No abstract

An investigation on internal arc fault tests in gas-insulated metal enclosed MV switchgears is described and discussed. The influence of different gases (SF 6 and air) and electrode materials (Cu, Al, and Fe) has been put into evidence. This is highly relevant based on the fact that IEC Standards and utilities technical specifications allows that internal arc fault tests are performed with air replacing SF 6 with some precautions. The experimental tests were carried out at NEFI High Power Laboratory in Skien, Norway. Full-scale test objects representative of typical gas insulated metal enclosed MV switchgear were prepared, filled with air or SF 6 and tested. The short circuit current was 16 kA rms and the duration was 1 second. Electric input energy, internal gas pressure in different locations, opening time of bursting discs and temperature were acquired. The arc was ignited at the end of the simplified, but still representative, 400 mm long bus bars that were located in the middle of the metal enclosure. The full-scale experiments were additionally analyzed by means of Finite Element Method. The comparison between experiments and simulations show that there was possible to set up an arc model and tune it in to the results. The main difference between testing of units filled with air versus SF 6 is the significant faster pressure increase in air. As a result of this and the fact that the bursting discs need a certain time to open, the pressure inside the test object will be higher at the time the bursting discs open when testing with air. The maximum pressure reached during a test with air however may be equal or lower than in SF 6 depending on dimensional parameters of bursting disc and encapsulation. Furthermore it is evident that there is a clear difference in the exhaust characteristics of SF 6 and air from an internal arc test.

MV switchgear experiences a rise in temperature during normal operation due to ohmic losses in conductors and contacts. If the temperature rise is too high, the switching device may be degraded. The focus of this paper is to find a value for the total heat transfer coefficient that may be applied to estimate the temperature of critical parts (open/close contact) of the load break switch in an enclosed MV switchgear, relative to the surrounding air (inside the enclosure) for future design. The values for the total heat transfer coefficient (including all transfer mechanisms) showed a relatively strong dependence on the surface emissivity and the actual design of the switch, but was less dependent on temperature changes within the relevant temperature range. Based on our findings, it is reasonable to assume that the total heat transfer coefficients may be applied in a first approximation of the temperature rise of a load break switch contacts relative to the surrounding air inside an enclosure. Further refinement could be obtained by taking the actual design of the switch into consideration, especially details influencing the emissivity and design elements influencing the heat conduction to adjacent conductor parts.

Arc fault tests of medium voltage switchgear have been performed with reduced volume and arc energy. This has been done in order to investigate if small scale experiments can be used to predict the pressure build-up during full scale arc fault test. Between 40 and 50 % of the arc energy was transferred to the gas in both small scale and full scale tests. The results show that it is reasonable to assume that small scale testing down to 10 % can be used to predict the pressure rise in a full scale test with a single phase arc with about 10 % precision. However, the scaling method of the arc energy seems to be important.

An arc fault inside metal enclosed switchgear will cause the pressure to rise and vaporization of electrode material may contribute to the pressure rise. An experimental study of high current arc erosion on copper electrodes in air has been performed, with an evaluation of fraction lost by gross melting and vaporization. All experiments were performed at NEFI High Voltage Laboratory in Skien, Norway. The measured mass loss from vaporization in our experiments seems to be negligible compared to erosion by gross melting.

No abstract

Full scale arc fault experiments are expensive, time consuming, and demand high power laboratory facilities. This paper presents a literature study performed to evaluate the possibilities of performing representative experiments on small scale by reducing the size of the test object. Previous model experiments have shown that the arc energy must be scaled down by the same factor as the volume in order to produce the same pressure build-up during the tests. The arc energy is given by the arc voltage and current, which both are functions of the arc length. In arc fault experiments one usually assumes a linear voltage drop along the length of the arc and experiments have shown that a voltage gradient of 20-30 V/cm can be assumed for arcs in switchgears filled with air. This value will however depend on volume, pressure, and electrode material. Experiments also indicate that there is a limit for short arcs where the linear voltage drop no longer can be assumed. To reveal possible consequences this may have on small scale experiments, a more systematic experimental study is required.

No abstract