基于GASA-SVR的矿井瓦斯涌出量预测研究
|
摘要
针对煤矿工作面瓦斯涌出量的多影响因素、非线性、时变性和不确定性等特点,提出了遗传模拟退火算法(GASA)与回归型支持向量机(SVR)的耦合算法(GASA-SVR)用于瓦斯涌出量预测。利用煤层瓦斯含量、深度、厚度、倾角等12个参数作为主要影响因素,经过归一化处理后作为回归型支持向量机训练和测试样本。采用遗传模拟退火算法寻找最优的惩罚参数和核函数参数,同时引入自适应交叉和变异概念,建立瓦斯涌出量的非线性拟合模型,并利用矿井实测历史数据进行试验,结果表明该预测模型比传统的神经网络模型具有更理想的精度和稳定性,可为煤矿瓦斯爆炸的防治提供可靠的理论依据。
Genetic simulated annealing algorithm( GASA) and regression support vector machine( SVM) coupling algorithm( GASA-SVR) were presented to predict gas emission in the view of the coal mine characteristics such as complicated,time varying and nonlinear etc. 12 parameters,such as gas content,depth,thickness and dip angle of coal seam,were used as the main influencing factors,the training and test samples were trained and tested as regression support vector machine. Genetic simulated annealing algorithm was used to find the optimal penalty parameter and kernel parameter. Simultaneously,the concept of adaptive crossover and mutation is introduced,the nonlinear fitting model of gas emission is established,and the experiment was carried out by using the measured historical data of the mine. The results show that the prediction model has better accuracy and stability than the traditional neural network model,which can provide a reliable theoretical basis for the prevention and control of coal mine gas explosion.
Genetic simulated annealing algorithm( GASA) and regression support vector machine( SVM) coupling algorithm( GASA-SVR) were presented to predict gas emission in the view of the coal mine characteristics such as complicated,time varying and nonlinear etc. 12 parameters,such as gas content,depth,thickness and dip angle of coal seam,were used as the main influencing factors,the training and test samples were trained and tested as regression support vector machine. Genetic simulated annealing algorithm was used to find the optimal penalty parameter and kernel parameter. Simultaneously,the concept of adaptive crossover and mutation is introduced,the nonlinear fitting model of gas emission is established,and the experiment was carried out by using the measured historical data of the mine. The results show that the prediction model has better accuracy and stability than the traditional neural network model,which can provide a reliable theoretical basis for the prevention and control of coal mine gas explosion.
引文
[1]李贤功.中国煤矿重大瓦斯事故致因复杂性机理及其管控研究[D].中国矿业大学,2010.
[2]尹光志,李铭辉,李文璞,等.基于改进BP神经网络的煤体瓦斯渗透率预测模型[J].煤炭学报,2013,38(7):1179-1184.
[3]潘玉民,张晓宇,张全柱,等.基于量子粒子群优化的小波神经网络预测模型[J].信息与控制,2012,41(6):741-746.
[4]付华,王馨蕊,杨本臣,等.基于MPSO-CWLS-SVM的瓦斯涌出量预测[J].传感技术学报,2014(11):1568-1572.
[5]付华,于翔,卢万杰.基于蚁群粒子群混合算法与LS-SVM瓦斯涌出量预测[J].传感技术学报,2016,29(3).
[6]付华,刘汀,张胜强,等.基于SOM-RBF算法的瓦斯涌出量动态预测模型研究[J].传感技术学报,2015(8):1255-1261.
[7]李敬花,余峰,樊付见.基于遗传模拟退火融合算法的船舶分段装配序列优化[J].计算机集成制造系统,2013,19(1):39-45.
[8]樊晓宇,王玉宝,卢国伟等.基于遗传模拟退火算法的压力传感器温度补偿系统[J].传感技术学报,2008,21(9):1532-1535.
[9]Vladimir Cherkassky,Ma Yunqian.Practical Selection of SVM Parameters and Noise Estimation for SVM Regression[J].Neural Networks,2004,17(1):113-126.
[10]Chih-Chung Chang,Chih-Jen Lin.Trainingν-Support Vector Regression:Theory and Algorithms[J].Neural Computation,2002,14(8):1959-1977.
[11]王泉夫.基于WSN的工作面监控及瓦斯浓度预测关键技术研究[D].中国矿业大学,2009.
[12]苏静波.工程结构不确定性区间分析方法及其应用研究[D].南京:河海大学,2006.
[13]袁澎,艾芊,赵媛媛.基于改进的遗传-模拟退火算法和误差度分析原理的PMU多目标优化配置[J].中国电机工程学报,2014,34(13):2178-2187.
[14]罗景峰.基于均匀设计的冷却进度表参数设定[J].科学技术与工程,2009,9(9):2468:2470.
[15]张昊,陶然,李志勇,等.基于自适应模拟退火遗传算法的特征选择方法[J].兵工学报,2009,30(1):81-85.
[16]付华,丰盛成,刘晶,等.基于DE-EDA-SVM的瓦斯浓度预测建模仿真研究[J].传感技术学报,2016,29(2):285-289.
[17]李瑾,刘金朋,王建军.采用支持向量机和模拟退火算法的中长期负荷预测方法[J].中国电机工程学报,2011,31(16):63-66.
[18]付华,姜伟,单欣欣.基于耦合算法的煤矿瓦斯涌出量预测模型研究[J].煤炭学报,2012,37(4):654-658.
[2]尹光志,李铭辉,李文璞,等.基于改进BP神经网络的煤体瓦斯渗透率预测模型[J].煤炭学报,2013,38(7):1179-1184.
[3]潘玉民,张晓宇,张全柱,等.基于量子粒子群优化的小波神经网络预测模型[J].信息与控制,2012,41(6):741-746.
[4]付华,王馨蕊,杨本臣,等.基于MPSO-CWLS-SVM的瓦斯涌出量预测[J].传感技术学报,2014(11):1568-1572.
[5]付华,于翔,卢万杰.基于蚁群粒子群混合算法与LS-SVM瓦斯涌出量预测[J].传感技术学报,2016,29(3).
[6]付华,刘汀,张胜强,等.基于SOM-RBF算法的瓦斯涌出量动态预测模型研究[J].传感技术学报,2015(8):1255-1261.
[7]李敬花,余峰,樊付见.基于遗传模拟退火融合算法的船舶分段装配序列优化[J].计算机集成制造系统,2013,19(1):39-45.
[8]樊晓宇,王玉宝,卢国伟等.基于遗传模拟退火算法的压力传感器温度补偿系统[J].传感技术学报,2008,21(9):1532-1535.
[9]Vladimir Cherkassky,Ma Yunqian.Practical Selection of SVM Parameters and Noise Estimation for SVM Regression[J].Neural Networks,2004,17(1):113-126.
[10]Chih-Chung Chang,Chih-Jen Lin.Trainingν-Support Vector Regression:Theory and Algorithms[J].Neural Computation,2002,14(8):1959-1977.
[11]王泉夫.基于WSN的工作面监控及瓦斯浓度预测关键技术研究[D].中国矿业大学,2009.
[12]苏静波.工程结构不确定性区间分析方法及其应用研究[D].南京:河海大学,2006.
[13]袁澎,艾芊,赵媛媛.基于改进的遗传-模拟退火算法和误差度分析原理的PMU多目标优化配置[J].中国电机工程学报,2014,34(13):2178-2187.
[14]罗景峰.基于均匀设计的冷却进度表参数设定[J].科学技术与工程,2009,9(9):2468:2470.
[15]张昊,陶然,李志勇,等.基于自适应模拟退火遗传算法的特征选择方法[J].兵工学报,2009,30(1):81-85.
[16]付华,丰盛成,刘晶,等.基于DE-EDA-SVM的瓦斯浓度预测建模仿真研究[J].传感技术学报,2016,29(2):285-289.
[17]李瑾,刘金朋,王建军.采用支持向量机和模拟退火算法的中长期负荷预测方法[J].中国电机工程学报,2011,31(16):63-66.
[18]付华,姜伟,单欣欣.基于耦合算法的煤矿瓦斯涌出量预测模型研究[J].煤炭学报,2012,37(4):654-658.