Introduction Hydrogen can be released during the thermal decomposition of organic materials; therefore, monitoring its level in the working industrial high-voltage transformer oil allows you to identify the development of degenerative processes in advance, because these processes can lead to an accident in the future. In experiments has shown that highly sensitive and small-sized field effect gas sensor based on the metal-insulator-semiconductor (MIS) structure can be used for measuring of H2 in oil with direct contact of its structure with transformer oil. Hydrogen Diffusion in Transformer Oil Numerical estimates show that with the diffusion coefficient D = 4*10-9 m2/s [1] (at room temperature), the time required for H2 saturation in a transformer oil with 10 cm depth to a level 0,9C0 is more than 1 year. In reality, there are convective flows occurred due to the temperature gradient which are combined with diffusion. Nevertheless, such a large diffusion time must be taken into account when designing the experimental setup and the obtained results interpreting. Field-Effect Gas Sensor A schematic representation of the field-effect gas sensor is shown on Figure 1 and described in work [2]. The measuring H2 concentration with the field-effect gas sensor is as follows. When hydrogen molecules from the external environment interact with the metal electrode, H2 molecules decompose into atoms and diffuse to the metal-insulator interface. It is known that on the surface of palladium hydrogen atoms have a dipole moment. The electric field of dipoles leads to a change in the electric field in the dielectric and in the surface layer of the semiconductor. As a result, the capacitance of the MIS-structure changes, the value of which is fixed by the electronic unit while maintaining a constant bias voltage on the MIS-structure. Results and Conclusions The setup used in experiment (Fig. 2) provides similar to a transformer conditions of H2 diffusion. Both field-effect sensors were preliminarily calibrated for H2 in air (in the concentration range 5-100 ppm) at the sensing element temperatures of 23 and 100°С. The optimal temperature of the sensor structure is 100-150°C. However, in experiments, the main sensor could not be maintained at this temperature due to intense heat exchange with oil. For this reason, its heating element was turned off and equal to the temperature of the oil and was 23°C. The reference sensor was maintained at a temperature of 100°C. Relative humidity above oil was 25%. The experiment was carried out as follows. Through drainage tubes, fixed H2 concentrations in the range from 5 ppm to 100 ppm were fed into the container. After the mixture was supplied, the tubes were blocked and, subsequently, the H2 concentration in the vessel was monitored by sensor (Fig. 2 (3)). To accelerate the process of oil saturation with H2, a mixing device was used, the rotation frequency of which was selected to exclude the formation of air bubbles in the oil. Figure 3 shows the dependence of the response of the sensors (main and reference) on the concentration of H2. Table 1 shows the measurement results.The experiment showed that the temperature of the sensor has a significant effect on the detection time of H2. Immersion of the sensor in oil leads to an increase in the noise level of the signal and a decrease in the sensitivity of the sensor in the low concentrations range. Also, the presence of transformer oil does not change the functional dependence of the sensor capacity on the H2 concentration. The response time to H2 is significantly less than the time of its diffusion in oil, which makes it possible to use the sensor as a sensitive element in the control systems for the level of dissolved H2 in the oil of a working transformer. Acknowledgment This work was funded by the Russian Science Foundation under Grant no 18-79-10230 from 08.08.2018.
Read full abstract