Abstract Body

Liquid hydrogen and liquid nitrogen could be used as coolants for superconducting cables in a future power grid [1]. The demand for liquid nitrogen and liquid hydrogen is expected to increase as various superconducting devices become more widespread in the future. In addition, liquid air and liquid hydrogen have potential as energy storage technologies in the transition to renewable energy. Liquid level sensors are needed to manage these low-temperature fluids. Existing liquid level sensors include differential pressure sensor [2], capacitance sensor [3], and superconducting sensor [4]. However, most of known level sensors have components and electrical wiring in contact with the liquid. This may result in loss of liquid due to heat transfer and potential accident risk due to electrical surge current etc. In this study, an all-optical liquid level sensor that does not contact the liquid is investigated.

Figure 1 shows the measurement principle. Three optical fibers, a fiber bundle, and a lens are attached to the sensor. The optical fibers are arranged in an equilateral triangular shape, and light is irradiated from their tips onto the liquid surface. The reflected image from the liquid surface is a triangle of three light points, which is observed by a camera through the lens and a fiber bundle. The size of the reflected image depends on the distance between the sensor and the liquid surface.

Figure 2 shows the overall view of the experimental apparatus, and Figure 3 shows the all-optical sensor constructed. In this experiment, liquid nitrogen is used as the cryogenic fluid. The sensor is installed at the mouth of the liquid nitrogen tank, and the reflected image displayed on the outer surface of the fiber bundle is observed by a camera. From the observed reflected image, the positions of the light points are measured using image processing. From there, we compute the size of the reflected image of the triangle, from which we obtain the level using a calibration curve [5].

The three optical fibers that illuminate the liquid surface can be controlled individually. This is because the measurement of a nonstationary liquid surface requires the reflected image of each light spot. When the liquid surface is nonstationary, the reflected image also fluctuates and if three light sources are simultaneously in operation, the reflected image of the light points overlap each other, making it difficult to determine their positions. By separately observing the reflected image of each light spot, we should be able to average the data and determine the position of each light spot when the liquid surface is nonstationary.

For this purpose, we use an optical coupler shown in Figure 4. Red semiconductor lasers are used as light sources, and the coupler is designed to couple the lasers and the optical fibers for illumination. Each of the three laser beams is focused through the lens onto the end surface of an optical fiber, providing a stable light supply. The light sources are connected to a PC via an interface (NI USB-6002), and can be turned on and off by a Python program. In this experiment, the laser is programmed to be turned on for 0.1 second at a time, and the camera used to observe the reflected images take 100 frames per second, making it possible to capture approximately 30 reflected images of each light spot per second.

Figure 5 shows the power supply for the laser. We make the lasers blink during the measurement by turning the power lines on and off, for the lack of control lines. Since any pulses or spikes in the power supply will shorten the life of the laser, the power supply unit incorporates a low-pass filter circuit with the Bessel characteristic, which has the slowest attenuation response among filter characteristics, to ensure smooth change of the voltage in the power line to the laser.

Figure 6 shows the observed reflection images. The distance between the liquid surface and the sensor was 255 mm. The lasers of the light sources blink in sequence. Figure 6(a) shows three reflection images when the liquid surface is stationary. The three spots in these images constitute an equilateral triangle, when combined. Figure 6(b) shows reflection images when the liquid surface is nonstationary. Determination of the positions of the light spots in reflected images when the liquid surface is nonstationary will be a subject for future work.

References

[1] Paul M. Grant, Chauncey Starr, and Thomas J. Overbye, A Power Grid for the Hydrogen Economy. Scientific American (June 2006).
[2] W. A. Olsen, A Survey of Mass and Level Gauging Techniques for Liquid Hydrogen. In: K.D. Timmerhaus (eds), Advances in Cryogenic Engineering, vol 8. Springer, Boston, MA. (1963).
[3] Koichi Matsumoto, Masamitsu Sobue, Kai Asamoto, Yuta Nishimura, Satoshi Abe, and Takenori Numazawa, Capacitive level sensor for liquid hydrogen, Cryogenics 51, 114-115 (2011).
[4] Ch. Haberstroh and G. Zick, A Superconductive MgB 2 Level Sensor for Liquid Hydrogen, AIP Conference Proceedings 823, 679 (2006).
[5] Muneo Futamura, Toshiyuki Oikawa, Shigeo Miura, and Hiroshi Okamoto, All-optical non-contact level sensor for liquid hydrogen, J. Phys. Conf. Ser, in press.