A deep learning approach to anomaly detection in geological carbon sequestration sites using pressure measurements

Abstract Carbon capture and storage (CCS) has been extensively investigated as a potential engineering measure to reduce anthropogenic carbon emission to the atmosphere. Real-time monitoring of the safety and integrity of carbon storage reservoirs is a critical aspect of any commercial-scale CCS deployment. Pressure-based sensing is cost effective, suitable for real-time monitoring, and scalable to large monitoring networks. However, questions remain on how to best harness intelligent information from the high-frequency pressure monitoring sensors to support real-time decisions. This work presents a deep-learning-based framework for analyzing and detecting anomalies in pressure data streams by using a convolutional long short-term memory (ConvLSTM) neural network model, which allows for the fusion of both static and dynamic reservoir data. In ConvLSTM, the convolutional neural network (CNN) is used for spatial pattern mining and the LSTM is used for temporal pattern recognition. The performance of the ConvLSTM model for real-time anomaly detection is demonstrated using a set of pressure monitoring data collected from Cranfield, Mississippi, an active enhanced-oil-recovery field. The anomaly detection model is trained using bottom-hole pressure data acquired from the base experiment (without leak event) and then tested on pressure data collected during a series of controlled CO 2 release experiments (with artificially created leak events). Results show that the ConvLSTM neural network model successfully detected anomalies in the pressure time series obtained from the controlled release experiments. Inclusion of static information into the model further improves the robustness of ConvLSTM.

[1]  Stefan Bachu,et al.  Sequestration of CO2 in geological media: criteria and approach for site selection in response to climate change , 2000 .

[2]  Sally M. Benson,et al.  Rapid detection and characterization of surface CO2 leakage through the real-time measurement of δ13δ13 C signatures in CO2 flux from the ground , 2009 .

[3]  Curtis M. Oldenburg,et al.  An improved strategy to detect CO2 leakage for verification of geologic carbon sequestration , 2005 .

[4]  Henrik Stahl,et al.  Detection of CO2 leakage from a simulated sub-seabed storage site using three different types of pCO2 sensors , 2015 .

[5]  Alexander Y. Sun,et al.  Metamodeling-based approach for risk assessment and cost estimation: Application to geological carbon sequestration planning , 2018, Comput. Geosci..

[6]  Mehdi Zeidouni,et al.  Analytical model of leakage through fault to overlying formations , 2012 .

[7]  Juan Song,et al.  Multimodal Gesture Recognition Using 3-D Convolution and Convolutional LSTM , 2017, IEEE Access.

[8]  Demis Hassabis,et al.  Mastering the game of Go without human knowledge , 2017, Nature.

[9]  R. Stuart Haszeldine,et al.  Carbon Capture and Storage: How Green Can Black Be? , 2009, Science.

[10]  Lei Wang,et al.  An Analytical Model for Assessing Stability of Pre-Existing Faults in Caprock Caused by Fluid Injection and Extraction in a Reservoir , 2016, Rock Mechanics and Rock Engineering.

[11]  Ming Yang,et al.  3D Convolutional Neural Networks for Human Action Recognition , 2010, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[12]  M. Small,et al.  Toward an adaptive monitoring design for leakage risk – Closing the loop of monitoring and modeling , 2018, International Journal of Greenhouse Gas Control.

[13]  Andy Chadwick,et al.  The state of the art in monitoring and verification—Ten years on , 2015 .

[14]  Z. Zhong,et al.  Application of mixed kernels function (MKF) based support vector regression model (SVR) for CO2 - Reservoir oil minimum miscibility pressure prediction , 2016 .

[15]  J. Carrera,et al.  Geologic carbon storage is unlikely to trigger large earthquakes and reactivate faults through which CO2 could leak , 2015, Proceedings of the National Academy of Sciences.

[16]  Dongxiao Zhang,et al.  Assessing leakage detectability at geologic CO2 sequestration sites using the probabilistic collocation method , 2013 .

[17]  J. Rutqvist,et al.  Probabilistic Analysis of Fracture Reactivation Associated with Deep Underground CO2 Injection , 2013, Rock Mechanics and Rock Engineering.

[18]  Mojdeh Delshad,et al.  Modeling and simulation of carbon sequestration at Cranfield incorporating new physical models , 2013 .

[19]  Ryan Noble,et al.  An assessment of near surface CO2 leakage detection techniques under Australian conditions , 2014 .

[20]  Timothy R. Carr,et al.  Dew point pressure prediction based on mixed-kernels-function support vector machine in gas-condensate reservoir , 2018, Fuel.

[21]  Alexander Y. Sun,et al.  A harmonic pulse testing method for leakage detection in deep subsurface storage formations , 2015 .

[22]  Qian Yang,et al.  Building complex event processing capability for intelligent environmental monitoring , 2019, Environ. Model. Softw..

[23]  Jean-Philippe Nicot,et al.  Static and dynamic reservoir modeling for geological CO2 sequestration at Cranfield, Mississippi, U.S.A. , 2013 .

[24]  Yan Liu,et al.  Recurrent Neural Networks for Multivariate Time Series with Missing Values , 2016, Scientific Reports.

[25]  Jef Caers,et al.  Direct forecasting of subsurface flow response from non-linear dynamic data by linear least-squares in canonical functional principal component space , 2014 .

[26]  Yoshua Bengio,et al.  Convolutional networks for images, speech, and time series , 1998 .

[27]  S. Bachu,et al.  Evaluation of the Potential for Gas and CO2 Leakage Along Wellbores , 2009 .

[28]  Timothy R. Carr,et al.  Geostatistical 3D geological model construction to estimate the capacity of commercial scale injection and storage of CO2 in Jacksonburg-Stringtown oil field, West Virginia, USA , 2019, International Journal of Greenhouse Gas Control.

[29]  M. Celia,et al.  Analytical solutions for leakage rates through abandoned wells , 2004 .

[30]  Akand W. Islam,et al.  Using pulse testing for leakage detection in carbon storage reservoirs: A field demonstration , 2016 .

[31]  Katherine D. Romanak,et al.  Improving monitoring protocols for CO2 geological storage with technical advances in CO2 attribution monitoring , 2015 .

[32]  Susan D. Hovorka,et al.  Geochemical sensitivity to CO2 leakage: Detection in potable aquifers at carbon sequestration sites , 2014 .

[33]  Rajesh J. Pawar,et al.  Geologic CO2 sequestration monitoring design: A machine learning and uncertainty quantification based approach , 2018, Applied Energy.

[34]  Alexander Y. Sun,et al.  Inversion of pressure anomaly data for detecting leakage at geologic carbon sequestration sites , 2012 .

[35]  A. Sun,et al.  A laboratory validation study of the time-lapse oscillatory pumping test for leakage detection in geological repositories , 2017 .

[36]  Elizabeth J. Wilson,et al.  Causes and financial consequences of geologic CO2 storage reservoir leakage and interference with other subsurface resources , 2014 .

[37]  Atul K. Jain,et al.  Global Carbon Budget 2015 , 2015 .

[38]  Demis Hassabis,et al.  Mastering the game of Go with deep neural networks and tree search , 2016, Nature.

[39]  M. Wheeler,et al.  Utilization of multiobjective optimization for pulse testing dataset from a CO2-EOR/sequestration field , 2018, Journal of Petroleum Science and Engineering.

[40]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[41]  Susan D. Hovorka,et al.  Simulating the Cranfield geological carbon sequestration project with high-resolution static models and an accurate equation of state. , 2016 .

[42]  Jens T. Birkholzer,et al.  Early detection of brine and CO2 leakage through abandoned wells using pressure and surface-deformation monitoring data: Concept and demonstration , 2013 .

[43]  Alexander Y. Sun,et al.  A learning-based data-driven forecast approach for predicting future reservoir performance , 2018 .

[44]  Trevor Darrell,et al.  Long-term recurrent convolutional networks for visual recognition and description , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[45]  David J. Young,et al.  Wellbore integrity and corrosion of carbon steel in CO2 geologic storage environments: A literature review , 2013 .

[46]  Susan D. Hovorka,et al.  Monitoring a large-volume injection at Cranfield, Mississippi—Project design and recommendations , 2013 .

[47]  Eduard H. Hovy,et al.  End-to-end Sequence Labeling via Bi-directional LSTM-CNNs-CRF , 2016, ACL.

[48]  Dmitri Kavetski,et al.  Development of a hybrid process and system model for the assessment of wellbore leakage at a geologic CO2 sequestration site. , 2008, Environmental science & technology.

[49]  Dit-Yan Yeung,et al.  Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting , 2015, NIPS.