混合学习法前向网络多属性储层参数预测
|
摘要
应用前向网络描述地震属性和储层参数间的非线性映射关系时,经典的误差反向传播算法存在收敛速度慢,易陷入局部极值等诸多不足。研究了融合粒子群优化算法和误差反向传播算法的混合学习法前向网络多属性储层参数预测技术。粒子群优化算法是一种群体随机搜索演化计算技术,具有较快的收敛速度和较强的全局搜寻能力;误差反向传播算法本质上是梯度下降算法,注重局部搜索。混合学习法为两种学习算法交替执行,首先以粒子群优化算法训练网络,当误差能量在规定的迭代次数内不再发生变化时,采用误差反向传播算法实现局部寻优。理论函数逼近测试和实际储层参数预测实验说明了混合学习法具有学习时间短、求解效率高、可靠性强的优点,具有良好的应用前景。
When applying the feed-forward network to describe the nonline arrelationship between seismic attributes and reservoir parameters,the classical error back propagation(BP)learning algorithm has slow convergence and is easy to fall into local minima,and has many other deficiencies.This paper researches a new reservoir parameter prediction technique by seismic multi-attribute which is based on the integration of particle swarm optimization(PSO)algorithm and BPalgorithm of the feed-forward network.PSO is a group randomsearch evolutionary computation technique,has a faster convergence rate and strong global search ability;BP algorithm is essentially a gradient descent algorithm,focusing on local search.Hybrid algorithm trains network by PSO and BP alternatively.That is,when the learning error energy of PSO in required generation has not changed,using BP for local search disturbances.Function approximation test and practical reservoir parameter prediction experiment illustrate the hybrid learning method has a short training time,high efficiency and reliability,and has good prospect.
When applying the feed-forward network to describe the nonline arrelationship between seismic attributes and reservoir parameters,the classical error back propagation(BP)learning algorithm has slow convergence and is easy to fall into local minima,and has many other deficiencies.This paper researches a new reservoir parameter prediction technique by seismic multi-attribute which is based on the integration of particle swarm optimization(PSO)algorithm and BPalgorithm of the feed-forward network.PSO is a group randomsearch evolutionary computation technique,has a faster convergence rate and strong global search ability;BP algorithm is essentially a gradient descent algorithm,focusing on local search.Hybrid algorithm trains network by PSO and BP alternatively.That is,when the learning error energy of PSO in required generation has not changed,using BP for local search disturbances.Function approximation test and practical reservoir parameter prediction experiment illustrate the hybrid learning method has a short training time,high efficiency and reliability,and has good prospect.
引文
[1]Hampson D P,Schuelke J S,Quirein J A.Use of multi-attribute transforms to predict log properties from seismicdata[J].Geophysics,2001,66(1):220 236.
[2]Liu Zhengping,Liu Jiaqi.Seismic-controlled nonlinearextrapolation of well parameters using neural network[J].Geophysics,1998,63(6):2035 2041.
[3]吴大奎,李亚林,伍志明,等.地震、测井资料联合反演的神经网络算法研究[J].天然气工业,2004,24(3):55 57.
[4]杨辉廷,颜其彬,李敏,等.应用改进的神经网络学习方法预测储层参数[J].天然气工业,2005,25(6):37 39.
[5]冉启全,周雪峰,胡永乐.基于神经网络的多属性火成岩岩性反演技术[J].西南石油学院学报,2006,28(1):5 8,17.
[6]郑庆生,韩大匡.高阶神经网络在储层分布参数定量预测中的应用[J].地球物理学进展,2007,22(2):522 555.
[7]宋维奇,王小马,杜玉民,等.综合应用地震属性、测井数据反演储层参数[J].石油地球物理勘探,2002,37(5):491 494.
[8]李安东,宁俊瑞,刘振峰.用神经网络和地质统计学综合多元信息进行储层预测[J].石油与天然气地质,2010,31(4):493 498,503.
[9]WangXingjian,HuGuangmin,CaoJun-Xing.Applicationof multiple attributes fusion technology in the Su-14 wellblock[J].Applied Geophysics,2010,7(3):257 264.
[10]Kennedy J,Eberhart R C.Particle swarm optimiza-tion[C]//Proceedingsof IEEEInternationalConferenceonNeural Networks,Piscataway NJ:1995:1942 1948.
[11]高海兵,高亮,周驰,等.基于粒子群优化的神经网络训练算法研究[J].电子学报,2004,32(9):1572 1574.
[12]ZhangJingru,ZhangJun,LokTM,etal.AHybridparticleswarm optimization-back-propagation algorithm for feed-forward neural network training[J].Applied Mathematicsand Computation,2007,185(2):1026 1037.
[13]唐贤伦,庄陵,李银国,等.混合粒子群优化算法优化前向神经网络结构和参数[J].计算机应用研究,2007,24(12):91 93.
[14]Carvalho M,Ludermir T B.Particle Swarm Optimizationof neural network architectures and weights[C]//7th Inter-national Conference on Hybrid Intelligent Systems 2007,Kaiserslautern:2007:336 339.
[15]吴秋波,王允诚,赵秋亮,等.混沌惯性权值调整策略的粒子群优化算法[J].计算机工程与应用,2009,45(7):49 51.
[16]Feng Y,Yao Y M,Wang A X,et al.Compare with chaoticinertia weights in particle swarm optimization[C]//2007International Conference on Machine Learning and Cy-bernetics,Hong Kong:2007:329 333.
[17]魏海坤.神经网络结构设计的理论与方法[M].北京:国防工业出版社,2005.
[18]田雨波.混合神经网络技术[M].北京:科学出版社,2009.
[2]Liu Zhengping,Liu Jiaqi.Seismic-controlled nonlinearextrapolation of well parameters using neural network[J].Geophysics,1998,63(6):2035 2041.
[3]吴大奎,李亚林,伍志明,等.地震、测井资料联合反演的神经网络算法研究[J].天然气工业,2004,24(3):55 57.
[4]杨辉廷,颜其彬,李敏,等.应用改进的神经网络学习方法预测储层参数[J].天然气工业,2005,25(6):37 39.
[5]冉启全,周雪峰,胡永乐.基于神经网络的多属性火成岩岩性反演技术[J].西南石油学院学报,2006,28(1):5 8,17.
[6]郑庆生,韩大匡.高阶神经网络在储层分布参数定量预测中的应用[J].地球物理学进展,2007,22(2):522 555.
[7]宋维奇,王小马,杜玉民,等.综合应用地震属性、测井数据反演储层参数[J].石油地球物理勘探,2002,37(5):491 494.
[8]李安东,宁俊瑞,刘振峰.用神经网络和地质统计学综合多元信息进行储层预测[J].石油与天然气地质,2010,31(4):493 498,503.
[9]WangXingjian,HuGuangmin,CaoJun-Xing.Applicationof multiple attributes fusion technology in the Su-14 wellblock[J].Applied Geophysics,2010,7(3):257 264.
[10]Kennedy J,Eberhart R C.Particle swarm optimiza-tion[C]//Proceedingsof IEEEInternationalConferenceonNeural Networks,Piscataway NJ:1995:1942 1948.
[11]高海兵,高亮,周驰,等.基于粒子群优化的神经网络训练算法研究[J].电子学报,2004,32(9):1572 1574.
[12]ZhangJingru,ZhangJun,LokTM,etal.AHybridparticleswarm optimization-back-propagation algorithm for feed-forward neural network training[J].Applied Mathematicsand Computation,2007,185(2):1026 1037.
[13]唐贤伦,庄陵,李银国,等.混合粒子群优化算法优化前向神经网络结构和参数[J].计算机应用研究,2007,24(12):91 93.
[14]Carvalho M,Ludermir T B.Particle Swarm Optimizationof neural network architectures and weights[C]//7th Inter-national Conference on Hybrid Intelligent Systems 2007,Kaiserslautern:2007:336 339.
[15]吴秋波,王允诚,赵秋亮,等.混沌惯性权值调整策略的粒子群优化算法[J].计算机工程与应用,2009,45(7):49 51.
[16]Feng Y,Yao Y M,Wang A X,et al.Compare with chaoticinertia weights in particle swarm optimization[C]//2007International Conference on Machine Learning and Cy-bernetics,Hong Kong:2007:329 333.
[17]魏海坤.神经网络结构设计的理论与方法[M].北京:国防工业出版社,2005.
[18]田雨波.混合神经网络技术[M].北京:科学出版社,2009.
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