Development and Application of Ensemble Kalman Filter for Efficient Production Data Analysis and Accurate Property Estimation
- Wenting Yue (University of North Dakota) | John Yilin Wang (The Pennsylvania State University)
- Document ID
- Society of Petroleum Engineers
- SPE Western Regional Meeting, 22-26 April, Garden Grove, California, USA
- Publication Date
- Document Type
- Conference Paper
- 2018. Society of Petroleum Engineers
- 1.8 Formation Damage
- Production data analysis, Ensemble Kalman Filter, Field application, Reservoir simulation and history matching
- 0 in the last 30 days
- 75 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 8.50|
|SPE Non-Member Price:||USD 25.00|
Single-well production data analysis (PDA) is an important subject to understand reservoir productions. This is commonly done with traditional methods/analytical models. But analytical models suffer from limited accuracy and applicability issues due to the way production data is matched, the error involved, and the assumptions that sometimes over simplifying the problem. Therefore, in this work, the authors want to provide an assessment of Ensemble Kalman Filter (EnKF)-based production data analysis model on single-well production.
By assuming homogeneous and isotropic reservoir permeability in a single-layer reservoir, we first formulate the basic EnKF algorithm and link it with single-well reservoir model. Results indicate estimation of skin factor and reservoir permeability present accuracy issues: large error for some cases and uncertainty bounds do not cover true values. Based on our evaluations, we propose the method of increased initial uncertainty bound and over estimating initial observation noise variance to improve the estimation. The method is tested with synthetic models, and results indicate that mean estimate have better match with true value, and the true values fall within a reasonable uncertainty bounds of the ensemble data predictions. Production data from 31 wells in real field are used for further verification. Excellent data matches are obtained with EnKF.
The model in the work could provide a reasonable estimation of reservoir properties for both synthetic and real-field cases. We also show that statistical inconsistency and poor data matches are encountered when matching production data for some extreme cases when permeability of damaged/stimulated zone is drastically different from reservoir permeability. But this issue could be alleviated with the proposed method. The model and method in this study proves to be applicable for real field evaluation.
We present readers with an implementation of EnKF-based in single-well production data analysis to overcome accuracy and applicability issues related with traditional analytical methods. We documented the accuracy and efficiency one could expect when applying this method in both synthetic models and real-field data to evaluate skin factor, drainage area, and permeability. We also proposed and verified a methodology to improve estimation accuracy under some extreme cases when estimation of skin factor possess a problem. This paper could provide a guild to the readers when constructing their own production data analysis model.
|File Size||2 MB||Number of Pages||21|
Al Haddad S, Mancini EA (2013) Reservoir characterization, modeling, and evaluation of Upper Jurassic Smackover microbial carbonate and associated facies in Little Cedar Creek field, southwest Alabama, eastern Gulf coastal plain of the United States. Am Assoc Pet Geol Bull 97:2059–2083. doi: 10.1306/07081312187
Arps JJ (1945) Analysis of Decline Curves. Trans AIME 160:228–247. doi: 10.2118/945228-G
Bhattacharya S, Nikolaou M (2013) Analysis of Production History for Unconventional Gas Reservoirs With Statistical Methods. SPE J 18:878–896. doi: 10.2118/147658-PA
Burgers G, Jan van Leeuwen P, Evensen G (1998) Analysis Scheme in the Ensemble Kalman Filter. Mon Weather Rev 126:1719–1724. doi: 10.1175/1520-0493(1998)126<1719:ASITEK>2.0.CO;2
Evensen G (1994) Inverse methods and data assimilation in nonlinear ocean models. Phys D Nonlinear Phenom 77:108–129. doi: 10.1016/0167-2789(94)90130-9
Han M, Li G, Chen J (2014) Assimilating Microseismic and Well-Test Data by Use of EnKF for Accurate Reservoir Characterization. SPE J Preprint:. doi: 10.2118/171554-PA
Houtekamer PL, Mitchell HL (1998) Data Assimilation Using an Ensemble Kalman Filter Technique. Mon Weather Rev 126:796–811. doi: 10.1175/1520-0493(1998)126<0796:DAUAEK>2.0.CO;2
Li G, Han M, Banerjee R, Reynolds AC (2010) Integration of Well-Test Pressure Data Into Heterogeneous Geological Reservoir Models. SPE Reserv Eval Eng 13:496–508. doi: 10.2118/124055-PA
Mancini EA, Parcell WC, Ahr WM, et al (2008) Upper Jurassic updip stratigraphic trap and associated Smackover microbial and nearshore carbonate facies, eastern Gulf coastal plain. Am Assoc Pet Geol Bull 92:417–442. doi: 10.1306/11140707076
Naevdal G, Johnsen L, Aanonsen S, Vefring E (2005) Reservoir Monitoring and Continuous Model Updating Using Ensemble Kalman Filter. SPE J 10:66–74. doi: 10.2118/84372-PA
Yue W, Wang JY (2015) Feasibility of Waterflooding for a Carbonate Oil Field Through Whole-Field Simulation Studies. J Energy Resour Technol 137:64501. doi: 10.1115/1.4030401