The increasing deployment of distributed temperature and pressure measuring devices in intelligent well completions is providing the means to monitor the inflow profiles of wells without any well intervention. If the profiles of pressure and/or temperature are affected by the inflow profiles of the various phases being produced, it is possible to estimate these flow profiles by inverting the measured temperature and pressure profiles. This inversion process is particularly challenging for horizontal wells because the pressure drop along the well is usually small, and temperature changes, caused primarily by Joule-Thomson effects, are also small.
This paper presents an inversion method that interprets distributed temperature and pressure data to obtain flow rate profiles along horizontal wells. The inversion method, which is based on the Levenberg-Marquardt algorithm, is applied to minimize the differences between the measured profiles and the profiles calculated from a forward model of the well and reservoir flow system.
We present synthetic and field examples in this paper to illustrate how to use the inversion model to interpret the flow profile of a horizontal well. The synthetic examples show that even with single-phase oil production, the inflow profile can be estimated in many cases with the inversion method developed. The method is even more robust when water or gas is produced along discrete intervals in an oil production well because of the unique temperature signature of water or gas production. We applied the inversion method to temperature and pressure profiles measured with production logs in a North Sea horizontal oil producing well. The method successfully inverted pressure and temperature profiles and the profiles of oil and water flow rates determined compared well with the flowmeter derived profiles.
Introduction
In the past decades, thousands of wells have been drilled horizontally and in multiple directions to obtain larger contact volume with the reservoir. Because of the growing complexities of the recent well trajectories, running conventional production monitoring tools on appropriate locations has become difficult and costly. Flow rate, pressure, and temperature are the principle parameters we wish to measure through production logging. For the pressure and temperature measurements, continuous profiles of these in a complex well can be obtained accurately and inexpensively due to the advanced technology of fiber optics. Since the first fiber optic sensor was implemented in a well in Shell's Sleen Field in 19931, the use of distributed temperature sensors (DTS) and distributed pressure sensors (DPS) has become increasingly common for monitoring producing sections of horizontal wells2–4.
As for the flow rate measurement, metering flow rate is still difficult especially under the multiphase flow conditions that occur in most wells. For multi-phase flowing wells, despite the recent advancements in technologies and equipments, a comprehensive solution to measuring flow rates and holdups of the phases is evasive5. However, to take full advantages of intelligent wells, which can control inflow capacities from different producing sections without interventions, real-time monitoring of the downhole flow conditions such as flow rate profiles and locations of excessive water or gas influx is essential. Therefore, to realize the value of intelligent wells, downhole flow conditions are either measured or translated from measurable parameters (e.g. density, pressure, and/or temperature) in horizontal, multi-lateral, or multi-branching wells. Some interpretation studies have shown how production profiles can be obtained by matching such data6,7.
This research proposes an inversion method to obtain downhole inflow conditions from temperature and pressure profile data. We set the parameters to be estimated as productivities or inflow rates of each segment. From continuous temperature and pressure data along the well, we invert them into the desired inflow rates by applying the Levenberg-Marquardt algorithm8.