en
×

分享给微信好友或者朋友圈

使用微信“扫一扫”功能。
作者简介:

王健(1994-)男,博士研究生,研究方向为机器学习在石油天然气钻探开发中应用。E-mail: b20020065@s.upc.edu.cn。

通信作者:

徐加放(1973-),男,教授,博士,博士生导师,研究方向为水合物开发及防治,钻井液处理剂及体系,井壁稳定、防漏堵漏、储层敏感性评价及保护,钻井废弃物无害化处理及资源化利用,分子模拟与微流控技术等。E-mail: xjiafang@upc.edu.cn。

中图分类号:TE832

文献标识码:A

文章编号:1673-5005(2025)06-0172-09

DOI:10.3969/j.issn.1673-5005.2025.06.017

参考文献 1
李乐,朱倩谊,陈小康,等.纯水体系下水合物的生成及堵塞实验研究[J].西南石油大学学报(自然科学版),2019,41(4):152-158.LI Le,ZHU Qianyi,CHEN Xiaokang,et al.Experimental study on hydrate formation and plugging in pure water system[J].Journal of Southwest Petroleum University(Science & Technology Edition),2019,41(4):152-158.
参考文献 2
白云程,孙敬杰,罗向权,等.石油工程中的水合物抑制[J].特种油气藏,2006,13(2):5-8.BAI Yuncheng,SUN Jingjie,LUO Xiangquan,et al.Gas hydrate inhibition in oil engineering[J].Special Oil & Gas Reservoirs,2006,13(2):5-8.
参考文献 3
赵欣,邱正松,黄维安,等.天然气水合物热力学抑制剂作用机制及优化设计[J].石油学报,2015,36(6):760-766.ZHAO Xin,QIU Zhengsong,HUANG Weian,et al.Mechanism and optimization design of thermodynamic inhibitor for natural gas hydrate[J].Acta Petrolei Sinica,2015,36(6):760-766.
参考文献 4
顾锋,赵会军,王树立,等.盐类体系中天然气水合物相平衡条件的研究[J].石油与天然气化工,2008,37(2):149-151.GU Feng,ZHAO Huijun,WANG Shuli,et al.Study on phase equilibrium conditions of natural gas hydrate in salt system[J].Chemical Engineering of Oil and Gas,2008,37(2):149-151.
参考文献 5
杨明军,宋永臣,刘瑜,等.多孔介质及盐度对甲烷水合物相平衡影响[J].大连理工大学学报,2011,51(1):31-35.YANG Mingjun,SONG Yongchen,LIU Yu,et al.Effect of porou media and salinity on phase equilibrium of methane hydrate[J].Journal of Dalian University of Technology,2011,51(1):31-35.
参考文献 6
啜世阳,王华宁,张宁,等.天然气水合物温度-压力-盐浓度三维相平衡曲面方程[J].石油与天然气化工,2019,48(5):49-55.CHUAI Shiyang,WANG Huaning,ZHANG Ning,et al.Three-dimensional phase equilibrium surface equation of temperature-pressure-salt concentration of natural gas hydrate[J].Chemical Engineering of Oil and Gas,2019,48(5):49-55.
参考文献 7
耿菲凡,董波,周训,等.盐度对甲烷水合物分解特性影响的LBM模拟[J].计算力学学报,2024,41(5):929-934,962.GENG Feifan,DONG Bo,ZHOU Xun,et al.LBM simulation of the influence of salinity on decomposition characteristics of methane hydrate[J].Chinese Journal of Computational Mechanics,2024,41(5):929-934,962.
参考文献 8
汪敏,杨桃,唐洪明,等.迁移深度神经网络的页岩总孔隙度预测[J].西南石油大学学报(自然科学版),2023,45(6):69-79.WANG Min,YANG Tao,TANG Hongming,et al.Prediction of total porosity of shale by migration depth neural network[J].Journal of Southwest Petroleum University(Science & Technology Edition),2023,45(6):69-79.
参考文献 9
SONG Y,ZHOU H,WANG P,et al.Prediction of clathrate hydrate phase equilibria using gradient boosted regression trees and deep neural networks[J].The Journal of Chemical Thermodynamics,2019,135:86-96.
参考文献 10
董晓霞,冯少柯.基于神经网络聚类分析的深层页岩储层岩相识别:以川南筇竹寺组为例[J].西南石油大学学报(自然科学版),2024,46(6):61-73.DONG Xiaoxia,FENG Shaoke.Lithofacies identification in deep shale reservoirs via neural network clustering analysis:a case study of the qiongzhusi formation in the southern Sichuan Basin[J].Journal of Southwest Petroleum University(Science & Technology Edition),2024,46(6):61-73.
参考文献 11
MEHRIZADEH M.Prediction of gas hydrate formation using empirical equations and data-driven models[J].Materials Today:Proceedings,2021,42:1592-1598.
参考文献 12
REBAI N,HADJADJ A,BENMOUNAH A,et al.Prediction of natural gas hydrates formation using a combination of thermodynamic and neural network modeling[J].Journal of Petroleum Science and Engineering,2019,182:106270.
参考文献 13
于丽丽,刘永红,蔡宝平,等.基于小波神经网络的双电极同步伺服放电加工工艺效果预测[J].开云电竞投注学报(自然科学版),2008,32(4):87-90.YU Lili,LIU Yonghong,CAI Baoping,et al.Effect prediction of dual-electrode synchronous servo discharge machining based on wavelet neural network[J].Journal of China University of Petroleum(Edition of Natural Science),2008,32(4):87-90.
参考文献 14
ZHAO Y,DU X,XIA G,et al.A novel algorithm for wavelet neural networks with application to enhanced PID controller design[J].Neurocomputing,2015,158:257-267.
参考文献 15
ALI Y,ALY H H.Short term wind speed forecasting using artificial and wavelet neural networks with and without wavelet filtered data based on feature selections technique[J].Engineering Applications of Artificial Intelligence,2024,133:108201.
参考文献 16
王健,徐加放,王博闻,等.基于GA-Elman神经网络的煤层气临界解吸压力预测[J].开云电竞投注学报(自然科学版),2024,48(5):138-145.WANG Jian,XU Jiafang,WANG Bowen,et al.Prediction of critical desorption pressure of coalbed methane based on GA-Elman neural network[J].Journal of China University of Petroleum(Edition of Natural Science),2024,48(5):138-145.
参考文献 17
李伟,袁新安,曲萌,等.基于GA-BP神经网络的ACFM实时高精度裂纹反演算法[J].开云电竞投注学报(自然科学版),2016,40(5):128-134.LI Wei,YUAN Xinan,QU Meng,et al.ACFM real-time high-precision crack inversion algorithm based on GA-BP neural network[J].Journal of China University of Petroleum(Edition of Natural Science),2016,40(5):128-134.
参考文献 18
WU Y,GAO R,YANG J.Prediction of coal and gas outburst:a method based on the BP neural network optimized by GASA[J].Process Safety and Environmental Protection,2020,133:64-72.
参考文献 19
李华,荣婕妤.基于AdaBoost方法的近红外光谱建模分析[J].吉林师范大学学报(自然科学版),2022,43(3):83-89.LI Hua,RONG Jieyu.Modeling and analysis of near infrared spectroscopy based on AdaBoost method[J].Journal of Jilin Normal University(Natural Science Edition),2022,43(3):83-89.
参考文献 20
肖捡花.交通领域下在线集成时间序列预测方法及应用研究[D].长沙:湖南大学,2020.XIAO Jianhua.Research on online ensemble time series prediction method and application in transportation field[D].Changsha:Hunan University,2020.
参考文献 21
邱腾煌,钱玉宝,季威,等.基于GA-BPNN混合智能模型的钻速预测[J].电子测量技术,2024,47(15):177-186.QIU Tenghuang,QIAN Yubao,JI Wei,et al.Prediction of ROP based on GA-BPNN hybrid intelligent model[J].Electronic Measurement Technology,2024,47(15):177-186.
参考文献 22
苏莫婷.基于机器学习方法的天然气价格和消费量预测研究[D].重庆:重庆大学,2022.SU Moting.Research on prediction of natural gas price and consumption based on machine learning methods[D].Chongqing:Chongqing University,2022.
参考文献 23
SLOAN E D Jr,KOH C A,KOH C A.Clathrate hydrates of natural gases[M].Boca Raton:CRC Press,2007:461-519.
参考文献 24
GAHLEN P,MAINKA R,STOMMEL M.Prediction of anisotropic foam stiffness properties by a neural network[J].International Journal of Mechanical Sciences,2023,249:108245.
参考文献 25
宋先知,朱硕,李根生,等.基于BP-LSTM双输入网络的大钩载荷与转盘扭矩预测[J].开云电竞投注学报(自然科学版),2022,46(2):76-84.SONG Xianzhi,ZHU Shuo,LI Gensheng,et al.Prediction of hook load and turntable torque based on BP-LSTM dual-input network[J].Journal of China University of Petroleum(Edition of Natural Science),2022,46(2):76-84.
参考文献 26
ZHANG X,HOU L,LIU J,et al.Energy consumption prediction for crude oil pipelines based on integrating mechanism analysis and data mining[J].Energy,2022,254:124382.
参考文献 27
YE K,WANG J,GAO H,et al.Optimization of lapping process parameters of CP-Ti based on PSO with mutation and BPNN[J].The International Journal of Advanced Manufacturing Technology,2021,117(9):2859-2866.
参考文献 28
CHENG S,WU Y,LI Y,et al.TWD-SFNN:three-way decisions with a single hidden layer feedforward neural network[J].Information Sciences,2021,579:15-32.
参考文献 29
YAN H,ZHANG J,ZHOU N,et al.Application of hybrid artificial intelligence model to predict coal strength alteration during CO2 geological sequestration in coal seams[J].Science of the Total Environment,2020,711:135029.
目录contents

    摘要

    以NaCl浓度预测为例,将核主成分分析(KPCA)处理后的温度、压力和气体组分作为输入参数,利用小波神经网络(WNN)对NaCl浓度进行预测,并通过遗传退火算法(GASA)和AdaBoost算法对WNN进行优化,建立AdaBoost-GASA-WNN水合物抑制剂NaCl浓度预测模型。结果表明,经过KPCA处理后模型的均方误差(eMSE)降低了3.87,优化后模型的eMSE进一步降低到9.51,与ELM、KNN、RF模型和数据拟合方法相比,eMSE分别低5.8、17.74、2.91和8.81,预测效果最好。

    Abstract

    Taking the prediction of NaCl concentration as an example, the temperature, pressure, and gas components processed by kernel principal component analysis (KPCA) were used as input parameters. The wavelet neural network (WNN) was utilized to predict the NaCl concentration, and the WNN was optimized by the Genetic annealing algorithm (GASA) and AdaBoost algorithm. Then, the AdaBoost-GASA-WNN prediction model for the concentration of the hydrate inhibitor NaCl was established. The results show that the man square error (eMSE) of the model is decreased by 3.87 after KPCA processing, and the eMSE of the optimized model is further reduced to 9.51. Compared with ELM, KNN, RF, and data fitting methods, the eMSE is declined by 5.8, 17.74, 2.91, and 8.81, respectively. And the prediction effect is the best.

  • 天然气水合物堵塞是海洋油气钻井过程中经常遇到的问题,同时也是石油天然气管道运输中最严重的故障之一[1],会造成管道和设备的堵塞,甚至产生环境和安全问题,因此需要采用适当的方法来防止水合物生成。目前常用的处理方法之一是添加相应的抑制剂,其中无机盐类由于其良好的抑制效果和低廉的价格常常被用来抑制水合物生成,然而其添加量至关重要,加量过少无法有效抑制水合物生成,加量过多又会在增加成本的同时对设备造成腐蚀,另外在钻井过程中也会对钻井液性能产生影响[2],因此对盐组分含量进行评估预测非常重要。对此,许多专家学者进行了相应的研究,包括盐分对水活度的影响[3],对水合物相平衡[4]及其曲线的影响[5]等。随着研究的深入,一些盐分存在下的相平衡预测模型也逐渐被提出,如啜世阳等[6]建立了盐浓度、天然气水合物温度和压力的三维相平衡方程;耿菲凡等[7]基于格子玻尔兹曼方法建立了孔隙尺度下甲烷水合物在盐水中分解的数值模型。虽然上述研究可以计算盐分存在下水合物相平衡条件,但是其公式较为复杂且无法对盐浓度进行直接计算,现场人员应用困难。随着人工智能技术的不断发展,神经网络预测方法因其使用简单方便,被广泛应用到油气邻域[8-10],其中一些研究人员使用神经网络来预测各种条件下水合物的相平衡温度和压力,如Mehrizadeh[11]提出了人工神经网络和自适应神经模糊推理系统(ANFIS)用于预测不同水合物形成系统中的水合物形成压力,另外有学者建立了人工智能预测模型来纠正热力学模型中的预测误差[12]等。虽然人工智能在水合物研究领域应用较为广泛,但是目前还没有针对无机盐抑制剂浓度预测的研究。笔者以NaCl为例,利用小波神经网络(WNN)建立水合物无机盐抑制剂浓度预测模型,同时通过遗传退火算法(GASA)对其权值和阈值进行优化,并利用AdaBoost算法进行集成优化,建立AdaBoost-GASA-WNN水合物无机盐抑制剂浓度预测模型。

  • 1 预测模型的建立

  • 1.1 WNN模型

  • WNN是基于反向传播算法且以小波基函数作为隐层传递函数的前馈神经网络[13],其通过引入小波分析,兼顾了小波理论和前馈神经网络的优点,具备了神经网络的自学习能力和小波变换良好的时频局域化性质,使得WNN具有较强的容错能力、泛化能力和学习能力[14],比传统的多层神经网络更有效,因此选用WNN作为基础预测模型,其结构如图1所示[15]

  • 图1 WNN结构

  • Fig.1 Structure of WNN

  • 1.2 GASA优化算法

  • 由于WNN属于前馈神经网络,其权值和阈值采用误差反向传播算法来进行调整,因此其易陷入局部极值而无法找到全局最优解[16],需要相应的优化算法来对其进行优化。GA是模拟生物进化的全局优化方法,具有较好的全局寻优能力,可对神经网络的阈值和权值进行优化找到全局最优解,提升预测精度[17]。SA是一种局部搜索能力很强的优化算法,可以以一定的概率接受次优解从而跳出局部极值,接近全局最优值。因此将SA算法引入到GA算法当中,在每执行完一次遗传迭代操作,计算出每个个体的适应度后,再利用SA算法对优化后的个体进行进一步筛选,以一定的概率接受次优解,进而有效保证遗传迭代的物种多样性,避免早熟现象的发生,使得算法既具有良好的全局控制能力,又具有理想的局部搜索能力,大大提升优化效果[18],其流程如2图所示。

  • 图2 GASA算法流程

  • Fig.2 Flowchart of GASA algorithm

  • 1.3 AdaBoost算法

  • 在建立水合物盐抑制剂浓度预测模型过程中,由于参数间的关系较为复杂且训练数据集有限,单一的GASA-WNN模型在某些情况下无法很好地反映输入输出参数间的非线性复杂关系。AdaBoost算法是一种自适应增强算法(图3),算法里每个样本点的权重都会得到充分考虑,在解决一些复杂问题时有显著优势[19],是使用最为广泛的集成学习预测模型之一[20]。因此,以GASA-WNN模型为弱预测器,利用AdaBoost算法将多个GASA-WNN模型进行组合,在迭代过程中通过增加误差大的样本的权重,使得GASA-WNN模型重点关注误差大的样本,从而逐步修正整体模型的预测偏差,最终形成一个可以全面反映输入输出参数间非线性复杂关系且预测精度高的集成模型,能够有效减少训练错误率,提升模型稳定性,提高模型的整体性能。

  • 图3 AdaBoost算法流程

  • Fig.3 AdaBoost algorithm flowchart

  • 1.4 评价指标

  • 为全面评估模型的性能,选取4个关键评价指标,分别为均方误差(eMSE)、均方根误差(eRMSE)、平均绝对误差(eMAE)和决定系数(R[21-22]对模型进行评估,其计算公式为

  • R=i=1n yi-y-i=1n yi'-y'¯i=1n yi-y-2i=1n yi'-y'¯2,
    (1)
  • eMAE=1nn=1n yi-yi',
    (2)
  • eRMSE=1ni=1n yi-yi'2,
    (3)
  • eMSE=1ni=1n yi-yi'2.

  • eMSE=1ni=1n yi-yi'2.
    (4)
  • 式中,yi为第i个数据实际值,y′i 为其对应预测值; y'¯y-分别为预测数据和实际数据的平均值; n为样本个数。

  • 2 数据分析与处理

  • 通过查阅相关文献[23],共收集了272组相关数据,同时对数据进行了如表1所示的统计分析。另外,通过相关性分析,计算了上述参数间的相关性,结果如图4所示。

  • 由图4可知,各个参数与NaCl相关性由强到弱分别为:H2S、T、C2H6、CH4、C4H10、CO2p和C3H8,其中T、C2H6、CO2和H2S与NaCl呈正相关,其余参数呈负相关,因此在利用本模型进行预测时尤其需要取准H2S和T等数据的值。另外由表1可知,气体参数的最小值和中间值都为0,最大值为100,说明在收集的数据中纯组分气体相对较多,相关数据分布不均匀,这种情况很可能会对模型的性能造成影响。因此为了降低数据分布不均的影响和消除冗余信息,利用核主成分分析对上述输入数据进行了处理,同时对比分析了不同核函数的处理结果,选择处理后贡献率高于95%或维度不超过原始维度的数据作为输入数据,结果见表2。

  • 表1 数据特征

  • Table1 Parameter characteristics of dataset

  • 图4 输入输出参数相关系数矩阵

  • Fig.4 Matrix of degree of correlation between various factors

  • 由表2可知,除了指数核和拉普拉斯核主成分分析处理后的数据维度不变以外,其他核主成分分析方法在一定程度上降低了数据维度。为选取合适的数据处理方法,利用不同核主成分分析后的数据和未处理数据建立了对应的预测模型来对比分析预测效果。另外由于数据量有限,为了更加准确地对各个数据集进行分析评价和后续选取最佳隐层节点数和传递函数,采用如图5所示的K-折交叉验证[24](本文中K值设置为10)方法,把数据集分成10份,每次把一份数据当作测试集,其余当作训练集,依次循环,从而对各个模型进行全面系统的评价,不同核主成分分析处理方法对应预测结果见表3。

  • 表2 不同核主成分分析结果

  • Table2 Results of different kernel principal component analysis

  • 图5 交叉验证示意图

  • Fig.5 Cross validation schematic diagram

  • 表3 预测结果

  • Table3 Prediction results

  • 由表3可知,拉普拉斯核函数处理后的数据精度最高,因此后续建模利用拉普拉斯核函数处理后的数据。另外为了消除各输入参数维度对预测性能的影响,对相关数据进行了归一化处理[25-26],把所有的参数都归一化到区间[0,1]内,从而消除参数间数量级差别的影响。归一化计算公式为

  • xi'=xi-xminxmax-xmin.
    (5)
  • 式中,xi为原始数据; xminxmax分别为同类数据中的最小值和最大值; x′i为归一化后的数据。

  • 3 模型参数的确定

  • 3.1 隐层节点数

  • 合适的隐层节点数对模型的准确性十分重要。选择合适隐层节点数的经验公式[27-28]

  • k=n+m+a.
    (6)
  • 式中,m为输入层节点数;n为输出层节点数;a为1到10的任意数。

  • 通过式(6)可知,隐层节点数取值范围为[313],针对这一范围分别建立不同隐层节点数的WNN预测模型,并利用相关数据进行训练和测试,同时进行计算分析,结果见图6。可以看出,当隐层节点数为11时,预测效果最好。

  • 图6 不同隐层节点数预测结果

  • Fig.6 Results for different numbers of hidden layer nodes

  • 3.2 小波函数

  • 在WNN预测模型建立过程中,选择不同的小波传递函数也会对预测效果产生一定的影响。对此选取4种常用的小波传递函数进行对比分析,结果见表4。

  • 表4 不同传递函数预测结果

  • Table4 Prediction results of different wavelet functions

  • 由表4可知,当小波传递函数为Mymorlet函数时,整体预测效果最好。另外模型使用的遗传算法的个体数目为20,遗传代数设置为50,交叉及变异概率分别为0.7和0.01,模拟退火算法中冷却系数为0.8,初始和终止温度分别为10和1。

  • 4 模型结果分析

  • 4.1 模型优化

  • 为分析优化算法的优化效果,计算不同优化算法下预测模型的预测结果,如表5及图7所示。

  • 表5 不同优化算法下的预测结果

  • Table5 Prediction results under different optimization algorithms

  • 模型的预测精度从低到高依次为WNN、GA-WNN、GASA-WNN、AdaBoost-GASA-WNN。与其他模型相比,AdaBoost-GASA-WNN预测模型的预测结果误差最小,最接近实际值,表明AdaBoost-GASA-WNN模型具有最高的预测精度。在构建AdaBoost-GASA-WNN模型的过程中,GA可以优化WNN的权值和阈值,从而提高WNN的预测精度,而SA可以解决GA梯度消失的问题,找到全局最优解,进一步提升预测精度,AdaBoost算法可以通过调整权重,使预测精度高的弱预测器的权重变得更高,进而使得预测模型的预测精度和泛化能力都得到了一定的提高[29]

  • 图7 模型优化效果对比

  • Fig.7 Comparative analysis of model optimization effects

  • 4.2 模型预测

  • 为了分析优化后模型的预测效果,对比分析优化后模型的训练和测试结果(图8)。可以看出,无论是训练结果还是测试结果,少部分数据拟合效果较差,大部分数据的预测值接近实际值,说明预测效果相对较好,可以对w(NaCl)进行较为精准的预测。

  • 图8 AdaBoost-GASA-WNN预测模型结果

  • Fig.8 Results of AdaBoost-GASA-WNN prediction model

  • 5 模型应用

  • 为了验证所开发模型的稳健性和可靠性,利用验证数据来验证模型的效果,同时使用相同的数据建立了极限学习机(ELM)、随机森林(RF)、K邻近拟合(KNN)预测模型和常用的数据拟合方法进行预测并对结果进行对比分析,结果见图9和表6。

  • 图9 不同预测模型的预测结果

  • Fig.9 Prediction results of different prediction models

  • 由图9及表6可知,AdaBoost-GASA-WNN模型的ReMAE分别为0.81和2.32,预测精度最高,RF次之,ELM和数据拟合方法预测结果误差相对较大,KNN预测误差最大。另外,AdaBoost-GASA-WNN模型预测结果的绝对误差也远远低于其他预测模型,其最大值为8.57,比其他模型中预测精度最高的RF低0.91,说明AdaBoost-GASA-WNN预测模型的稳定性也优于其他模型。综合分析可知,所提出的混合预测模型整体预测精度较高,可为管道油气输送及钻探过程中的水合物防治提供一定的参考。

  • 表6 不同模型预测结果

  • Table6 Analysis of prediction results of different models

  • 6 结论

  • (1)通过相关性分析,各个参数与NaCl相关性由强到弱分别为H2S、T、C2H6、CH4、C4H10、CO2p和C3H8,因此在利用本模型对NaCl抑制剂加量进行计算时,尤其需要取准H2S和T等参数的值。

  • (2)对输入参数进行KPCA处理后模型的eMSE降低了3.86,利用GASA和AdaBoost算法优化后模型的eMSE进一步降低了6.37,有效提升了模型的预测性能。

  • (1)通过相关性分析,各个参数与NaCl相关性由强到弱分别为H2S、T、C2H6、CH4、C4H10、CO2p和C3H8,因此在利用本模型对NaCl抑制剂加量进行计算时,尤其需要取准H2S和T等参数的值。

  • (2)对输入参数进行KPCA处理后模型的eMSE降低了3.86,利用GASA和AdaBoost算法优化后模型的eMSE进一步降低了6.37,有效提升了模型的预测性能。

  • (3)AdaBoost-GASA-WNN模型预测的eMAE为2.32,与其他预测模型相比误差最小,是预测w(NaCl)的一种高效准确的方法。由于数据量有限,模型只对NaCl进行了预测,未来将加入其他无机盐种类对模型进行进一步训练以提升模型的适应能力。

  • 参考文献

    • [1] 李乐,朱倩谊,陈小康,等.纯水体系下水合物的生成及堵塞实验研究[J].西南石油大学学报(自然科学版),2019,41(4):152-158.LI Le,ZHU Qianyi,CHEN Xiaokang,et al.Experimental study on hydrate formation and plugging in pure water system[J].Journal of Southwest Petroleum University(Science & Technology Edition),2019,41(4):152-158.

    • [2] 白云程,孙敬杰,罗向权,等.石油工程中的水合物抑制[J].特种油气藏,2006,13(2):5-8.BAI Yuncheng,SUN Jingjie,LUO Xiangquan,et al.Gas hydrate inhibition in oil engineering[J].Special Oil & Gas Reservoirs,2006,13(2):5-8.

    • [3] 赵欣,邱正松,黄维安,等.天然气水合物热力学抑制剂作用机制及优化设计[J].石油学报,2015,36(6):760-766.ZHAO Xin,QIU Zhengsong,HUANG Weian,et al.Mechanism and optimization design of thermodynamic inhibitor for natural gas hydrate[J].Acta Petrolei Sinica,2015,36(6):760-766.

    • [4] 顾锋,赵会军,王树立,等.盐类体系中天然气水合物相平衡条件的研究[J].石油与天然气化工,2008,37(2):149-151.GU Feng,ZHAO Huijun,WANG Shuli,et al.Study on phase equilibrium conditions of natural gas hydrate in salt system[J].Chemical Engineering of Oil and Gas,2008,37(2):149-151.

    • [5] 杨明军,宋永臣,刘瑜,等.多孔介质及盐度对甲烷水合物相平衡影响[J].大连理工大学学报,2011,51(1):31-35.YANG Mingjun,SONG Yongchen,LIU Yu,et al.Effect of porou media and salinity on phase equilibrium of methane hydrate[J].Journal of Dalian University of Technology,2011,51(1):31-35.

    • [6] 啜世阳,王华宁,张宁,等.天然气水合物温度-压力-盐浓度三维相平衡曲面方程[J].石油与天然气化工,2019,48(5):49-55.CHUAI Shiyang,WANG Huaning,ZHANG Ning,et al.Three-dimensional phase equilibrium surface equation of temperature-pressure-salt concentration of natural gas hydrate[J].Chemical Engineering of Oil and Gas,2019,48(5):49-55.

    • [7] 耿菲凡,董波,周训,等.盐度对甲烷水合物分解特性影响的LBM模拟[J].计算力学学报,2024,41(5):929-934,962.GENG Feifan,DONG Bo,ZHOU Xun,et al.LBM simulation of the influence of salinity on decomposition characteristics of methane hydrate[J].Chinese Journal of Computational Mechanics,2024,41(5):929-934,962.

    • [8] 汪敏,杨桃,唐洪明,等.迁移深度神经网络的页岩总孔隙度预测[J].西南石油大学学报(自然科学版),2023,45(6):69-79.WANG Min,YANG Tao,TANG Hongming,et al.Prediction of total porosity of shale by migration depth neural network[J].Journal of Southwest Petroleum University(Science & Technology Edition),2023,45(6):69-79.

    • [9] SONG Y,ZHOU H,WANG P,et al.Prediction of clathrate hydrate phase equilibria using gradient boosted regression trees and deep neural networks[J].The Journal of Chemical Thermodynamics,2019,135:86-96.

    • [10] 董晓霞,冯少柯.基于神经网络聚类分析的深层页岩储层岩相识别:以川南筇竹寺组为例[J].西南石油大学学报(自然科学版),2024,46(6):61-73.DONG Xiaoxia,FENG Shaoke.Lithofacies identification in deep shale reservoirs via neural network clustering analysis:a case study of the qiongzhusi formation in the southern Sichuan Basin[J].Journal of Southwest Petroleum University(Science & Technology Edition),2024,46(6):61-73.

    • [11] MEHRIZADEH M.Prediction of gas hydrate formation using empirical equations and data-driven models[J].Materials Today:Proceedings,2021,42:1592-1598.

    • [12] REBAI N,HADJADJ A,BENMOUNAH A,et al.Prediction of natural gas hydrates formation using a combination of thermodynamic and neural network modeling[J].Journal of Petroleum Science and Engineering,2019,182:106270.

    • [13] 于丽丽,刘永红,蔡宝平,等.基于小波神经网络的双电极同步伺服放电加工工艺效果预测[J].开云电竞投注学报(自然科学版),2008,32(4):87-90.YU Lili,LIU Yonghong,CAI Baoping,et al.Effect prediction of dual-electrode synchronous servo discharge machining based on wavelet neural network[J].Journal of China University of Petroleum(Edition of Natural Science),2008,32(4):87-90.

    • [14] ZHAO Y,DU X,XIA G,et al.A novel algorithm for wavelet neural networks with application to enhanced PID controller design[J].Neurocomputing,2015,158:257-267.

    • [15] ALI Y,ALY H H.Short term wind speed forecasting using artificial and wavelet neural networks with and without wavelet filtered data based on feature selections technique[J].Engineering Applications of Artificial Intelligence,2024,133:108201.

    • [16] 王健,徐加放,王博闻,等.基于GA-Elman神经网络的煤层气临界解吸压力预测[J].开云电竞投注学报(自然科学版),2024,48(5):138-145.WANG Jian,XU Jiafang,WANG Bowen,et al.Prediction of critical desorption pressure of coalbed methane based on GA-Elman neural network[J].Journal of China University of Petroleum(Edition of Natural Science),2024,48(5):138-145.

    • [17] 李伟,袁新安,曲萌,等.基于GA-BP神经网络的ACFM实时高精度裂纹反演算法[J].开云电竞投注学报(自然科学版),2016,40(5):128-134.LI Wei,YUAN Xinan,QU Meng,et al.ACFM real-time high-precision crack inversion algorithm based on GA-BP neural network[J].Journal of China University of Petroleum(Edition of Natural Science),2016,40(5):128-134.

    • [18] WU Y,GAO R,YANG J.Prediction of coal and gas outburst:a method based on the BP neural network optimized by GASA[J].Process Safety and Environmental Protection,2020,133:64-72.

    • [19] 李华,荣婕妤.基于AdaBoost方法的近红外光谱建模分析[J].吉林师范大学学报(自然科学版),2022,43(3):83-89.LI Hua,RONG Jieyu.Modeling and analysis of near infrared spectroscopy based on AdaBoost method[J].Journal of Jilin Normal University(Natural Science Edition),2022,43(3):83-89.

    • [20] 肖捡花.交通领域下在线集成时间序列预测方法及应用研究[D].长沙:湖南大学,2020.XIAO Jianhua.Research on online ensemble time series prediction method and application in transportation field[D].Changsha:Hunan University,2020.

    • [21] 邱腾煌,钱玉宝,季威,等.基于GA-BPNN混合智能模型的钻速预测[J].电子测量技术,2024,47(15):177-186.QIU Tenghuang,QIAN Yubao,JI Wei,et al.Prediction of ROP based on GA-BPNN hybrid intelligent model[J].Electronic Measurement Technology,2024,47(15):177-186.

    • [22] 苏莫婷.基于机器学习方法的天然气价格和消费量预测研究[D].重庆:重庆大学,2022.SU Moting.Research on prediction of natural gas price and consumption based on machine learning methods[D].Chongqing:Chongqing University,2022.

    • [23] SLOAN E D Jr,KOH C A,KOH C A.Clathrate hydrates of natural gases[M].Boca Raton:CRC Press,2007:461-519.

    • [24] GAHLEN P,MAINKA R,STOMMEL M.Prediction of anisotropic foam stiffness properties by a neural network[J].International Journal of Mechanical Sciences,2023,249:108245.

    • [25] 宋先知,朱硕,李根生,等.基于BP-LSTM双输入网络的大钩载荷与转盘扭矩预测[J].开云电竞投注学报(自然科学版),2022,46(2):76-84.SONG Xianzhi,ZHU Shuo,LI Gensheng,et al.Prediction of hook load and turntable torque based on BP-LSTM dual-input network[J].Journal of China University of Petroleum(Edition of Natural Science),2022,46(2):76-84.

    • [26] ZHANG X,HOU L,LIU J,et al.Energy consumption prediction for crude oil pipelines based on integrating mechanism analysis and data mining[J].Energy,2022,254:124382.

    • [27] YE K,WANG J,GAO H,et al.Optimization of lapping process parameters of CP-Ti based on PSO with mutation and BPNN[J].The International Journal of Advanced Manufacturing Technology,2021,117(9):2859-2866.

    • [28] CHENG S,WU Y,LI Y,et al.TWD-SFNN:three-way decisions with a single hidden layer feedforward neural network[J].Information Sciences,2021,579:15-32.

    • [29] YAN H,ZHANG J,ZHOU N,et al.Application of hybrid artificial intelligence model to predict coal strength alteration during CO2 geological sequestration in coal seams[J].Science of the Total Environment,2020,711:135029.

  • 参考文献

    • [1] 李乐,朱倩谊,陈小康,等.纯水体系下水合物的生成及堵塞实验研究[J].西南石油大学学报(自然科学版),2019,41(4):152-158.LI Le,ZHU Qianyi,CHEN Xiaokang,et al.Experimental study on hydrate formation and plugging in pure water system[J].Journal of Southwest Petroleum University(Science & Technology Edition),2019,41(4):152-158.

    • [2] 白云程,孙敬杰,罗向权,等.石油工程中的水合物抑制[J].特种油气藏,2006,13(2):5-8.BAI Yuncheng,SUN Jingjie,LUO Xiangquan,et al.Gas hydrate inhibition in oil engineering[J].Special Oil & Gas Reservoirs,2006,13(2):5-8.

    • [3] 赵欣,邱正松,黄维安,等.天然气水合物热力学抑制剂作用机制及优化设计[J].石油学报,2015,36(6):760-766.ZHAO Xin,QIU Zhengsong,HUANG Weian,et al.Mechanism and optimization design of thermodynamic inhibitor for natural gas hydrate[J].Acta Petrolei Sinica,2015,36(6):760-766.

    • [4] 顾锋,赵会军,王树立,等.盐类体系中天然气水合物相平衡条件的研究[J].石油与天然气化工,2008,37(2):149-151.GU Feng,ZHAO Huijun,WANG Shuli,et al.Study on phase equilibrium conditions of natural gas hydrate in salt system[J].Chemical Engineering of Oil and Gas,2008,37(2):149-151.

    • [5] 杨明军,宋永臣,刘瑜,等.多孔介质及盐度对甲烷水合物相平衡影响[J].大连理工大学学报,2011,51(1):31-35.YANG Mingjun,SONG Yongchen,LIU Yu,et al.Effect of porou media and salinity on phase equilibrium of methane hydrate[J].Journal of Dalian University of Technology,2011,51(1):31-35.

    • [6] 啜世阳,王华宁,张宁,等.天然气水合物温度-压力-盐浓度三维相平衡曲面方程[J].石油与天然气化工,2019,48(5):49-55.CHUAI Shiyang,WANG Huaning,ZHANG Ning,et al.Three-dimensional phase equilibrium surface equation of temperature-pressure-salt concentration of natural gas hydrate[J].Chemical Engineering of Oil and Gas,2019,48(5):49-55.

    • [7] 耿菲凡,董波,周训,等.盐度对甲烷水合物分解特性影响的LBM模拟[J].计算力学学报,2024,41(5):929-934,962.GENG Feifan,DONG Bo,ZHOU Xun,et al.LBM simulation of the influence of salinity on decomposition characteristics of methane hydrate[J].Chinese Journal of Computational Mechanics,2024,41(5):929-934,962.

    • [8] 汪敏,杨桃,唐洪明,等.迁移深度神经网络的页岩总孔隙度预测[J].西南石油大学学报(自然科学版),2023,45(6):69-79.WANG Min,YANG Tao,TANG Hongming,et al.Prediction of total porosity of shale by migration depth neural network[J].Journal of Southwest Petroleum University(Science & Technology Edition),2023,45(6):69-79.

    • [9] SONG Y,ZHOU H,WANG P,et al.Prediction of clathrate hydrate phase equilibria using gradient boosted regression trees and deep neural networks[J].The Journal of Chemical Thermodynamics,2019,135:86-96.

    • [10] 董晓霞,冯少柯.基于神经网络聚类分析的深层页岩储层岩相识别:以川南筇竹寺组为例[J].西南石油大学学报(自然科学版),2024,46(6):61-73.DONG Xiaoxia,FENG Shaoke.Lithofacies identification in deep shale reservoirs via neural network clustering analysis:a case study of the qiongzhusi formation in the southern Sichuan Basin[J].Journal of Southwest Petroleum University(Science & Technology Edition),2024,46(6):61-73.

    • [11] MEHRIZADEH M.Prediction of gas hydrate formation using empirical equations and data-driven models[J].Materials Today:Proceedings,2021,42:1592-1598.

    • [12] REBAI N,HADJADJ A,BENMOUNAH A,et al.Prediction of natural gas hydrates formation using a combination of thermodynamic and neural network modeling[J].Journal of Petroleum Science and Engineering,2019,182:106270.

    • [13] 于丽丽,刘永红,蔡宝平,等.基于小波神经网络的双电极同步伺服放电加工工艺效果预测[J].开云电竞投注学报(自然科学版),2008,32(4):87-90.YU Lili,LIU Yonghong,CAI Baoping,et al.Effect prediction of dual-electrode synchronous servo discharge machining based on wavelet neural network[J].Journal of China University of Petroleum(Edition of Natural Science),2008,32(4):87-90.

    • [14] ZHAO Y,DU X,XIA G,et al.A novel algorithm for wavelet neural networks with application to enhanced PID controller design[J].Neurocomputing,2015,158:257-267.

    • [15] ALI Y,ALY H H.Short term wind speed forecasting using artificial and wavelet neural networks with and without wavelet filtered data based on feature selections technique[J].Engineering Applications of Artificial Intelligence,2024,133:108201.

    • [16] 王健,徐加放,王博闻,等.基于GA-Elman神经网络的煤层气临界解吸压力预测[J].开云电竞投注学报(自然科学版),2024,48(5):138-145.WANG Jian,XU Jiafang,WANG Bowen,et al.Prediction of critical desorption pressure of coalbed methane based on GA-Elman neural network[J].Journal of China University of Petroleum(Edition of Natural Science),2024,48(5):138-145.

    • [17] 李伟,袁新安,曲萌,等.基于GA-BP神经网络的ACFM实时高精度裂纹反演算法[J].开云电竞投注学报(自然科学版),2016,40(5):128-134.LI Wei,YUAN Xinan,QU Meng,et al.ACFM real-time high-precision crack inversion algorithm based on GA-BP neural network[J].Journal of China University of Petroleum(Edition of Natural Science),2016,40(5):128-134.

    • [18] WU Y,GAO R,YANG J.Prediction of coal and gas outburst:a method based on the BP neural network optimized by GASA[J].Process Safety and Environmental Protection,2020,133:64-72.

    • [19] 李华,荣婕妤.基于AdaBoost方法的近红外光谱建模分析[J].吉林师范大学学报(自然科学版),2022,43(3):83-89.LI Hua,RONG Jieyu.Modeling and analysis of near infrared spectroscopy based on AdaBoost method[J].Journal of Jilin Normal University(Natural Science Edition),2022,43(3):83-89.

    • [20] 肖捡花.交通领域下在线集成时间序列预测方法及应用研究[D].长沙:湖南大学,2020.XIAO Jianhua.Research on online ensemble time series prediction method and application in transportation field[D].Changsha:Hunan University,2020.

    • [21] 邱腾煌,钱玉宝,季威,等.基于GA-BPNN混合智能模型的钻速预测[J].电子测量技术,2024,47(15):177-186.QIU Tenghuang,QIAN Yubao,JI Wei,et al.Prediction of ROP based on GA-BPNN hybrid intelligent model[J].Electronic Measurement Technology,2024,47(15):177-186.

    • [22] 苏莫婷.基于机器学习方法的天然气价格和消费量预测研究[D].重庆:重庆大学,2022.SU Moting.Research on prediction of natural gas price and consumption based on machine learning methods[D].Chongqing:Chongqing University,2022.

    • [23] SLOAN E D Jr,KOH C A,KOH C A.Clathrate hydrates of natural gases[M].Boca Raton:CRC Press,2007:461-519.

    • [24] GAHLEN P,MAINKA R,STOMMEL M.Prediction of anisotropic foam stiffness properties by a neural network[J].International Journal of Mechanical Sciences,2023,249:108245.

    • [25] 宋先知,朱硕,李根生,等.基于BP-LSTM双输入网络的大钩载荷与转盘扭矩预测[J].开云电竞投注学报(自然科学版),2022,46(2):76-84.SONG Xianzhi,ZHU Shuo,LI Gensheng,et al.Prediction of hook load and turntable torque based on BP-LSTM dual-input network[J].Journal of China University of Petroleum(Edition of Natural Science),2022,46(2):76-84.

    • [26] ZHANG X,HOU L,LIU J,et al.Energy consumption prediction for crude oil pipelines based on integrating mechanism analysis and data mining[J].Energy,2022,254:124382.

    • [27] YE K,WANG J,GAO H,et al.Optimization of lapping process parameters of CP-Ti based on PSO with mutation and BPNN[J].The International Journal of Advanced Manufacturing Technology,2021,117(9):2859-2866.

    • [28] CHENG S,WU Y,LI Y,et al.TWD-SFNN:three-way decisions with a single hidden layer feedforward neural network[J].Information Sciences,2021,579:15-32.

    • [29] YAN H,ZHANG J,ZHOU N,et al.Application of hybrid artificial intelligence model to predict coal strength alteration during CO2 geological sequestration in coal seams[J].Science of the Total Environment,2020,711:135029.

  • Baidu
    map