Characterization of Production through a Fracture Cell Using Acoustic Data

Abstract

Abstract Distributed acoustic sensing (DAS) is an emerging technology heralded as a means of direct interrogation of flow distribution, a diagnosis tool for stimulation operation success, and production distribution in horizontal multi-stage fractured wells. A challenge that has come about from the use of distributed acoustic sensing system is processing, interpretation, and diagnosis of the vast amounts of data produced. Unlike distributed temperature sensing, the data provided by DAS is originally presented in sound amplitude as a function of time. For a defined time interval, the data is first translated to amplitude as a function of frequency, and then the characteristics of both amplitude and frequency is related to flow conditions. Using the characteristics of amplitude and frequency to understand flow distribution is mostly qualitative in the field today. This study started with building a physical simulator in the lab and flowing nitrogen gas through a parallel-plate fracture cell with proppant packs of different permeabilities. A horizontal wellbore section was connected to the fracture cell through a perforation to simulate fractured well flow. Data was collected using an array of microphones distributed along the length of the simulated horizontal wellbore and a directional hydrophone at the toe of the simulated wellbore. The amplitude-frequency characteristic of this acoustic signal is obtained from raw measurements of acoustic pressure via Fast Fourier Transform. A relationship between flow rate and sound pressure level is proposed based on the observation of experiments. With the lab observation, we then modeled acoustic wave behavior in a domain consisting of a perforation and a pipe. Computational fluid dynamics simulation was carried out to simulate the gas flow through the perforation, and the simulated acoustic signals were compared with the lab-observed signals. Modeling of the fluid flow is subdivided into two parts: fluid flow through the proppant pack near the perforation, and fluid flow through the perforation into the wellbore. The proppant pack is described as a set of spherical particles which allows creation of an analogy of the porous structure. In the second part of the simulation, we assume steady state fluid flow through the perforation into the wellbore. This fluid flow creates a noise which is measured along the center of the wellbore. Different types of proppant packing were considered during simulation of the first part of the model. A two-size mesh model was used as the most complex type of packing in this research. Simulations of fluid flow through these structures were conducted for incompressible and compressible fluids. These results allow defining the pressure profile in the inlet of the perforation. For the second part of modeling the main pressure drop is observed along the perforation. Noise is induced from fluid flow out of this perforation with a small diameter in comparison with the diameter of the wellbore. Simulation of 10 seconds of fluid flow was conducted in order to collect enough data for acoustic measurements. Acoustic wave peaks are observed in the amplitude-frequency characteristic plot, and these peaks respond to the different types of fluid and flow rates. Sound pressure level is obtained from amplitude-frequency characteristics and is compared with experimental results.