针对传统注水开发油藏产量预测方法存在的问题,提出了基于人工神经网络的预测模型,阐述了模拟工作流程,并进行了算例分析。提出了基于流体物理学和测量数据随机组合的特征提取方法,以提高模型的预测效果。优选贝叶斯正则化算法作为模型的训练算法,该算法一般耗时较长,但能对产油量、产气量、产水量等嘈杂数据集进行良好泛化。通过计算均方误差及决定系数、绘制误差分布直方图及模拟数据-验证数据交会图等方式进行模型评价。用90%的历史数据训练、验证、测试目标模型结构,然后用其余10%数据进行盲测。研究表明,提出的流体产量预测模型决定系数超过0.9,模拟结果与实际数据吻合程度高,输入信息少,计算成本低。图20表8参34
As the conventional prediction methods for production of waterflooding reservoirs have some drawbacks, a production forecasting model based on artificial neural network was proposed, the simulation process by this method was presented, and some examples were illustrated. A workflow that involves a physics-based extraction of features was proposed for fluid production forecasting to improve the prediction effect. The Bayesian regularization algorithm was selected as the training algorithm of the model. This algorithm, although taking longer time, can better generalize oil, gas and water production data sets. The model was evaluated by calculating mean square error and determination coefficient, drawing error distribution histogram and the cross-plot between simulation data and verification data etc. The model structure was trained, validated and tested with 90% of the historical data, and blindly evaluated using the remaining. The predictive model consumes minimal information and computational cost and is capable of predicting fluid production rate with a coefficient of determination of more than 0.9, which has the simulation results consistent with the practical data.
[1] ZHONG H, WU K, JI D, et al.Using least square support vector machines to approximate single phase flow[R]. SPE 185881-MS, 2017.
[2] NEGASH B M.Artificial neural network and inverse solution method for assisted history matching of a reservoir model[J]. International Journal of Applied Engineering Research, 2017, 12: 2952-2962.
[3] WEI Chenji, WANG Yuhe, DING Yutao, et al.Uncertainty assessment in production forecasting and optimization for a giant multi-layered sandstone reservoir using optimized artificial neural network technology[R]. SPE 186069-MS, 2017.
[4] NEGASH B M, VASANT P M, JUFAR S R.Application of artificial neural networks for calibration of a reservoir model[J]. Intelligent Decision Technologies, 2018, 12(2): 1-13.
[5] NEGASH B M.History matching of the PUNQ-S3 reservoir model using proxy modeling and multi-objective optimizations[R]. Kuala Lumpur, Malaysia: 2016 International Conference on Industrial Engineering and Operations Management, 2016.
[6] WANTAWIN M, YU W, DACHANUWATTANA S, et al.An iterative response-surface methodology by use of high-degree-polynomial proxy models for integrated history matching and probabilistic forecasting applied to shale-gas reservoirs[R]. SPE 187938-PA, 2017.
[7] WANG D, NIU D, LI H A.Predicting waterflooding performance in low-permeability reservoirs with linear dynamical systems[R]. SPE 185960-PA, 2017.
[8] TIAN C, HORNE R N.Inferring interwell connectivity using production data[R]. SPE 181556-MS, 2016.
[9] SAPUTELLI L.Technology focus: Petroleum data analytics[J]. Journal of Petroleum Technology, 2017, 69(10): 87.
[10] MOHAGHEGH S D.Shale analytics: Data-driven analytics in unconventional resources[M]. Berlin: Springer International Publishing, 2017.
[11] MIŠKUF M. Data mining in cloud usage data with Matlab’s statistics and machine learning toolbox[R].Herl’any, Slovakia: 2017 IEEE 15th International Symposium on Applied Machine Intelligence and Informatics (SAMI), 2017.
[12] ANIFOWOSE F A, LABADIN J, ABDULRAHEEM A.Ensemble machine learning: An untapped modeling paradigm for petroleum reservoir characterization[J]. Journal of Petroleum Science and Engineering, 2017, 151: 480-487.
[13] XIE Y, ZHU C, ZHOU W, et al.Evaluation of machine learning methods for formation lithology identification: A comparison of tuning processes and model performances[J]. Journal of Petroleum Science and Engineering, 2018, 160: 182-193.
[14] FULFORD D S, BOWIE B, BERRY M E, et al.Machine learning as a reliable technology for evaluating time/rate performance of unconventional wells[J]. SPE Economics & Management, 2016, 8(1): 23-39.
[15] CAO Q.Data driven production forecasting using machine learning[R]. Buenos Aires, Argentina: SPE Argentina Exploration and Production of Unconventional Resources Symposium, 2016.
[16] LI Y, HAN Y.Decline curve analysis for production forecasting based on machine learning[R]. SPE 189205-MS, 2017.
[17] TEMIZEL C, NABIZADEH M, KADKHODAEI N, et al.Data-driven optimization of injection/production in waterflood operations[R]. SPE 187468-MS, 2017.
[18] RAMIREZ A M.Prediction of PVT properties in crude oil using machine learning techniques (MLT)[R]. Buenos Aires, Argentina: SPE Latin America and Caribbean Petroleum Engineering Conference, 2017.
[19] MOGHADASI L, RANAEE E, INZOLI F, et al.Petrophysical well log analysis through intelligent methods[R]. SPE 185922-MS, 2017.
[20] ZHENG J, LEUNG J Y, SAWATZKY R P, et al.An AI-based workflow for estimating shale barrier configurations from SAGD production histories[R]. SPE 184984-MS, 2017.
[21] AZIZI S, AHMADLOO E.Prediction of void fraction for gas-liquid flow in horizontal, upward and downward inclined pipes using artificial neural network[J]. International Journal of Multiphase Flow, 2016, 87: 35-44.
[22] MA Z, LEUNG J Y, ZANON S.Integration of artificial intelligence and production data analysis for shale heterogeneity characterization in steam-assisted gravity-drainage reservoirs[J]. Journal of Petroleum Science and Engineering, 2017, 163: 139-155.
[23] KIM P.MATLAB deep learning: With machine learning, neural networks and artificial intelligence[M]. Berkeley, CA: Apress, 2017.
[24] ARAS S, KOCAKOÇ İ D.A new model selection strategy in time series forecasting with artificial neural networks: IHTS[J]. Neurocomputing, 2016, 174: 974-987.
[25] LÓPEZ-IÑESTA E, GRIMALDO F, AREVALILLO-HERRÁEZ M. Combining feature extraction and expansion to improve classification based similarity learning[J]. Pattern Recognition Letters, 2017, 93: 95-103.
[26] SHI Xiancheng, FENG Yucheng, ZENG Jinsong, et al.Chaos time-series prediction based on an improved recursive levenberg-marquardt algorithm[J]. CHAOS SOLITONS AND FRACTALS, 2017, 100: 57-61.
[27] XIAN Zhang, AWOTUNDE A A.Improvement of Levenberg-Marquardt algorithm during history fitting for reservoir simulation[J]. Petroleum Exploration and Development, 2016, 43(5): 876-885.
[28] UTAMI E, WIBAWA A, BIYANTO T R, et al. Reservoir zone prediction using logging data - multi well based on levenberg-marquardt method[R]. Langkawi, Malaysia: 2017 IEEE Symposium on Computer Applications & Industrial Electronics (ISCAIE), 2017.
[29] SIMM J, ARANY A, ZAKERI P, et al. Macau: Scalable Bayesian factorization with high-dimensional side information using MCMC[R]. Tokyo:2017 IEEE 27th International Workshop on Machine Learning for Signal Processing (MLSP), 2017.
[30] GUÉRILLOT D R, BRUYELLE J. Uncertainty assessment in production forecast with an optimal artificial neural network[R]. SPE 183921-MS, 2017.
[31] LU P, LIU H, SERRATELLA C, et al.Assessment of data-driven, machine learning techniques for machinery prognostics of offshore assets[R]. OTC 27577-PT, 2017.
[32] AMIRIAN E, DEJAM M, CHEN Z, et al.Performance forecasting for polymer flooding in heavy oil reservoirs[J]. Fuel, 2018, 216: 83-100.
[33] NEGASH B M, TUFA L D, RAMASAMY M, et al.System identification based proxy model of a reservoir under water injection[J]. Modelling and Simulation in Engineering, 2017(1):1-10.
[34] HAGAN M T.Neural network design[M]. Boulder, Colorado, USA: Campus Publishing Service, University of Colorado, 2002.
