Inflow Profiling and Production Optimization in Smart Wells Using Distributed Acoustic and Temperature Measurements

Advisors

Roland N. Horne, primary advisor
Khalid Aziz, advisor
Tapan Mukerji, advisor

Abstract

Current advances in the well completion technology have allowed for more complex smart well instrumentation with marginal additional cost. As an example, optical fibers can be run along horizontal wells to provide acoustic and temperature data that are distributed both in time and space. With such data at our disposal, an immediate evaluation of the well response is possible as changes occur in reservoir or well conditions. The combination of this continuous monitoring capability and downhole controls in smart wells, facilitates the implementation of efficient well production optimization. Most current work in distributed measurements looks at Distributed Acoustic Sensing (DAS) or Distributed Temperature Sensing (DTS) data individually, which limits inferences about the multiphase flow problem. The objective of this work was to look at the two sets of data together in the multiphase inflow profiling problem. By doing so, we examined what improvements could be achieved and what limitations persist compared to the conventional methods of looking at each inflow profiling method alone. The last research component focused on integrating several optimization procedures that take advantage of such smart completions. The study began by evaluating the performance of DAS in analyzing two-phase flow. This process begins by extracting the speed of sound within the fluid medium from the acoustic signal. Then, the phase fraction combination that corresponds to this speed of sound reading can be estimated. Another procedure was used to obtain similar results from DTS measurements. In this case, the in-situ phase fractions are correlated to the Joule-Thomson effect as reservoir fluids enter the wellbore. As both these procedures are limited to one- and two-phase flow applications, the theoretical background for solutions in three-phase flow problems was established by combining information from DAS and DTS. The flow profiling procedure was applied to several smart well production data sets that included real wells as well as synthetic models. For real single-phase flow examples, flow rates from different segments of the well were calculated and results were in close agreement with a surface flow meter for most sections of the well. For oil-water production examples, we were able to estimate the phase fractions along the well. However, accuracy of DAS results was dependent on the flow regime in the wellbore. In cases where both DAS and DTS were not available for the same well, a commercial compositional and thermal reservoir simulator was used to generate synthetic data for analysis. By applying the developed procedure, we found that cointerpretation of DAS and DTS data improves the profiling performance in two-phase flow and yields fair accuracy for in-situ three-phase fractions for all ranges of water cuts and gas volume fractions. In comparison, analyzing DAS or DTS individually is usually not sufficient to fully determine a three-phase flow problem. When the developed optimization procedure was applied for synthetic models completed with a typical smart well design, results showed that significant value could be realized by incorporating downhole flow rate measurements. For example, continuous downhole flow monitoring provides asset managers with more accurate allocation of their wells. Moreover, more accurate history matching of reservoir models is possible by using in-situ phase flows to calibrate existing models. With more accurate models, evaluating different flow scenarios is possible before applying them in the field. Finally, quick decisions to change the controls of the well are easier with the described optimization method as simple proxy models are trained after collecting more production rate samples. By comparison, full reservoir simulation model optimization takes too long a time to make their use practical in everyday applications.