50 Years of TMF Contacts Design Considerations

Transcription

1 Technical collection 50 Years of TMF Contacts Design Considerations Conferences publications H. Schellekens

2 ISBN XXIII rd Int. Symp. on Discharges and Electrical Insulation in Vacuum-Bucharest Years of TMF Contacts Design Considerations H. Schellekens Schneider Electric, Medium Voltage Development, Usine 38V, ZAC Champ Saint-Ange, Varces, France Abstract- The historical evolution of TMF or RMF contacts is discussed. Today the very basic guideline as prescribed by Harold. N. Schneider, 50 years ago, still prevails. This allows for simple, compact and robust contact designs. Better understanding of the thermodynamics of the arc has helped to design contacts coping with short circuit currents of up to 100 ka. Contact material improvement has helped to boost up the short circuit interruption performance, which today is equivalent to the AMF contact structures. Progress in the theoretical understanding of the arcing process, and the increasing base of experimental data permit to give guidelines to design a correct functioning TMF contact. I. INTRODUCTION A. History of TMF contacts Commercialization of sealed vacuum interrupters started in the 1950 s. The first VI s were mainly used as load break switches for capacitor bank switching. General Electric Company started development work on VI s in Within GEC, Harold N. Schneider was in 1958 the first to propose a compact contact design, the spiral contact, fig. 1a, to move the constricted arc [1]. At about the same time in 1962 Anthony A. Lake and Michael P. Reece [2] proposed a cup shaped contact which creates a sufficiently strong magnetic field to move the arc on the ridge, fig.1b. Here, the many slots that prevail in the cup extend into the ridge which forms the contact surface. As a variant the cup contact can be topped with a contact disc. The constricted arc, although rotating, projects vapour and droplets onto the surrounding walls. This is attributed by Richard L. Hundstad of Westinghouse Electric Company in 1972 to the radial component of the force on the arc [3]. He, therefore, proposed a folded petal contact system that reduces the radial force on the arc by bending the contact arms in the 3 rd dimension, fig. 1c. B. Relation between TMF, RMF and AMF Often TMF contacts are also called radial magnetic field or RMF contacts, as the magnetic field component that makes the arc rotate points in the radial direction. A TMF contact system is composed of a set of 2 non-identical contacts. AMF contacts are designed to create a mainly axial magnetic field between the contacts. This field tends to keep the arc in a diffuse and stable arcing mode. An AMF contact system is composed of a set of 2 identical contacts, which resemble the contacts of fig. 1b and 1c. C. TMF and Contact Materials The development of vacuum interrupters in general is strongly related to the development of materials; the TMF contacts are no exception to this rule. Three different time periods can be distinguished. In the beginning the main contact material was pure copper or copper-bismuth alloy, an optimized alloy to minimize the force necessary to break contact welds. These materials suffer from arc erosion, which create large droplets that influence negatively the dielectric characteristics of the VI. The development of sintered copper-chromium (CuCr) alloy (in the 1970 s) was a great improvement [4]. With this material the generation of large droplets was reduced as this material has a grain size in the order of 100 µm; so, only micro droplets are emitted by the arc. The grainy structure of the material, though, makes it more susceptible to gas adsorption, which interferes with current interruption in other ways. A CuCr contact fabricated by arc melting in a low pressure environment [5] created a virtually gas free material, which has been the standard for interruption performance since the mid 1980 s. Due to patent protection, this material was not available for the Fig. 1a. Spiral contact by Harold Schneider Fig. 1b. Cup contact by Anthony A. Lake and Michael P. Reece Fig. 1c. Folded petal contact by Richard Hundstad

3 large community. Only recently, a vacuum casting technique [6] has reduced the cost of the fabrication of gas-free CuCr and mass production has popularized its application. D. Renewed interest for TMF At the end of the 1970 s AMF contacts became popular. Higher current ratings and higher voltage ratings and longer electrical switching live could be attained with the AMF technology. The success of AMF attracted more research, which led to a reasonably good understanding of the arc behavior [7,8,9]. As a consequence, the application of TMF contacts regressed relatively. Yet development on TMF contacts continued. For low voltage applications the interruption current was raised to 100kA [10]. By applying the development methodology, which was so successful for the AMF contacts [11], a breakthrough was obtained in the design of TMF contacts. As a consequence, near to identical interruption performance was obtained for TMF and AMF contacts. Recently also the first generator breaker based on TMF contacts with rated current of 75 ka at 15 kv has been commercialized. [12] II. ARC MODELS Models of arc motion are used to relate contact design to interruption performance. To describe the arc motion across the contact surface, an obvious approach would be to use electro-mechanics: the Lorentz-force on a solid conductor to explain the arc motion. Yet, the arc can not be treated as a solid conductor, as the plasma is composed of electrons, ions and gas. Movement of these species within the plasma is far more complex and the displacement of the arc becomes less evident. Below different arc model categories are presented that deal with arc motion. The arc is mainly characterized by the physical evaporation process at the anode and cathode foot point. The ionized species are confined to the arc by the strong self-magnetic field. The neutral particles are lost from the arc due to diffusion. Evaporation has to compensate for this loss (energy balance). The neutral mass loss represents a loss of momentum (force balance). This leads to a direct relation between arc speed and contact design [13]. Scaling laws have been derived that relate these parameters to interruption capability [14]. C. 2D and 3D arc models A 2D plasma dynamic model combines full plasma modeling with an exact description of the plasma contact interaction including thermal energy balance in the contacts. This model gives an impressively realistic image of how the arc could move fig. 4. [15] In the phenomenological arc model Nike [9,16], the arc is governed by a minimum energy criterion and adapts its size to the prevailing magnetic field. In a transverse magnetic field, the arc reduces its cross-section and moves in the amperian direction. Fig. 5 shows the arc for three sequential time steps in a folded-petal contact system. The time necessary to move between the positions depends on arc voltage and inductivity of the contact system [17]. Fig. 2. Solid arc (blue) Fig. 3. Direction of Lorentz model of TMF contact force for a solid arc body geometry with 4 curved on 4 curved petal contact for petals. any possible position. A. Arc models based on force balance equations 3D magnetic field modeling tools, make it possible to calculate the force on a solid body for complex electrode geometries. Fig. 2 shows a typical 3D finite element model of the arc between curved contacts. This approach allows adapting the contact geometry to, for instance, maximize the force on the arc in the rotating direction, fig. 3 [11]. B. Models based on energy balance equations A sophisticated approach mixes a force balance model of the arc with an energy balance model for the contacts. Fig. 4. Arc motion predicted by plasma dynamic model Fig. 5. Arc motion predicted by 3D model Nike. D. Comparison between the models In table I a comparison is given of the above presented models in terms of their predictive powers. All models are capable to predict the arc speed. Only models that include a description of the arc plasma behaviour are capable to predict the arc diameter. Energy balance equations relate the arc speed to the material properties of the contact material. Only extended, 2D/3D physical models can explain the arc motion from one contact arm to the other. With respect to scaling laws, force calculations contain not sufficient elements. The 2D/3D physical models mask the essential relationships in their numerical output but multiple simulations allow revealing scaling laws. Albeit 2D/3D physical models predict arc movement, in [15] arc motion is conditioned by the ability of the arc to generate a hot electron emissive contact surface up front, whereas in [16] the

4 main current displaces, due to electrodynamics and arc energy balance, which results in an apparent arc motion. TABLE I : Comparison of predictive powers of the models. Type of Model Force Force and Calculation Energy 2D and 3D models Balance Arc Diameter No Yes Yes Arc Speed Yes Yes Yes Time to Cross a contact slot No No Yes Scaling laws No Yes Yes III. DESIGN CONSIDERATIONS The design considerations given below are based on experimental results, theoretical models and numerical simulations. A. Contact shape The contact system has to be assessed [11] on the effectiveness of its current interruption capability. This is promoted by its ability to move the arc and to force it to rotate. It is evident that a contact fails if the arc does not run at all, or suddenly stops and stays immobile. To achieve this the main design rule, as formulated by Harold Schneider [1], has to be followed: The disc (contact) should be slotted from the outer periphery inward, and the slot configuration should be such that the current path between the conductor and an arc terminal located at substantially any angular point on the outer peripheral region has a net component extending generally tangentially with respect to the periphery in the vicinity of the arc. All three TMF contact types of fig. 1 obey to this rule. Fig. 6. Anode arc root diameter variation with arc current for a Siemens type cup contact. B. Width of the contact arm The width of the arc on a contrate cup contact. is given in fig. 6 for the anode side of the arc. The width of the arc at the cathode side is square root of 2 larger. This sets a lower limit for the width of the contact arms [13]. C. Number of contact slots The time to make one revolution is composed of the time for the arc to cross the total circumferential distance with the proper arc speed and of the time to cross each of the contact slots. The proper arc speed depends on parameters like momentary arc current, arc diameter, local magnetic field and contact distance. The local magnetic field strength depends on the form of the contacts. In some designs, like the folded petal, this field is maximized. The proper arc speed is between 100 to 1000 m/s for a vacuum arc. Fig. 7 shows the evolution of time depending on the angular position of the arc; moments of rapid displacement across the contact arms are interrupted by the larger time it takes to cross the contact slot [17]. The number of slots has a direct influence on the time to make a complete revolution. Fig. 7. Evolution of time depending on the angular arc position. The proper arc speed is 1030m/s and the apparent speed is 290 m/s. D. Contact diameter Interruption performance is intimately related to contact diameter, D, and contact distance, d. The theoretical relation of eq. 1 is derived by [14]. The constant C RMF depends on contact geometry and material. I = C D (1) max RMF d 0.7 E. Width of slots The slots in the contact are an essential feature to create their form. Yet, the width of the slots has no influence on the behaviour of the TMF arc or on its ability to cross the slot. The latter solely depends on the length of the contact arm. Due to contact erosion the slot fills up progressively with melt. Therefore an appropriate choice of slot width should be made depending on the expected lifetime of the contacts. F. Contact thickness Contact thickness has a direct influence on the magnetic field in the arc. For a stationary or stalled arc the field reduces with increased contact thickness; so, to set an arc in motion thin contacts should be favored. For the running arc current flow is close to the surface and the contact thickness has nearly no influence on its speed. Fig. 8 shows the current distribution on a rail contact an arc speed of 50 m/s. G. Contact distance The contact distance has a direct influence on the arc speed. On the contact arm the constricted arc will always run. Yet, a minimum contact gap is necessary to make the arc move across the slot. With increasing contact distance the proper speed of the arc as well as the number of revolutions will increase. The rotation frequency depends on the contact distance and on the number of contact arms.

5 10000 Fig. 8. Current distribution in a rail contact for arc speed of 50 m/s. Current flow close to contact surface. H. Contact heating As the constricted arc is running on the edge of the contact, the contact heating is very time and place dependent. The notion of the relative heating time defined as the ratio of heating time and cooling time becomes useful in order to compare different contact designs. A heating period is followed by a cooling period. The heating time is the ratio of arc diameter and proper arc speed; typically for a current of 50 ka the arc has 8 mm diameter and the proper arc speed of 200 m/s this time is 40µs. As the arc stalls at the end of the contact arm, the heating time at this position is of the order of 100µs, fig.7. The cooling time for both positions is determined by the time of one complete revolution 550µs. For this example the relative heating time at the end of the contact arm is 100/550= 18%. Reduction of the relative heating time is an efficient way to increase the interruption performance. I. Shield heating As the arc heats the contact surface, the contact surface temperature will increase and can surpass easily the melting temperature. Due to the high magnetic pressure in the arc the liquid will be ejected from the arc root. This process of erosion cooling is one of the advantages for TMF contacts with respect to AMF contacts. The disadvantage is the deposition of the erosion products on the shields. During the arcing process the shields are under a constant spray of liquid material. This material will cool upon impact with the shield. Depending on the thermal accommodation between shield and contact material, the shield will heat up beyond the melting temperature and start to erode. For a correct performance of the VI the shield thickness and composition are important parameters. J. Electrical endurance Contact erosion increases exponentially with current on TMF contacts. Therefore, the nominal short circuit current rating depends on the final application. Fig. 9 compares the electrical endurance for AMF and TMF contacts for a 20kA rated vacuum interrupter. K. Opening speed A minimum contact distance, ~ 4 mm, is necessary before the arc crosses a slot. So, the opening speed of the circuit breaker should be designed such that this minimum opening is attained in a short time to facilitate the interruption process. Operations AMF TMF Current [ka] Fig. 9. Electrical endurance of AMF and TMF contacts. IV. CONCLUSIONS Improved contact materials and a better understanding of the multi-physics governing the TMF help in the conception of compact and efficient vacuum interrupters. In the continuing process to reduce the cost of vacuum interrupters the TMF technology is a challenging alternative to the AMF technology for applications where electrical endurance is of second concern. ACKNOWLEDGMENT The author thanks the students Julien Fontchastagner, David Gonin and Guillaume Gomez for their valuable contributions. REFERENCES 1. H. N. Schneider, US. Patent 2,949,520, filed A. A. Lake and M. P. Reece, UK Patent 997,384, filed R. L. Hundstad, US. Patent 3,845,262, filed A. A. Robinson, British Patent 1,194,674, R Muller, Arc-Melted CuCr Alloys as Contact Materials for Vacuum Interrupters, Siemens Forsch. Entwicklungsber , , B. Miao, Z. Yan, Z. Yumin, L. Guoxun, D. Shenlin, H. Yu, Two new Cu-Cr alloy contact materials, XIXth ISDEIV, p , Xi an, China 7. S. Yanabu, S. Souma, T. Tamagawa, S. Yamashita, T. Tsutsumi, Vacuum arcs under axial magnetic field and its interrupting ability, Proc. IEE, 126, p , H. Fink, M. Heimbach, W. Shang, A new contact design based on a quadrupolar axial magnetic field and its characteristics, European Trans. on Electrical Power Engineering, H. Schellekens, "Arc Behaviour in Axial Magnetic Field Vacuum Interrupters Equipped With an External Coil", Proc. XVIIIth ISDEIV 2, , Eindhoven, H. Fink and R. Renz, Future trends in vacuum technology applications, Proc. XXth ISDEIV, 25-29, Tours, E. Dullni, E. Schade, W. Shang, Vacuum arcs driven by cross-magnetic fields (RMF), ISDEIV pp , Tours, Eaton / Cutler-Hammer Product Bulletin BR E Medium Voltage Generator Vacuum Circuit Breakers, W. Haas and W. Hartmann, Investigation of arc roots of constricted high current vacuum arcs, IEEE transactions on plasma science, Vol. PS-27, No. 4, pp , August R. Renz, Thermodynamic Models for TMF and AMF vacuum arcs, XXIIth ISDEIV, pp , Matsue, T. Delachaux, O. Fritz, D. Gentsch, E. Schade and D.L. Shmelev, Numerical simulation of a moving high-current vacuum arc driven by a transverse magnetic field, XXIIth ISDEIV, pp , Matsue, H. Schellekens, Phenomenological arc model NIKE, private communication Current zero club meeting 2003, Grenoble. 17. J. Fontchastagner, O. Chadebec, H. Schellekens, G. Meunier, and V. Mazauric Coupling of an Electrical Arc Model With FEM for Vacuum Interrupter Designs, IEEE TRANSACTIONS ON MAGNETICS, VOL. 41, NO. 5, MAY of author: hans.schellekens@schneider-electric.com