- Email: [email protected]
Electrical and structural R&D activities on high voltage dc solid insulator in vacuum
Fusion Engineering and Design 96–97 (2015) 563–567
Contents lists available at ScienceDirect
Fusion Engineering and Design journal homepage: www.elsevier.com/locate/fusengdes
Electrical and structural R&D activities on high voltage dc solid insulator in vacuum N. Pilan a,∗ , D. Marcuzzi a , A. Rizzolo a , L. Grando a , G. Gambetta a , S. Dalla Rosa c , V. Kraemer d , T. Quirmbach d , G. Chitarin a , R. Gobbo b , G. Pesavento b , A. De Lorenzi a , L. Lotto a , R. Rizzieri a , M. Fincato a , L. Romanato a , L. Trevisan a , V. Cervaro a , L. Franchin a a
Consorzio RFX, Associazione EURATOM-ENEA sulla Fusione, Corso Stati Uniti 4, I-35127 Padova, Italy DII, Università di Padova, v. Gradenigo 6/A, I-35131 Padova, Italy c Umicore – Italbras S.p.A., Strada del Balsego, n.6, 36100 Vicenza, Italy d FRIATEC Ceramics Division, Steinzeugstrasse 50, 68229 Mannheim, Germany b
h i g h l i g h t s • • • •
A thorough R&D activity on the MITICA post insulator prototypes is being carried out. The design has been numerically verified considering both mechanical and electrical aspects. Experimental validation has been started, with positive results in both involved fields. Alternative design solutions thickness have been proposed and successfully tested.
a r t i c l e
i n f o
Article history: Received 19 September 2014 Received in revised form 18 May 2015 Accepted 20 May 2015 Available online 30 May 2015 Keywords: Insulator High voltage Ceramic Vacuum Electrostatic accelerator
a b s t r a c t This paper describes the R&D work performed in support of the design of the alumina insulators for the MITICA Neutral Beam Injector. The ceramic insulators are critical elements, both from the structural and electrical point of view, of the 1 MV electrostatic accelerator of the MITICA injector, as they are required to sustain both the mechanical loads due to the cantilevered weight of the ion source and the high electric field between the accelerator grids. This paper presents the results of numerical simulations and experimental tests on prototypes that have been carried out to validate the insulator design under realistic operating conditions. © 2015 Consorzio RFX. Published by Elsevier B.V. All rights reserved.
1. Introduction The MITICA electrostatic accelerator is based on the MAMuG (multi-aperture-multi-grid concept [1,2]). This accelerator is composed of five stages and includes seven main grids, polarized at potentials ranging from −1 MV to ground (0 V) with five main stages of 200 kV each (see Fig. 1). The negative ion beam, extracted from the RF ion source, is composed of 1280 individual beamlets which are guided and focused by an intense electric field through the apertures of the accelerating grids. The electrical insulation between the grids relies, both
∗ Corresponding author. E-mail address: [email protected] (N. Pilan). http://dx.doi.org/10.1016/j.fusengdes.2015.05.057 0920-3796/© 2015 Consorzio RFX. Published by Elsevier B.V. All rights reserved.
on vacuum gaps and ceramic insulators which are part of the sustaining structure. Five sets of alumina cylindrical insulators are located between the accelerator grids; the insulators are therefore subjected to 200 kVdc and to structural loads. The ion source and the multi-stage accelerator constitute a cantilevered structure sustained from the grounded stage. The present work describes the R&D activities carried out to identify a reliable insulator design considering several constraints: withstanding voltage, structural and manufacturing issues. 2. Insulator description The insulator is a cylinder made of high purity alumina C799 [3] according with the ITER requirements [4], the cylinder has a diameter of 130 mm and a length of 280 mm.
564
N. Pilan et al. / Fusion Engineering and Design 96–97 (2015) 563–567
3.9–3.95 g/cm3 , average grain size 10–15 m. A M14 carbon steel screw (class 12.8) has been screwed in the tapped hole in the alumina and it has been pulled by a tensile machine up to 100 kN (the limit of the machine) without any issues. The same thread has been subjected to a severe fatigue test: an oscillating load between 65 and 85 kN has been applied until failure, which occurred in the screw (made by carbon steel) at about half million cycles. The functionality of the threaded ceramic hole was not compromised. The second phase of the R&D campaign was focused on the manufacturing and testing of the full prototype, for a thorough mechanical and electrical validation of the design, with reference to the numerical verification described in the next sections. 3. Structural design and prototype tests
Fig. 1. Vertical section view of the half MITICA beam source electrostatic accelerator. The insulators are the horizontal axis cylinders in light blue. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)
Each ceramic insulator is connected to the structure by means of a couple of stainless steel (SS) flanges which are directly bolted on the frames sustaining the accelerating grids. A central dowel is present on each flange to assure coaxial positioning between the stainless steel flanges and to prevent any relative displacement due to shear stresses between ceramic and metal flanges. The low friction coefficient between ceramic and SS is not sufficient to guarantee a rigid bolted joint between the two materials. The joining between metal flanges and ceramic insulator is not a straightforward choice because both electrostatic and structural requirements have to be satisfied at the same time. In the reference configuration, the alumina is joined to the SS flanges by screws, six threaded holes M14 are manufactured directly on the sintered ceramic; the blind holes are metalized with a suitable conductive layer as well as the interface between the ceramic and the metallic flange. The metallization is necessary to prevent any electric field concentration at the triple junctions where vacuum, metal and ceramic are in close contact. A set of 10 alumina small specimens were manufactured to test the single ceramic tapped holes (Fig. 2). The specimens were made by Friatec F99.7: aluminum oxide purity 99.7%, density
Fig. 2. Tensile test on a threaded hole ceramic specimen M14.
The overall cantilevered weight of the beam source (BS) assembly is 150 kN. This implies not negligible mechanical loads transferred, through each insulator, from the BS to the anchor points located at ground potential. The ceramic insulators are loaded in different manner from the structural point of view; for this reason, the static load distributions due to the gravity have been
Fig. 3. 3D finite element (FE) static model of the electrostatic accelerator, 5 M elements. The most stressed insulators are located in the bottom part of the structure near the support of the tilting system.
Fig. 4. Sketch of the equipment designed to apply shear forces and bending moments to the insulator prototype
N. Pilan et al. / Fusion Engineering and Design 96–97 (2015) 563–567
565
Fig. 5. Insulator FEM models. Model “a” features a ceramic solid cylinder, model “b” a hollow cylinder, model “c” an “H” shape cylinder.
analyzed. A full 3D finite element model of the electrostatic accelerator has been developed to identify the forces and the moment acting on the most loaded insulator. Fig. 3 shows the results of FEM analyses. The most loaded insulators are located in the bottom part of the grounded frame where the accelerator leans against the BS tilting mechanism [1]. Table 1 shows the loads acting on the insulator ends. The distance between the flat surfaces of the flanges is 330 mm; the axial symmetry axis is orientated from flange “A” to flange “B.” The loads reported in Table 1 correspond to an equilibrated system of forces and moments.
Table 1 Structural (static) loads on the most stressed insulators. Flange “A” Fx [kN] Fy [kN] Fz [kN] Mx [MN mm] My [MN mm] Mz [MN mm]
4.55 −26.01 2.44 −0.7 0.059 2.69
Flange “B” −4.55 26.01 −2.44 −0.096 −0.059 −1.18
A dedicated tool has been designed and manufactured, to generate a set of equilibrated moments and forces according to the loads indicated in Table 1, in order to be applied in a standard tensile test machine (Figs. 4 and 6). The tool in Fig. 4 allows the application of the system of forces described by the following equations:
⎧ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨
FxA = −FxB = F · sin (˛) = T FyA = −FyB = F · cos (˛) = N FzA = FzB = 0 MxA = MxB = 0
MyA = MyB = 0 ⎪ ⎪ ⎪ ⎪ ⎪ MzA = FyA · a − FxA · c = F · [cos (˛) · a − sin (˛) · c] ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ MzB = FyB · b + FxB · c = −F · [cos (˛) · b + sin (˛) · c] ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ a−b ⎩ tan (˛) = L+2·c
(1)
566
N. Pilan et al. / Fusion Engineering and Design 96–97 (2015) 563–567
Fig. 6. Insulator prototype during the test.
Table 2 Insulator structural loads, reference design case. Face “A” Fx [kN] Fy [kN] Mz [MN mm]
Face “B” −9.66 −52.03 −2.37
9.66 52.03 5.56
The reference load case necessary to design and verify the insulator from the mechanical point of view has been obtained multiplying the loads reported in Table 1 by a factor −2 and the load orientations have been adjusted in order to have a system of load compatible with Eq. (1). The possible effects of dynamic loads, due to earthquakes and handlings, have been estimated amplifying the static stresses by a factor 2, while the negative sign has been adopted to consider a traction load case rather than a compressive one, since the brittle materials strength is lower when subjected to positive stresses. These considerations have generated the reference load case reported in Table 2. The load case, shown in Table 2, can be obtained if F = 52.9 kN, ˛ = 10.51◦ , a = 118 mm, b = 34.5 mm, and c = 60 mm. The minimal force F = 53 kN has been set as the acceptance level for the tensile tests. An additional detailed 3D FE structural model of the single insulator has also been developed in order to simulate the loading of the mechanical test. Fig. 5a shows the results in terms of 1st principal stress distribution 1 . The simulation was carried out considering all the details of the real geometry such as bolts pre-stresses, contacts and notch effects. Fig. 5b and c show alternative options, respectively a hollow cylinder and an “H shape.” The failure probability has been evaluated with Eq. (2) according to [5] in all models. Pf = 1 − exp
m 1 −
V
H · 0
+
m 2 H · 0
+
m 3 H · 0
dV
(2)
where i i = 1, 2, 3 are the principal stresses, H( i ) is a step function defined to account for the greater strength of the material under compressive loading, if i > 0, H = 1 else H = −8, 0 and m are, respectively, the material strength referred to 1 mm3 and the Weibull modulus. Assuming 0 = 250 MPa and m = 10 the failure probability results low (<10−4 ) in all cases, with regard to the reference load. A first set of 7 solid cylinders was manufactured; 5 insulators were tested from the mechanical point of view and one of them was tested electrically in vacuum. Moreover, a second set of 3 hollow cylinders was then procured, and two were mechanically tested. The mechanical tensile tests showed very different results between the two geometry solutions: the solid insulators got broken with a tensile force F in the order of 30–35 kN, thus lower than the goal. Instead, both the hollow cylinder prototypes passed successfully the mechanical test, withstanding an external load F = 80 kN, without any structural damage. The proof tests have been carried out applying the maximum load for 2 min being aware that subcritical crack growth (SCCG) [6,7] might raise some issues after a relatively long time. With this regard, the verification of the insulator lifetime considering SCCG is still under consideration, as it is also depending on radiation effects [8], whose details (type, “actual” quantity under ITER-relevant operation conditions, and related consequences on the material) are not entirely known. Ongoing discussion to justify completely the difference of mechanical performance highlighted so far the following points. In the solid cylinder geometry, the central hole for the dowel and the six threaded holes for the bolts have the same depth, and the broken section passes mostly through this area. In the hollow cylinder, the central hole represents a smoother boundary condition, since no notch effect is induced. In the solid cylinder insulator, the larger transversal “thickness” (i.e. the full diameter, uninterrupted by the central hole) may cause a less uniform temperature during the sinterization process, in particular during cooling down, leading to higher inhomogeneity of mechanical properties or even high residual stresses. With this regard, it has been found in literature [9] that the machining operations and the relevant residual stresses on a thick solid block of sintered B4 C, weakened the structural behavior of such big solid part. The analysis of these aspects leads to the two alternative configurations to the solid cylinder, already showing remarkable benefit. In both alternative geometries the thickness of the sintered “wall” has been limited to 50 mm, which is about the thickness of the single-thread samples described in Section 2. The numerical simulations show similar results from the mechanical point of view. The H-shape configuration has been taken into consideration for an overall optimization of the mechanical and electrical performances. The final structural validation of the design is progressing with the manufacturing of a larger number of insulators with the described geometries (Fig. 5), that shall complete the mechanical tests, with the same procedure already described.
4. Electrostatic design and prototype test The insulators have been design to withstand a nominal voltage of 200 kVdc in vacuum with a background pressure in the range 10–5 to 5.10–2 Pa; the insulator (and also each couple of electrodes inside the Beam Source vessel) has to work on the left side of the Paschen curve, i.e. with a product between the electrode gap length and pressure lower than 0.01 Pa m to prevent any gas discharge. The electric field distributions have been calculated for the solid insulator considering both electrostatic and resistive models. In both cases the electric fields at the triple junctions, in the solid alumina, along the interface ceramic-vacuum and the electric field on metal surfaces have been checked.
N. Pilan et al. / Fusion Engineering and Design 96–97 (2015) 563–567
567
The solid insulator prototype (Fig. 7) has been electrically tested in high vacuum at the high voltage Padova test facility (HVPTF) [11]. 240 kVdc were reached after 140 h of high voltage conditioning; such voltage was withstood without any breakdown for 6.5 h, no gases were injected during the test. Fig. 8 reports the whole conditioning history where voltage, pressure, X ray and current are reported as function of time. The same test is foreseen considering also the alternative geometries. 5. Conclusions A thorough R&D activity on the MITICA post insulator prototypes is being carried out. The insulator design has been numerically verified considering both mechanical and electrical aspects. Experimental validation has been started, with positive results in both involved fields. Alternative design solutions based on a reduction of the sintered ceramic thickness has been proposed and successfully tested from the mechanical point of view. The goal of the activity in the next period is the identification and validation of the final configuration, completing positively the electrical and mechanical tests. The conclusion of the R&D campaign will allow also the finalization of the technical specification for the procurement of the insulators for the MITICA Beam Source. Fig. 7. Electric field maps, electric field distribution the left side is energized at +200 kV while the right side is grounded. Picture of the model “a” insulator prototype during the HV tests.
Acknowledgments The project has been funded with support from Fusion for Energy. This publication reflects the views only of the author, and Fusion for Energy cannot be held responsible for any use which may be made of the information contained therein. The views and opinions expressed herein do not necessarily reflect those of the ITER Organization. References
Fig. 8. High Voltage conditioning history, voltage, pressure, X-ray and current signals vs. time (h).
The electric field in the sintered ceramic has been limited to 8.5 kV/mm, considering a safety factor 2, on the alumina dielectric strength, which is 17 kV/mm as reported in [3]. The electric fields along the insulator surface and in the triple junctions have been limited according to the data reported in Table 1 of [10] Fig. 7 shows the electrostatic field map distribution calculated by the stationary finite element 2D axial-symmetric model during the high voltage test (model “a”). The left side of the insulator is connected to a positive power supply while the other is connected to ground.
[1] P. Zaccaria, et al., Progress in the MITICA beam source design, Rev. Sci. Instrum. 83 (2012) 02B108. [2] R. Hemsworth, et al., Status of the ITER heating neutral beam system, Nucl. Fusion 49 (2009) 045006. [3] C EI EN 60672-3, Ceramic and glass-insulating materials. Part 3. Specifications for individual materials (1994). [4] ITER Handbook Materials. [5] P. Stanley, et al., “Application of the 4th Weibull Equations in the Design of Brittle Components” Fracture Mechanics of Ceramics 3, Plenum Press, New York, 1978. [6] A.H. De Aza, et al., Crack growth resistance of alumina, zirconia and zirconia toughened alumina ceramics for joint prostheses, Biomaterials 23 (2002) 937–945. [7] J.-L. Le, Z.P. Baˇzant, M.Z. Bazant, Subcritical crack growth law and its consequences for lifetime statistics and size effect of quasibrittle structures, J. Phys. D: Appl. Phys. 42 (2009) 21. [8] G.P. Pells, Boothby, The effects of ␥-irr[iation on subcritical crack growth in alumina, J. Nucl. Mater. 256 (1998) 25–34. [9] D.W. Richerson, Modern Ceramic Engineering: Properties, Processing, and Use in Design, 2nd ed., New York, M. Dekker, 1992. [10] A. Masiello, Adaptation of the 1 MV bushing to the SINGAP concept for the ITER NB injector test bed, Nucl. Fusion 46 (2006) S340. [11] A. De Lorenzi, et al., HVPTF – the high voltage laboratory for the ITER Neutral Beam test facility, Fusion Eng. Des. 86 (2011) 6–8.
Contents lists available at ScienceDirect
Fusion Engineering and Design journal homepage: www.elsevier.com/locate/fusengdes
Electrical and structural R&D activities on high voltage dc solid insulator in vacuum N. Pilan a,∗ , D. Marcuzzi a , A. Rizzolo a , L. Grando a , G. Gambetta a , S. Dalla Rosa c , V. Kraemer d , T. Quirmbach d , G. Chitarin a , R. Gobbo b , G. Pesavento b , A. De Lorenzi a , L. Lotto a , R. Rizzieri a , M. Fincato a , L. Romanato a , L. Trevisan a , V. Cervaro a , L. Franchin a a
Consorzio RFX, Associazione EURATOM-ENEA sulla Fusione, Corso Stati Uniti 4, I-35127 Padova, Italy DII, Università di Padova, v. Gradenigo 6/A, I-35131 Padova, Italy c Umicore – Italbras S.p.A., Strada del Balsego, n.6, 36100 Vicenza, Italy d FRIATEC Ceramics Division, Steinzeugstrasse 50, 68229 Mannheim, Germany b
h i g h l i g h t s • • • •
A thorough R&D activity on the MITICA post insulator prototypes is being carried out. The design has been numerically verified considering both mechanical and electrical aspects. Experimental validation has been started, with positive results in both involved fields. Alternative design solutions thickness have been proposed and successfully tested.
a r t i c l e
i n f o
Article history: Received 19 September 2014 Received in revised form 18 May 2015 Accepted 20 May 2015 Available online 30 May 2015 Keywords: Insulator High voltage Ceramic Vacuum Electrostatic accelerator
a b s t r a c t This paper describes the R&D work performed in support of the design of the alumina insulators for the MITICA Neutral Beam Injector. The ceramic insulators are critical elements, both from the structural and electrical point of view, of the 1 MV electrostatic accelerator of the MITICA injector, as they are required to sustain both the mechanical loads due to the cantilevered weight of the ion source and the high electric field between the accelerator grids. This paper presents the results of numerical simulations and experimental tests on prototypes that have been carried out to validate the insulator design under realistic operating conditions. © 2015 Consorzio RFX. Published by Elsevier B.V. All rights reserved.
1. Introduction The MITICA electrostatic accelerator is based on the MAMuG (multi-aperture-multi-grid concept [1,2]). This accelerator is composed of five stages and includes seven main grids, polarized at potentials ranging from −1 MV to ground (0 V) with five main stages of 200 kV each (see Fig. 1). The negative ion beam, extracted from the RF ion source, is composed of 1280 individual beamlets which are guided and focused by an intense electric field through the apertures of the accelerating grids. The electrical insulation between the grids relies, both
∗ Corresponding author. E-mail address: [email protected] (N. Pilan). http://dx.doi.org/10.1016/j.fusengdes.2015.05.057 0920-3796/© 2015 Consorzio RFX. Published by Elsevier B.V. All rights reserved.
on vacuum gaps and ceramic insulators which are part of the sustaining structure. Five sets of alumina cylindrical insulators are located between the accelerator grids; the insulators are therefore subjected to 200 kVdc and to structural loads. The ion source and the multi-stage accelerator constitute a cantilevered structure sustained from the grounded stage. The present work describes the R&D activities carried out to identify a reliable insulator design considering several constraints: withstanding voltage, structural and manufacturing issues. 2. Insulator description The insulator is a cylinder made of high purity alumina C799 [3] according with the ITER requirements [4], the cylinder has a diameter of 130 mm and a length of 280 mm.
564
N. Pilan et al. / Fusion Engineering and Design 96–97 (2015) 563–567
3.9–3.95 g/cm3 , average grain size 10–15 m. A M14 carbon steel screw (class 12.8) has been screwed in the tapped hole in the alumina and it has been pulled by a tensile machine up to 100 kN (the limit of the machine) without any issues. The same thread has been subjected to a severe fatigue test: an oscillating load between 65 and 85 kN has been applied until failure, which occurred in the screw (made by carbon steel) at about half million cycles. The functionality of the threaded ceramic hole was not compromised. The second phase of the R&D campaign was focused on the manufacturing and testing of the full prototype, for a thorough mechanical and electrical validation of the design, with reference to the numerical verification described in the next sections. 3. Structural design and prototype tests
Fig. 1. Vertical section view of the half MITICA beam source electrostatic accelerator. The insulators are the horizontal axis cylinders in light blue. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)
Each ceramic insulator is connected to the structure by means of a couple of stainless steel (SS) flanges which are directly bolted on the frames sustaining the accelerating grids. A central dowel is present on each flange to assure coaxial positioning between the stainless steel flanges and to prevent any relative displacement due to shear stresses between ceramic and metal flanges. The low friction coefficient between ceramic and SS is not sufficient to guarantee a rigid bolted joint between the two materials. The joining between metal flanges and ceramic insulator is not a straightforward choice because both electrostatic and structural requirements have to be satisfied at the same time. In the reference configuration, the alumina is joined to the SS flanges by screws, six threaded holes M14 are manufactured directly on the sintered ceramic; the blind holes are metalized with a suitable conductive layer as well as the interface between the ceramic and the metallic flange. The metallization is necessary to prevent any electric field concentration at the triple junctions where vacuum, metal and ceramic are in close contact. A set of 10 alumina small specimens were manufactured to test the single ceramic tapped holes (Fig. 2). The specimens were made by Friatec F99.7: aluminum oxide purity 99.7%, density
Fig. 2. Tensile test on a threaded hole ceramic specimen M14.
The overall cantilevered weight of the beam source (BS) assembly is 150 kN. This implies not negligible mechanical loads transferred, through each insulator, from the BS to the anchor points located at ground potential. The ceramic insulators are loaded in different manner from the structural point of view; for this reason, the static load distributions due to the gravity have been
Fig. 3. 3D finite element (FE) static model of the electrostatic accelerator, 5 M elements. The most stressed insulators are located in the bottom part of the structure near the support of the tilting system.
Fig. 4. Sketch of the equipment designed to apply shear forces and bending moments to the insulator prototype
N. Pilan et al. / Fusion Engineering and Design 96–97 (2015) 563–567
565
Fig. 5. Insulator FEM models. Model “a” features a ceramic solid cylinder, model “b” a hollow cylinder, model “c” an “H” shape cylinder.
analyzed. A full 3D finite element model of the electrostatic accelerator has been developed to identify the forces and the moment acting on the most loaded insulator. Fig. 3 shows the results of FEM analyses. The most loaded insulators are located in the bottom part of the grounded frame where the accelerator leans against the BS tilting mechanism [1]. Table 1 shows the loads acting on the insulator ends. The distance between the flat surfaces of the flanges is 330 mm; the axial symmetry axis is orientated from flange “A” to flange “B.” The loads reported in Table 1 correspond to an equilibrated system of forces and moments.
Table 1 Structural (static) loads on the most stressed insulators. Flange “A” Fx [kN] Fy [kN] Fz [kN] Mx [MN mm] My [MN mm] Mz [MN mm]
4.55 −26.01 2.44 −0.7 0.059 2.69
Flange “B” −4.55 26.01 −2.44 −0.096 −0.059 −1.18
A dedicated tool has been designed and manufactured, to generate a set of equilibrated moments and forces according to the loads indicated in Table 1, in order to be applied in a standard tensile test machine (Figs. 4 and 6). The tool in Fig. 4 allows the application of the system of forces described by the following equations:
⎧ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨
FxA = −FxB = F · sin (˛) = T FyA = −FyB = F · cos (˛) = N FzA = FzB = 0 MxA = MxB = 0
MyA = MyB = 0 ⎪ ⎪ ⎪ ⎪ ⎪ MzA = FyA · a − FxA · c = F · [cos (˛) · a − sin (˛) · c] ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ MzB = FyB · b + FxB · c = −F · [cos (˛) · b + sin (˛) · c] ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ a−b ⎩ tan (˛) = L+2·c
(1)
566
N. Pilan et al. / Fusion Engineering and Design 96–97 (2015) 563–567
Fig. 6. Insulator prototype during the test.
Table 2 Insulator structural loads, reference design case. Face “A” Fx [kN] Fy [kN] Mz [MN mm]
Face “B” −9.66 −52.03 −2.37
9.66 52.03 5.56
The reference load case necessary to design and verify the insulator from the mechanical point of view has been obtained multiplying the loads reported in Table 1 by a factor −2 and the load orientations have been adjusted in order to have a system of load compatible with Eq. (1). The possible effects of dynamic loads, due to earthquakes and handlings, have been estimated amplifying the static stresses by a factor 2, while the negative sign has been adopted to consider a traction load case rather than a compressive one, since the brittle materials strength is lower when subjected to positive stresses. These considerations have generated the reference load case reported in Table 2. The load case, shown in Table 2, can be obtained if F = 52.9 kN, ˛ = 10.51◦ , a = 118 mm, b = 34.5 mm, and c = 60 mm. The minimal force F = 53 kN has been set as the acceptance level for the tensile tests. An additional detailed 3D FE structural model of the single insulator has also been developed in order to simulate the loading of the mechanical test. Fig. 5a shows the results in terms of 1st principal stress distribution 1 . The simulation was carried out considering all the details of the real geometry such as bolts pre-stresses, contacts and notch effects. Fig. 5b and c show alternative options, respectively a hollow cylinder and an “H shape.” The failure probability has been evaluated with Eq. (2) according to [5] in all models. Pf = 1 − exp
m 1 −
V
H · 0
+
m 2 H · 0
+
m 3 H · 0
dV
(2)
where i i = 1, 2, 3 are the principal stresses, H( i ) is a step function defined to account for the greater strength of the material under compressive loading, if i > 0, H = 1 else H = −8, 0 and m are, respectively, the material strength referred to 1 mm3 and the Weibull modulus. Assuming 0 = 250 MPa and m = 10 the failure probability results low (<10−4 ) in all cases, with regard to the reference load. A first set of 7 solid cylinders was manufactured; 5 insulators were tested from the mechanical point of view and one of them was tested electrically in vacuum. Moreover, a second set of 3 hollow cylinders was then procured, and two were mechanically tested. The mechanical tensile tests showed very different results between the two geometry solutions: the solid insulators got broken with a tensile force F in the order of 30–35 kN, thus lower than the goal. Instead, both the hollow cylinder prototypes passed successfully the mechanical test, withstanding an external load F = 80 kN, without any structural damage. The proof tests have been carried out applying the maximum load for 2 min being aware that subcritical crack growth (SCCG) [6,7] might raise some issues after a relatively long time. With this regard, the verification of the insulator lifetime considering SCCG is still under consideration, as it is also depending on radiation effects [8], whose details (type, “actual” quantity under ITER-relevant operation conditions, and related consequences on the material) are not entirely known. Ongoing discussion to justify completely the difference of mechanical performance highlighted so far the following points. In the solid cylinder geometry, the central hole for the dowel and the six threaded holes for the bolts have the same depth, and the broken section passes mostly through this area. In the hollow cylinder, the central hole represents a smoother boundary condition, since no notch effect is induced. In the solid cylinder insulator, the larger transversal “thickness” (i.e. the full diameter, uninterrupted by the central hole) may cause a less uniform temperature during the sinterization process, in particular during cooling down, leading to higher inhomogeneity of mechanical properties or even high residual stresses. With this regard, it has been found in literature [9] that the machining operations and the relevant residual stresses on a thick solid block of sintered B4 C, weakened the structural behavior of such big solid part. The analysis of these aspects leads to the two alternative configurations to the solid cylinder, already showing remarkable benefit. In both alternative geometries the thickness of the sintered “wall” has been limited to 50 mm, which is about the thickness of the single-thread samples described in Section 2. The numerical simulations show similar results from the mechanical point of view. The H-shape configuration has been taken into consideration for an overall optimization of the mechanical and electrical performances. The final structural validation of the design is progressing with the manufacturing of a larger number of insulators with the described geometries (Fig. 5), that shall complete the mechanical tests, with the same procedure already described.
4. Electrostatic design and prototype test The insulators have been design to withstand a nominal voltage of 200 kVdc in vacuum with a background pressure in the range 10–5 to 5.10–2 Pa; the insulator (and also each couple of electrodes inside the Beam Source vessel) has to work on the left side of the Paschen curve, i.e. with a product between the electrode gap length and pressure lower than 0.01 Pa m to prevent any gas discharge. The electric field distributions have been calculated for the solid insulator considering both electrostatic and resistive models. In both cases the electric fields at the triple junctions, in the solid alumina, along the interface ceramic-vacuum and the electric field on metal surfaces have been checked.
N. Pilan et al. / Fusion Engineering and Design 96–97 (2015) 563–567
567
The solid insulator prototype (Fig. 7) has been electrically tested in high vacuum at the high voltage Padova test facility (HVPTF) [11]. 240 kVdc were reached after 140 h of high voltage conditioning; such voltage was withstood without any breakdown for 6.5 h, no gases were injected during the test. Fig. 8 reports the whole conditioning history where voltage, pressure, X ray and current are reported as function of time. The same test is foreseen considering also the alternative geometries. 5. Conclusions A thorough R&D activity on the MITICA post insulator prototypes is being carried out. The insulator design has been numerically verified considering both mechanical and electrical aspects. Experimental validation has been started, with positive results in both involved fields. Alternative design solutions based on a reduction of the sintered ceramic thickness has been proposed and successfully tested from the mechanical point of view. The goal of the activity in the next period is the identification and validation of the final configuration, completing positively the electrical and mechanical tests. The conclusion of the R&D campaign will allow also the finalization of the technical specification for the procurement of the insulators for the MITICA Beam Source. Fig. 7. Electric field maps, electric field distribution the left side is energized at +200 kV while the right side is grounded. Picture of the model “a” insulator prototype during the HV tests.
Acknowledgments The project has been funded with support from Fusion for Energy. This publication reflects the views only of the author, and Fusion for Energy cannot be held responsible for any use which may be made of the information contained therein. The views and opinions expressed herein do not necessarily reflect those of the ITER Organization. References
Fig. 8. High Voltage conditioning history, voltage, pressure, X-ray and current signals vs. time (h).
The electric field in the sintered ceramic has been limited to 8.5 kV/mm, considering a safety factor 2, on the alumina dielectric strength, which is 17 kV/mm as reported in [3]. The electric fields along the insulator surface and in the triple junctions have been limited according to the data reported in Table 1 of [10] Fig. 7 shows the electrostatic field map distribution calculated by the stationary finite element 2D axial-symmetric model during the high voltage test (model “a”). The left side of the insulator is connected to a positive power supply while the other is connected to ground.
[1] P. Zaccaria, et al., Progress in the MITICA beam source design, Rev. Sci. Instrum. 83 (2012) 02B108. [2] R. Hemsworth, et al., Status of the ITER heating neutral beam system, Nucl. Fusion 49 (2009) 045006. [3] C EI EN 60672-3, Ceramic and glass-insulating materials. Part 3. Specifications for individual materials (1994). [4] ITER Handbook Materials. [5] P. Stanley, et al., “Application of the 4th Weibull Equations in the Design of Brittle Components” Fracture Mechanics of Ceramics 3, Plenum Press, New York, 1978. [6] A.H. De Aza, et al., Crack growth resistance of alumina, zirconia and zirconia toughened alumina ceramics for joint prostheses, Biomaterials 23 (2002) 937–945. [7] J.-L. Le, Z.P. Baˇzant, M.Z. Bazant, Subcritical crack growth law and its consequences for lifetime statistics and size effect of quasibrittle structures, J. Phys. D: Appl. Phys. 42 (2009) 21. [8] G.P. Pells, Boothby, The effects of ␥-irr[iation on subcritical crack growth in alumina, J. Nucl. Mater. 256 (1998) 25–34. [9] D.W. Richerson, Modern Ceramic Engineering: Properties, Processing, and Use in Design, 2nd ed., New York, M. Dekker, 1992. [10] A. Masiello, Adaptation of the 1 MV bushing to the SINGAP concept for the ITER NB injector test bed, Nucl. Fusion 46 (2006) S340. [11] A. De Lorenzi, et al., HVPTF – the high voltage laboratory for the ITER Neutral Beam test facility, Fusion Eng. Des. 86 (2011) 6–8.