Weihua,Hu Zhikun,Peng Xiaoqi.A coupled exercise algorithm of forward neural network combined with gradient search and chaotic optimization search[J].Journal of Central South University of Technology(Natural Science),2002,33(6):629–631.(in Chinese))
[11]李兵,蒋慰孙.混沌优化方法及其应用[J].控制理论与应用,1997,14(4):613–615.(Li Bing,Jiang Weisun.Chaos optimization method and its application[J].Control Theory and Applications,1997,14(4):613–615.(in Chinese))
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心