Data-Driven Reservoir Management of a Giant Mature Oilfield in the Middle East

Abstract

A novel approach to reservoir management applied to a mature giant oilfield in the Middle East is presented. This is a prolific brown field producing from multiple horizons with production data going back to mid-1970s. Periphery water injection in this filed started in mid-1980s. The field includes more than 400 producers and injectors. The production wells are deviated (slanted) or horizontal and have been completed in multiple formations. An empirical, full field reservoir management technology, based on a data-driven reservoir model was used for this study. The model was conditioned to all available types of field data (measurements) such as production and injection history, well configurations, well-head pressure, completion details, well logs, core analysis, time-lapse saturation logs, and well tests. The well tests were used to estimates the static reservoir pressure as a function of space and time. Time-lapse saturation (Pulse-Neutron) logs were available for a large number of wells indicating the state of water saturation in multiple locations in the reservoir at different times. The data-driven, full field model was trained and history matched using machine learning technology based on data from all wells between 1975 and 2001. The history matched model was deployed in predictive mode to generate (forecast) production from 2002 to 2010 and the results was compared with historical production (Blind History Match). Finally future production from the field (2011 to 2014) was forecasted. The main challenge in this study was to simultaneously history match static reservoir pressure, water saturation and production rates (constraining well-head pressure) for all the wells in the field. History matches on a well-by-well basis and for the entire asset is presented. The quality of the matches clearly demonstrates the value that can be added to any given asset using pattern recognition technologies to build empirical reservoir management tools. This model was used to identify infill locations and water injection schedule in this field. RESERVOIR MANAGEMENT Reservoir management has been defined as use of financial, technological, and human resources, to minimizing capital investments and operating expenses and to maximize economic recovery of oil and gas from a reservoir. The purpose of reservoir management is to control operations in order to obtain the maximum possible economic recovery from a reservoir on the basis of facts, information, and knowledge (Thakur 1996). Historically, tools that have been successfully and effectively used in reservoir management integrate geology, petrophysics, geophysics and petroleum engineering throughout the life cycle of a hydrocarbon asset. Through the use of technologies such as remote sensors and simulation modeling, reservoir management can improve production rates and increase the total amount of oil and gas recovered from a field (Chevron 2012). Reservoir simulation and modeling has proven to be one of the most effective instruments that can integrate data and expertise fr...

LSTM Based EFAST Global Sensitivity Analysis for Interwell Connectivity Evaluation Using Injection and Production Fluctuation Data

“…With more than twenty years of application in O&G field, ANN has been considered and utilized as an effective approach for reservoir characterization, management and optimization, such as optimization of well location [35], [36], optimization of oil recovery [37], reservoir monitoring and management [38], [39], and history matching [38], [40], [41].…”

Characterizing Reservoir Connectivity and Forecasting Waterflood Performance Using Data-Driven and Reduced-Physics Models

“…Several areas of application included reservoir characterization (Artun and Mohaghegh 2011;Raeesi et al 2012;Alizadeh et al 2012), candidate well selection for hydraulic fracturing treatments (Mohaghegh et al 1996), well-placement/trajectory optimization (Centilmen et al 1999;Doraisamy et al 2000;Johnson and Rogers 2001;Guyaguler and Horne 2000;Yeten et al 2003;Gokcesu et al 2005;Mohaghegh et al 2006), screening and optimization of secondary/enhanced oil recovery processes (Ayala and Ertekin 2005;Patel et al 2005;Demiryurek et al 2008;Artun et al 2010Artun et al , 2012Parada and Ertekin 2012;Amirian et al 2013), history matching (Cullick et al 2006Silva et al 2007;Zhao et al 2015), reservoir modeling, monitoring and management (Zangl et al 2006;Mohaghegh 2011;Mohaghegh et al 2014;Zhao et al 2015;Kalantari-Dhaghi et al 2015;Esmaili and Mohaghegh 2016). Most of these problems presented in the literature are based on development of artificial neural network (ANN) based proxy models that can accurately mimic reservoir models within a reasonable amount of accuracy and computational efficiency.…”

Modeling, analysis, and screening of cyclic pressure pulsing with nitrogen in hydraulically fractured wells

“…Artificial neural networks (ANN) are very powerful in extracting non-linear and complex relationships between input and output patterns. Several areas of application included reservoir characterization (Artun and Mohaghegh 2011;Raeesi et al 2012;Alizadeh et al 2012;Artun 2016), candidate well selection for hydraulic fracturing treatments (Mohaghegh et al 1996), field development (Centilmen et al 1999;Doraisamy et al 2000;Mohaghegh et al 1996), well-placement and trajectory optimization Rogers 2011, Guyaguler 2002;Yeten et al 2003), scheduling of cyclic steam injection processes (Patel et al 2005), screening and optimization of secondary/enhanced oil recovery (Ayala and Ertekin 2005;Artun et al 2010Artun et al , 2011bArtun et al , 2012Parada and Ertekin 2012;Amirian et al 2013), history matching (Cullick et al 2006;Silva et al 2007;Zhao et al 2015), underground-gas-storage management (Zangl et al 2006), reservoir monitoring and management (Zhao et al 2015;Mohaghegh et al 2014), and modeling of shale-gas reservoirs (Kalantari-Dhaghi et al 2015;Esmaili and Mohaghegh 2015).…”