Sand management/control

Downhole Sand-Ingress Detection With Fiber-Optic Distributed Acoustic Sensors

This paper shows results from use of a new technology that uses in-well-conveyed fiber-optic distributed acoustic sensing (DAS) for the detection of sand-ingress zones across the reservoir section throughout the production period in real time.

Fig. 1—DAS-system sound field filtered to monitor sand transport.

There is currently no proven technology available in the market that accurately identifies downhole sand-ingress locations in real time. In this paper, the authors present results from use of a new technology that uses in-well-conveyed fiber-optic distributed acoustic sensing (DAS) for the detection of sand-ingress zones across the reservoir section throughout the production period in real time.


Mechanical sand-control systems are not always fully effective. The end result may be high sand production, which results in choking back the well and reducing hydrocarbon production significantly. In most cases, the precise sanding interval is unknown, making sand-remediation operations (such as remedial plug placements) often ineffective. A successful remediation requires identification of locations of sand entry to inform targeted sand-shutoff operations. However, no proven technology accurately identifies sand-ingress locations during well production in real time.

The technology described in this paper has now been used successfully

  • To monitor sand ingress during production (in real time) to inform and optimize oil production
  • To inform and implement a targeted sand-shutoff operation to increase hydrocarbon-production rates
  • To assess the reliability of the sand-control equipment during production to improve future designs

Sand Detection

While conventional surface acoustic sand detectors provide a delayed indication of onset of downhole sanding events, they do not provide information about the zones in the reservoir that are producing sand. A successful sand-­shutoff operation, however, requires knowledge and definitive identification of the zones (or depth sections) in the reservoir contributing to sanding and their relative concentrations.

DAS has been viewed as a potential candidate technology for downhole sand detection in recent years. DAS systems are intrinsic optical-fiber-based acoustic-sensing systems that use the backscatter component of the light injected into an optical fiber to detect acoustic perturbations along the length of the fiber. The fiber itself acts as the sensing element, with no additional transducers in the optical path, and measurements are taken along the length of the entire fiber, allowing for a true distributed measurement using a single fiber. The technology provides sensitivity to strain variations by monitoring changes in the length and index of refraction of the fiber induced by impinging acoustic pressure waves.

Technology Development

Flow-Loop Experimentation. The first step in developing a real-time sand-detection system was to develop a physical understanding of the acoustic fingerprint of sand ingress into hydrocarbon-producing wells (i.e., to have a description of what sand ingress into a production tubing “sounded” like on DAS optical fiber). This acoustic pattern of sand ingress was derived empirically with the help of experimental data from a multiphase flow-loop experiment. The results have now also been modeled through first principles to obtain a clearer understanding of the underlying physics. The flow loop comprised a test section with multiple fluid and sand-ingress ports. The fiber was attached to the outside of the test pipe to simulate the case in which the fiber is installed as part of the downhole completion. Known concentrations of sand were then injected along with known quantities of multiphase fluids, to simulate reservoir conditions better. The DAS data were then gatherered across the entire test section to analyze the characteristics of the acoustic data specifically at the sand-entry points and to compare these with zones where there was no sand injection but only multiphase fluids. These experiments formed the basis for uniquely identifying and extracting the sand-ingress fingerprint. This was followed by experiments to quantify the performance limits of the DAS system by varying the concentrations of injected sand at different locations across the test section.

Specific tests were also conducted to examine the variability of the sand-­ingress acoustic pattern with respect to variations typically observed during production, to simulate reservoir conditions better. The data were also analyzed to distinguish sand transportation acoustics from sand-ingress acoustics.

Field Trials. Following the flow-loop trials, the performance of the DAS system and the real-time sand-detection ­algorithm was tested with field data acquired on a well with a history of sand production. The candidate well had openhole-gravel-pack completions, and a fiber-optic cable was preinstalled across the reservoir interval. Three consecutive field trials were conducted at 6-month intervals.  

Field Trial 1. The acquired data were depth calibrated to ensure data alignment with the wellbore measured depth (MD) before commencement of data acquisition for sand-monitoring purposes. The DAS data were acquired from the entire length of the fiber that was installed permanently within the downhole completion assembly. The well was initially operated at a steady production rate and then ramped up to higher drawdown pressure before being brought back to initial production conditions. The data were then manually processed to check and analyze the data gathered across the reservoir section, to identify zones with acoustic signals with characteristics similar to those of the sand-ingress pattern as modeled and experimentally observed during the flow-loop trials. This indicated multiple depth zones with acoustics with characteristics similar to those of the sand-ingress pattern, and these zones of interest were then investigated further to verify conformance with the sand-­ingress pattern. The data were later reprocessed to filter out the background flow and instrumentation noise. The filtered sand acoustic signals were then averaged through time to construct a sand log across the entire reservoir section.

The DAS-system sand log presents the acoustic amplitude (filtered for sand ingress) as a function of depth across the reservoir interval. Reviewing the log indicates five distinct zones of sand ingress, with relatively higher sand-­ingress noise observed in Zones 3 and 4.

The DAS data were then further analyzed to study the transportation of the sand from the point of entry all the way up to the surface. Fig. 1 above displays a DAS sound field, filtered to monitor sand transportation over the entire well depth, acquired for a period of several hours during well ramp up. The sound field displays the calibrated measured depth along the y-axis and the time stamp along the x-axis; the acoustic intensity is shown by a color scale (with red being high acoustic intensity and blue being low). The plot also overlays the surface flowline acoustic-­sensor data (red data trace at the bottom of the illustration) aligned in time with the DAS sound field.

A few observations can be made from the sound field shown in Fig. 1:

  • Filtering the DAS sound fields to pick up sand-transport acoustics indicated sand slugs originating from approximately 3000-m MD and reaching the surface with velocities increasing at shallower depths. Correlating the DAS measurements at times when the sand slugs reach the surface with the data captured on the surface acoustic sensors at the production flowlines indicates a good match between the filtered DAS data and the surface sensor data.
  • A red carpet of high-intensity acoustics is observed near the surface. This is because of contamination by surface noise (caused by mechanical equipment on the surface) that may be filtered and removed through further processing. 

Field Trial 2. To check the repeatability of the measurement, a second field trial was conducted in the candidate well, 6 months after the first field trial. This time, the well was initially shut in and then ramped up in steps, holding the production rates stable for several hours at each choke setting. As part of the field trial, data-processing and real-time data-handling infrastructure was developed to enable real-time streaming of processed acoustic data. Back-end visualization systems were also built to visualize and manipulate the processed acoustic data in real time.

Sand logs were computed at each of the production rates (under stable drawdown conditions), and the computed logs were then integrated through time to construct drawdown-lapsed sand logs. These sand logs allow for a better representation of the temporal behavior of sand ingress across the sanding intervals identified downhole.

The first data trace, as seen in Fig. 5 of the complete paper, shows the sand log computed at a relatively low drawdown (of nearly 160 psi). When comparing these results with those of Field Trial 1, it may be observed that many of the sand-producing intervals identified in Field Trial 1 remain consistent in Field Trial 2. As the drawdown is increased progressively, more zones are seen to produce sand.

Field Trial 3. The results from Field Trials 1 and 2 were then used to execute a DAS-informed targeted sand-remediation operation. Selected sand-ingress zones as identified in Field Trials 1 and 2 were isolated with mechanical-patch technology, and the well was brought back into production. The DAS system was then reconnected to the optical fiber, and the data acquisition commenced. The complete end-to-end technology solution incorporating the real-time sand-detection algorithm was installed and tested as part of the field trial. Sand logs were also computed in real time to evaluate and compare the sanding behavior observed downhole post-remediation.

The results show a substantial reduction in the overall sand production (by more than 70%) after remediation, indicating a successful treatment. This was also confirmed through physical samples taken on the surface. The lower sanding rates have also enabled further ramp up of the well, consequently leading to a significant increase in oil production by more than 2,000 B/D.

This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 183329, “Downhole Sand-Ingress Detection With Fiber-Optic Distributed Acoustic Sensors,” by Pradyumna Thiruvenkatanathan, Tommy Langnes, Paul Beaumont, Daniel White, and Michael Webster, BP, prepared for the 2016 Abu Dhabi International Petroleum Exhibition and Conference, Abu Dhabi, 7–10 November. The paper has not been peer reviewed.