Clarkson, Christopher R.Ahmadi, Hossein2022-08-112022-08-112022-08-08Ahmadi, H. (2022). Analysis of Production Data from Communicating Multi-Fractured Horizontal Wells (Doctoral thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca.http://hdl.handle.net/1880/114929https://dx.doi.org/10.11575/PRISM/39975The most common technology used for the development of low-permeability (unconventional) oil and gas reservoirs is multi-fractured horizontal wells (MFHWs), which are designed to maximize fracture surface area in contact with the reservoir. In current development practice, multiple MFHWs are typically drilled from the same pad; over time, fracture and well (lateral) spacing have decreased, resulting in greater interference between wells (vertically and laterally). This interference, which commonly occurs through hydraulic fractures, can affect well performance, and complicate reservoir and fracture characterization efforts. Rate-transient analysis (RTA) is a reservoir engineering approach that is commonly used to derive reservoir/fracture properties for MFHWs producing from unconventional reservoirs. However, RTA methods/models are generally applicable to single wells and do not specifically account for inter-well communication through hydraulic fractures. It is the purpose of this dissertation to develop RTA methods/models that account for inter-well communication. In order to account for inter-well communication through fractures, both semi-analytical and machine learning approaches are developed for RTA of MFHWs. The semi-analytical approach developed herein utilizes the dynamic drainage area (DDA) concept. Two and three reservoir/fracture region semi-analytical DDA models are first developed to history match and forecast two MFHWs communicating through hydraulic fractures. These models, which account for two-phase flow (dry gas, water), and pressure-dependent porosity and permeability, are successfully verified with fully-numerical simulation results. Their practical applicability is also demonstrated using a field dataset consisting of six wells (drilled from two neighboring pads) exhibiting different degrees of communication. The final semi-analytical model developed is in the form of a new straight-line analysis (RTA) method, referred to as the “DDA-corrected” RTA model. When this method is applied to two wells that are communicating through hydraulic fractures, and are put on production at different times, two straight lines can be obtained for the first well that comes on production. Because the method assumes that the wells are fully connected, the slope of the first straight line (i.e., before the second well is put on production) corresponds to an equivalent fracture whose half-length is a portion of the total summation of individual fracture half-lengths depending on the permeabilities of individual fractures. This method, which also accounts for two-phase flow (gas, water), and pressure-dependent porosity and permeability, is successfully verified against fully-numerical simulation results, and its practical applicability is demonstrated with the same field case as the history match/forecasting models. Finally, machine learning (ML) algorithms, along with an optimization method, are employed to develop a history matching tool for three-phase flow of oil, gas and water associated with two communicating wells (staggered production). The basic concept used in ML algorithm training is that a change in the slope of the square-root of time plot for the first (parent) well occurs due to the second well coming on production. A dataset consisting of synthetic simulation cases is generated by varying some of the reservoir and fracture parameters to produce all the possible production scenarios for a two-well base model. After preprocessing the outliers, ML models are trained and tested. These models can generalize the relationship between the reservoir and fracture parameters and the square-root of time plot slope values for the parent well. The goal of developing a history matching tool is then attained by using the best performing ML model and an optimization algorithm which can estimate the reservoir and fracture parameters for a set of slopes obtained from the square-root of time plot of the parent well. Fundamentally, this dissertation advances RTA methods that can be applied to two wells that are in communication through hydraulic fractures.engUniversity of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission.reservoir engineeringpetroleum engineeringunconventional reservoirsrate transient analysismulti-fractured horizontal wellsdata analyticsEnergyEngineering--Petroleumdoctoral thesis