摘要
神经网络方法是处理非线性问题的有力工具,但当输入变量较多,输入变量间存在的多重共线性性会使得网络的建模效率下降。偏最小二乘回归方法通过提取对因变量解释性较强的成分,能较好地克服变量间的多重共线性。将两种方法相结合,建立了爆破振动峰值速度的偏最小二乘回归BP神经网络预测模型。利用偏最小二乘法对影响爆破振动的因素进行分析,提取出3个新综合变量,使BP网络的输入层节点数目由9个减少到3个,简化了网络结构,提高了计算速度,增强了网络稳定性。分析结果表明,耦合模型的平均预测误差为7.62%,相较于传统的萨氏公式及标准的BP神经网络模型其预测精度有了明显提高。
The neural network method is a powerful tool to deal with problems of nonlinearity,but when input variables are so many,the multicollinearity among variables can lead to a lower modeling efficiency.The partial leastsquare regression(PLSR) method can extract components with better interpretation to the dependent variables,thus it can overcome the multicollinearity among variables.Here,by combining the two methods,a BP neural network prediction model for peak velocity of blasting vibration based on PLSR was established.The affecting factors on blasting vibration were analyzed by means of PLSR,and three new synthesis variables were extracted.Since the input layer nodes of the BP neural network decreased from nine to three,the network structure was simplified and the netweork became,more efficient and more stable.The results showed that the average prediction error of the combined model is 7.62%,the new model is more accurate than Sadaovsk formula and a normal BP neural network modelbe.
引文
[1]唐海,石永强,李海波,等.基于神经网络的爆破振动速度峰值预报[J].岩石力学与工程学报,2007,26(增1):3533-3539.TANG Hai,SHI Yong-qiang,LI Hai-bo,et al.Prediction ofpeak velocity of blasting vibration based on neural network[J].Chinese Journal of Rock Mechanics and Engineering,2007,26(suppl.1):3533-3539.
[2]林从谋,逄焕东,王其升,等.隧道掘进爆破地震峰值神经网络预报研究[J].岩土力学,2004,25(增):125-126.LIN Cong-mou,PANG Huan-dong,WANG Qi-sheng,et al.Study on neural network prediction of peak amplitude ofblasting ground vibration for tunneling[J].Rock and SoilMechanics,2004,2(suppl):125-126.
[3]杨佑发,崔波.爆破震动速度峰值的预测[J].振动与冲击,2009,28(10):195-198.YANG You-fa,CUI Bo.Prediction for vibration intensity dueto blasting induced ground motions[J].Journal of Vibrationand Shock,2009,28(10):195-198.
[4]陈建宏,刘浪,周智勇,等.基于主成分分析与神经网络的采矿方法优选[J].中南大学学报(自然科学版),2010,41(5):1967-1972.CHEN Jian-hong,LIU Lang,ZHOU Zhi-yong,et al.Optimization of mining methods based on combination ofprincipal component analysis and neural networks[J].Journalof Central South University(Science and Technology),2010,41(5):1967-1972.
[5]方向,陆凡东,高振儒,等.中深孔爆破振动加速度峰值的遗传BP网络预测[J].解放军理工大学学报(自然科学版),2010,11(3):312-315.FANG Xiang,LU Fan-dong,GAO Zhen-ru,et al.Predictionon peak vibration acceleration value of medium-deep-holeblasting using genetic BP network[J].Journal of PLAUniversity of Science and Technology(Natural ScienceEdition),2010,11(3):312-315.
[6]张艺峰,姚道平,谢志招,等.L-M优化算法在爆破振动参数预测中的应用[J].地震学报,2008,30(5):540-544.ZHANG Yi-feng,YAO Dao-ping,XIE Zhi-zhao,et al.Application of the L-M optimization algorithm to predictingblast vibration parameters[J].Acta Seismologica Sinica,2008,30(5):540-544.
[7]史秀志.爆破振动信号时频分析与爆破振动预测研究[D].长沙:中南大学资源与安全工程学院,2007:94-127.
[8]王惠文.偏最小二乘回归方法及其应用[M].北京:国防工业出版社,1999.
[9]宋高阳.偏最小二乘回归的研究[D].杭州:浙江大学理学院,2009:4-9.
[10]葛哲学,孙志强.神经网络理论与MATLAB R2007实现[M].北京:电子工业出版社,2007.
[11]The MathWorks,Inc.Neural Network ToolboxTM User’sGuide(Version 7.0.2)[R].The Math Works,2011.
[12]沈花玉,王兆霞,高成耀,等.BP神经网络隐含层单元数的确定[J].天津理工大学学报,2008,24(5):13-15.SHEN Hua-yu,WANG Zhao-xia,GAO Cheng-yao,et al.Determining the number of BP neural network hidden layerunits[J].Journal of Tianjin University of Technology,2008,24(5):13-15.
[13]言志信,言浬,江平,等.爆破振动峰值速度预报方法探讨[J].振动与冲击,2010,29(5):179-182.YAN Zhi-xin,Yan Li,JIANG Ping,et al.Predictionmethods for blasting-induced ground vibration velocity[J].Journal of Vibration and Shock,2010,29(5):179-182.