EDFM-AI Automatic History Matching (AHM)
Figure 2: Non-intrusive EDFM technology bridging Fracture simulators and reservoir simulators.
SimTech AHM technology is now available and has been implemented in real field applications for assessing operators about future improved fracture designs and even new well placement optimization.
Actual Field Well applications
How does EDFM-AI work? (AHM Framework)
SimTech LLC offers the best solutions to the challenging needs of our clients by translating geoscience fracture evaluation output into multimillion decision field applications. We believe that synergy between geosciences and engineering is essential for understanding current field fracturing practices and optimizing their outcome in terms of net oi/gas production and investment returns. This integration cannot escape from the accelerating digital transformation of unconventional reservoir characterization -- and SimTech LLC has taken the lead with the EDFM technology. In that sense, SimTech has developed its own Automatic History Matching (AHM) technology called EDFM-AI to bridge the information from these disciplines and generate fast physics-based answers for the oil and gas field current inquiries.
Our AHM technology is a very powerful tool to obtain multiple reservoir realizations unlike traditional History Matching with single solutions which can be time-consuming and too risky for multimillion investment decisions. EDFM-AI can reduce the subsurface uncertainties and estimate production forecast reliably in a few days. These multiple HM solutions are quickly found by our EDFM-AI and are essential to improve the calibration of hydraulic fracture designs, estimate the impact of natural fractures, and provide assessment for infill well placement and optimal cluster spacing.x.
Our AHM Framework is based on a practical and efficient iterative workflow that can handle unconventional reservoirs complexities and that integrates four main stages: (1) EDFM preprocessing for the best fracture characterization, (2) Reservoir simulation for dynamic multiphase reservoir characterization (any commercial simulator can be used), (3) Neural Network application for generating Proxy Model based on uncertainties (sampling stochastic reservoir model realizations), and (4) Proxy-based Markov Chain Monte Carlo (MCMC) Algorithm for screening the best stochastic HM solutions based on multiple objective functions. (See Figure 1)
Our technology is based on Embedded Discrete Fracture Modeling (EDFM), which constitutes the optimum tool to model fractures in unconventional reservoirs in terms of accuracy, flexibility, and computational time efficiency when compared to traditional DPDK, LGR, or unstructured methods. Some EDFM applications showed simulation results about 20-times faster than conventional LGR. Essentially, EDFM can transfer 3rd party fracture simulator directly to 3rd party reservoir simulator like no other technology in the market (Figure 2), providing a smooth translation among Geoscience, Petrophysics, Well Completions, Geology into Reservoir, Production Engineering and Economics.
Figure 1: AHM Integrated Framework diagram