Drilling fault classification based on pressure and flowrate responses via ensemble classifier in Managed pressure drilling

Abstract Managed pressure drilling (MPD) has a promising prospect in complicated formation where drilling faults always happen. Automatical drilling fault classification is one of the key issues for MPD to effectively reduce drilling cost. In this paper, an unscented kalman filter based ensemble classifier (UKF-EC) based on pressure and flowrate responses is presented to achieve drilling fault classification automatically. According to the relationship between different drilling faults and characteristic parameters, a novel feature extraction strategy to weaken the influence of characteristic parameters’ reference value is proposed, and pressure-loss factors and flowrate factor which indicate the changes caused by drilling faults are chosen to be the input features of UKF-EC. The UKF-EC, which consists of a number of unscented kalman filter based individual classifiers (UKF-ICs) with identical topology trained by different sets of training data, is established to improve its generalization ability to abnormal data at various well depths. Laboratory experiment has verified the effectiveness of UKF-EC for drilling fault classification. Numerical simulation indicates that the UKF-EC is more accurate than the UKF-IC for abnormal data at various well depths and the number of UKF-ICs in the UKF-EC has great influence on accuracy. Via this ensemble classifier UKF-EC, drilling faults can be classified quickly and drilling efficiency can be enhanced.

[1]  Xiaodong Luo,et al.  Estimating observation error covariance matrix of seismic data from a perspective of image denoising , 2016, Computational Geosciences.

[2]  Shahab D. Mohaghegh,et al.  Design and Development of An Artificial Neural Network for Estimation of Formation Permeability , 1995 .

[3]  C. L. Dalton,et al.  The Benefits of Real-Time DownHole Pressure and Tension Data With Wired Composite Tubing , 2002 .

[4]  Chen Lu,et al.  Fault diagnosis for rotary machinery with selective ensemble neural networks , 2017, Mechanical Systems and Signal Processing.

[5]  Stuart L. Scott,et al.  A Neural Network Model for Prediction of Liquid Holdup in Two-Phase Horizontal Flow , 2002 .

[6]  Shuzhi Sam Ge,et al.  Sea state identification based on vessel motion response learning via multi-layer classifiers , 2018 .

[7]  Xin Yao,et al.  Diversity creation methods: a survey and categorisation , 2004, Inf. Fusion.

[8]  Turgay Ertekin,et al.  An Artificial Neural Network Based Relative Permeability Predictor , 2003 .

[9]  Hao Li,et al.  Gas kick during carbonate reservoirs drilling and its risk assessment , 2017 .

[10]  Tao Zhang,et al.  An innovative diagnosis method for lost circulation with unscented Kalman filter , 2018, Journal of Petroleum Science and Engineering.

[11]  Shuai Zhang,et al.  A new pattern recognition model for gas kick diagnosis in deepwater drilling , 2018, Journal of Petroleum Science and Engineering.

[12]  Youmin Tang,et al.  Sigma-Point Kalman Filter Data Assimilation Methods for Strongly Nonlinear Systems , 2009 .

[13]  Alípio Mário Jorge,et al.  Ensemble approaches for regression: A survey , 2012, CSUR.

[14]  Mike Nicolai,et al.  Simulation-driven machine learning: Bearing fault classification , 2018 .

[15]  H. A. Malki,et al.  Estimating Permeability by Use of Neural Networks in Thinly Bedded Shaly Gas Sands , 1996 .

[16]  Vikram Garaniya,et al.  Risk analysis of deepwater drilling operations using Bayesian network , 2015 .

[17]  Agnar Aamodt,et al.  Detection of Symptoms for Revealing Causes Leading to Drilling Failures , 2013 .

[18]  Ahmed K. Abbas,et al.  Intelligent decisions to stop or mitigate lost circulation based on machine learning , 2019, Energy.

[19]  Hongwei Yang,et al.  A new method for early gas kick detection based on the consistencies and differences of bottomhole pressures at two measured points , 2019, Journal of Petroleum Science and Engineering.

[20]  Fulei Chu,et al.  Deep convolutional neural network based planet bearing fault classification , 2019, Comput. Ind..

[21]  Benoit Daireaux,et al.  Early Symptom Detection on the Basis of Real-Time Evaluation of Downhole Conditions: Principles and Results From Several North Sea Drilling Operations , 2012 .

[22]  Jun Li,et al.  Model based fault diagnosis for drillstring washout using iterated unscented Kalman filter , 2019, Journal of Petroleum Science and Engineering.

[23]  Hamid Parvin,et al.  Proposing a classifier ensemble framework based on classifier selection and decision tree , 2015, Eng. Appl. Artif. Intell..

[24]  Jay H. Lee,et al.  Fault detection and classification using artificial neural networks , 2018 .

[25]  Hongwei Yang,et al.  A transient hydro-thermo-bubble model for gas kick simulation in deep water drilling based on oil-based mud , 2019, Applied Thermal Engineering.

[26]  Faisal Khan,et al.  Real time risk analysis of kick detection: Testing and validation , 2017, Reliab. Eng. Syst. Saf..

[27]  Steve William Nas,et al.  Advantages of Managed Pressure Drilling and the Recent Deployment of the Technology in Vietnam , 2010 .

[28]  R. A. Startzman,et al.  Neural Network Approach Predicts U.S. Natural Gas Production , 2003 .

[29]  John-Morten Godhavn,et al.  Control Requirements for Automatic Managed Pressure Drilling System , 2010 .

[30]  Nikunj C. Oza,et al.  Online Ensemble Learning , 2000, AAAI/IAAI.

[31]  Mogens Blanke,et al.  Incident detection and isolation in drilling using analytical redundancy relations , 2015 .

[32]  Don M. Hannegan.  Managed Pressure Drilling in Marine Environments - Case Studies , 2005 .

[33]  Yuanxin Wu,et al.  Comments on "Performance evaluation of UKF-based nonlinear filtering" , 2007, Autom..

[34]  Mogens Blanke,et al.  Fault diagnosis of downhole drilling incidents using adaptive observers and statistical change detection , 2015 .

[35]  Erwin Duvan Gomez,et al.  Successful Use of Managed Pressure Drilling (MPD) in Low Pressure, High Pressure, and Deeper Reservoirs in Mexico South , 2012 .

[36]  Tommy W. S. Chow,et al.  Weighted local and global regressive mapping: A new manifold learning method for machine fault classification , 2014, Eng. Appl. Artif. Intell..

[37]  Shujie Liu,et al.  An adaptive pattern recognition method for early diagnosis of drillstring washout based on dynamic hydraulic model , 2019 .

[38]  Tao Zhang,et al.  Numerical simulation of a new early gas kick detection method using UKF estimation and GLRT , 2019, Journal of Petroleum Science and Engineering.

[39]  Glenn-Ole Kaasa,et al.  A Framework for Fault Diagnosis in Managed Pressure Drilling Applied to Flow-Loop Data , 2013 .

[40]  Zhiqiang Ge,et al.  K-means Bayes algorithm for imbalanced fault classification and big data application , 2019, Journal of Process Control.

[41]  Mogens Blanke,et al.  Early Pack-Off Diagnosis in Drilling Using an Adaptive Observer and Statistical Change Detection , 2015 .

[42]  William Cox,et al.  SMART Kick Detection; First Step on the Well Control Automation Journey , 2015 .

[43]  Ioannis Hatzilygeroudis,et al.  Recognizing emotions in text using ensemble of classifiers , 2016, Eng. Appl. Artif. Intell..

[44]  W. W. Wylie,et al.  Drilling Kick Statistics , 1990 .

[45]  I. M. Moroz,et al.  Ensemble Kalman filter with the unscented transform , 2009, 0901.0461.

[46]  Fei Zhang,et al.  Inter-patient heartbeat classification based on region feature extraction and ensemble classifier , 2019, Biomed. Signal Process. Control..

[47]  Diego Cabrera,et al.  Observer-biased bearing condition monitoring: From fault detection to multi-fault classification , 2016, Eng. Appl. Artif. Intell..

[48]  Kaveh Ashenayi,et al.  Application of Artificial Neural Network to Pump Card Diagnosis , 1994 .

[49]  Zhongxiao Peng,et al.  Integrated model of BP neural network and CNN algorithm for automatic wear debris classification , 2019, Wear.

[50]  Lars Imsland,et al.  Drillstring Washout Diagnosis Using Friction Estimation and Statistical Change Detection , 2015, IEEE Transactions on Control Systems Technology.

[51]  Chang Ouk Kim,et al.  A Convolutional Neural Network for Fault Classification and Diagnosis in Semiconductor Manufacturing Processes , 2017, IEEE Transactions on Semiconductor Manufacturing.

[52]  Zhenhua Rui,et al.  Complex Lithofacies Identification Using Improved Probabilistic Neural Networks , 2018 .

[53]  Lars Imsland,et al.  Fault diagnosis in managed pressure drilling using nonlinear adaptive observers , 2013, 2013 European Control Conference (ECC).

[54]  Mahesh Chandra,et al.  An improved wavelet-based signal-denoising architecture with less hardware consumption , 2019 .

[55]  R. A. Arehart,et al.  Drill-Bit Diagnosis With Neural Networks , 1990 .

[56]  Don Reitsma.  Development of an Automated System for the Rapid Detection of Drilling Anomalies using Standpipe and Discharge Pressure , 2011 .