Abstract
Hydraulic fracturing of long horizontal wells has become a standard for development of unconventional reservoirs. Fiber-optic based monitoring techniques have been used for optimization of completions design. In addition, production logging can provide critical spatial information on well productivity and can help optimize cost for unconventional reservoirs. Wells in unconventional reservoirs can be low-rate, and have hundreds of spaced entry points which make production logging a challenge using conventional tools. We developed a novel method that solely relies on fiber-optic sensing measurements to provide production logging results with uncertainty estimations. The method uses Distributed Temperature Sensing to measure borehole temperatures during steady states and uses Distributed Acoustic Sensing to measure borehole flow velocities by tracking temperature slugging signals during transient states. A Markov-Chain-Monte-Carlo based stochastic inversion is applied to find the statistical distribution of possible total fluid production allocations that fit both temperature and velocity measurements. While temperature-based methods have been used in the industry with recognized uncertainties, our concept of using DAS to measure flow rates has been verified using flowloop experiments. The workflow using both DAS and DTS has been tested through synthetic models and on data collected from an unconventional oil producer.
Introduction
Conventional production logging tools measure the flow velocity and fluid phase inside a producing well bore to estimate the productivity distribution along the well (e.g. Hill, 1990). It provides valuable information for production problem diagnosis, reservoir surveillance, and completion optimization. Production logging measurements are usually made by using spinner-based flow meters and other sensors, which require the tool to make multiple trips through the well section of interest (e.g. Chadwick and Whittaker, 2011; Zhu et al., 2013). There are several problems associated with this technology. First, the tool set may partially block production flows and alter production allocations during the survey. Second, the tool requires borehole cleanup before the survey to prevent sensor damage, which significantly increases the cost. Finally, the measurements are less reliable in horizontal wells due to fluid segregation (Vu-Hoang et al., 2004).
