| Abstract: | Production data analysis and forecasting uses empirical decline curves and model-based rate transient analysis to understand the well performance. Some of the most popular decline curve analysis models used in production forecasting are stretched exponential decline, modified Duong decline, power law decline, and multi-segment Arps decline model. Production forecasting using multi-segment Arps remains a popular choice among the resource evaluators due to its simple application and long-term use since 1945. It divides the production data into three segments allowing the three segments to capture distinct flow regimes, including transient, transition flow, and boundary dominated flow, and generates a forecast of future production rates based on the Arps hyperbolic decline curve model. Although rigorous simulation studies probably provide the most accurate approach for forecasting future production, they could be time consuming, expensive, and may require extensive input data. Due to large number of wells under operation in the Permian basin, a single reservoir engineer may be tasked with analyzing hundreds of wells, and a rigorous reservoir simulation-based analysis for every well is not possible. In such cases, empirical models and type wells can be used to forecast future production. Although a significant number of analytical, semi-analytical, and empirical forecasting methods have been proposed for history matching and forecasting production from multi fractured horizontal wells (MFHWs) in low permeability reservoirs, the simplifying assumptions in their formulation assumes single-phase-flow. Also, only a limited number of empirical models have been published to forecast secondary phase production in unconventional reservoirs. Most of the published research presents complex numerical models focused on specific production periods for forecasting secondary phase production, which could be very difficult to implement for assessing field examples. This work aims to investigate several questions raised by the oil and gas industry for accurate production forecasting of oil and gas in unconventional reservoirs. The project starts by investigating if the conventional DCA models are applicable to the wells producing under high GOR/CGR and provides reserve evaluators with a simple empirical model that can be applied quickly and easily to accurately forecast secondary phase production in tight oil and liquid-rich gas reservoirs experiencing multi-phase flow. I also present the use of dynamic time warping (DTW) algorithm to identify wells with similar production profiles in an objective manner. This workflow is completely automated, and it can be used for efficient and effective identification of wells with similar production characteristics for enhanced type well construction. The electronic version of this dissertation is accessible from https://hdl.handle.net/1969.1/198129 |
|---|