The overall goal of this experiment is to provide high resolution flow field data to validate computational fluid dynamics computer codes, checking their ability to accurately simulate heat transfer in turbulent flows. This advanced temperature measurement technique can generate thousands of data points along a single sensor made of commercial telecom fiber. It's able to provide data density that's far beyond traditional sensors like thermocouples.
But unlike thermocouples, these fiber optic sensors are sensitive to strain as well as temperature. So handling, vibration, and changes in humidity influence the temperature's signal. The strain sensitivity is quite unlike traditional sensors.
And obtaining accurate data, requires special operating procedures and practices which we will demonstrate here. The distributed temperature sensor, or DTS will be demonstrated using this test section. It is a glass tank with two hexagonal air ducts attached to the base.
The air exits through a vent at the top after mixing the in interior. On the inside of the top panel, is a black polymer screen over the location of the DTS. Access the interior of the test section by removing panels at the top and side.
Strung across the top of the test section, is steel wire with a 20 millimeter pitch. Over the measurement area, the wires support the optical fiber, of the distributed temperature sensor. The optical fibers run perpendicular to the steel wire and are spaced 10 millimeters apart.
The start of the DTS has roughly a meter of fiber reserved for splicing and end termination. Segments are woven through adjacent wires from one side of the measurement region to the other. The fiber is looped to being the adjacent segment and the ends are secured to the test section with tape.
This is a schematic of the final test section layout from above. The steel wires cover a region that includes the ducts. The fiber of the distributed temperature sensor is woven through the wires in the area over the ducts.
To weave the optical fiber of the DTS follow these steps. Begin with a 50 meter spool of fiber, and unspool enough to create the reserve for splicing to a connector outside the tank. Anchor the fiber beyond the reserve to the test section lid.
Next, unspool about a half meter of fiber that will be used to start the first new sensor segments. Lay a few hundred millimeters of fiber onto the wire mesh and shape it into a loop. Being weaving the loop through the support wires alternating between going above and below adjacent wires.
When the opposite side of the tank is reached, the fiber for the first two sensor segments is ready to be taped into position. Pull the first segment taut and tape it into position on the underside of the lid. A ruler glued to the lid, is used to indicate position and control segment spacing.
We want this taut enough that it stays in position during flow but not so taut that deformation at the test section strains the sensor and causes measurement error. Now, form a loop of roughly 30 millimeters diameter in the unanchored portion of the fiber and tape it 10 millimeters from the first segment. Anchor the other end of the second segment to the tank lid.
unspool about one half meter of fiber to begin the next pair of sensor segments. Repeat the steps used for the first two detector segments to create additional segment pairs. With the sensing array in place, splice a connector to the end of the fiber.
Plug the sensor into the yellow delay cable that will link the sensor to the interrogator. Connect the delay cable to the interrogator and complete the configuration steps. The sensor interrogator is based on swift wavelength interferometry.
A low power tunable diode laser sends a narrow band signal into the fiber. The laser is swept across several nanometers, and the signal is split between reference and measurement legs. Scattered light from the sensor, is combined with the reference signal, to generate interference signals at the detectors.
The detector output is used to retrieve the release scattering signal. The sensor positions must now be mapped. To do this, connect a soldering iron to a variable transformer, set to about 40 percent.
At the computer, use the interrogator software to display live data on the screen. Take the soldering iron and approach the start of the first segment of the sensor, hold the iron near the sensor and briefly touch it to the end point. At the computer, stop data acquisition to freeze the peak.
Then zoom in on the temperature spike to record the position of the peak and the physical location. Collect data in the same way for the start and end points of each segment. After mapping the sensor positions, prepare one or more temperature standards.
In this case, use the thermocouple for the task. Select a region in the measurement zone of the DTS. Mount the thermocouple near the fiber of the DTS.
After completing work in the interior, close the tank by returning all panels to their positions. Continue by wrapping the tank in some form of insulation. Once wrapped, allow it to sit long enough to establish an isothermal atmosphere around the sensor.
When the system is ready, start the interrogator software and begin measurements. To get the baseline, first stop the measurement. Then select baseline and enter a file name in the provided field.
Select a measurement range and click on OK to being the baseline process. Simultaneously record the thermocouple temperature reading for calibration. Once the software has completed the baseline, go to the measurement mode.
The signal should now be zero along the entire length of the sensor and not drift overtime if she system is in a thermally stable state. At this point, go to the flow and heating controls to prepare for a run. With the air flow on, adjust the flow rates in the channels to match.
Adjust the power controller to heat one jet and establish the desired differential temperature between jets. Return to the system after it has run overnight to reach equilibrium. Examine the live DTS signal to access noise levels and choose the appropriate gauge length.
Use the distributed temperature system to log 2000 scans at 4 hertz. This is raw data of the measured temperature difference from the baseline temperature. The left end of the horizontal axis corresponds to a fiber segment on the boundary of the system denoted as the east end.
Moving to the right, the fiber loops back and forth toward the west end. Peaks occur where the sensor is directly above the hot jet. The noise toward the west end is due to flow induced vibration.
Here, the raw data is mapped onto positions in the test section. The point of view is from above the measurement plain on the DTS array. The outlines of the hexagonal channels are an A to orientation.
Linear interpolation between the data points allows the creation of a 2D temperature contour map to provide a sense of the thermal pattern beneath the lid. The red area indicates a warm region over the east jet but not centered on it. The sensor data can also be used to map the route mean square of the temperature fluctuations.
The magenta region is one of high temperature fluctuations and high turbulence where the two rising jets interact as they impinge on the lid. We've seen how a single fiber optic sensor, based on rayleigh scattering, can generate thousands of temperature measurements to provide a detailed picture of the full field that would be unobtainable with thermocouples. To ensure measurement accuracy, always pay special attention to strain control after the baseline and minimize flow induced vibrations through thoughtful design of the sensor support configuration.
This distributed sensing technique provides new possibilities for temperature measurements in opaque fluids and other applications that are unsuitable for optical techniques that rely on lasers and cameras.
