- Email: [email protected]
Divertor design for HL-2A tokamak modification
Journal of Nuclear Materials 415 (2011) S952–S956
Contents lists available at ScienceDirect
Journal of Nuclear Materials journal homepage: www.elsevier.com/locate/jnucmat
Divertor design for HL-2A tokamak modification Y.D. Pan ⇑, J.H. Zhang, W. Li, J.X. Li Southwestern Institute of Physics, P.O. Box 432, Chengdu 610041, China
a r t i c l e
i n f o
Article history: Available online 27 November 2010
a b s t r a c t This article introduces the new divertor design for the planned upgrade to HL-2A and details the optimization of the divertor parameters, including the target geometry, throat width, and the size and location of the pumping chamber entrance. The basic requirements are to provide tolerable power loads on the target plates and sufficient particle exhaust for the Ip = 1.2MA, 10–20 MW auxiliary heated reference discharge. The SOLPS5.0 code package is used to generate a database. A vertical target divertor configuration has been adopted as the first conceptual design, which gives a peak heat load on the divertor target of 4– 10 MW/m2 for an anticipated power flux into the SOL of 5–10 MW discharge. The pumping efficiency is found to increase by a factor of 3–4 when the divertor gap is narrowed from 5.5 to 2.0 cm. Ó 2010 Elsevier B.V. All rights reserved.
1. Introduction
2. Divertor geometry for HL-2A upgraded machine
The upgrade of the HL-2A tokamak to a modified device, HL-2M is a major challenge for the Southwestern Institute of Physics (SWIP). Its key divertor design technology research is quite important for the project. The planned tokamak machine is designed to operate with shaped plasma cross-sections and in the single null (SN) divertor configuration with major radius/minor radius/toroidal at R/a/Bt = 1.85 m/0.5 m/2.5 T. The divertor should be designed to accommodate the ohmic power coming from 1.2 MA plasma current and the auxiliary heating power ECRH + LHCD + NBI at the level of 10–20 MW in 5 s discharges. During the last decade, HL-2A, formerly the ASDEX machine, has some expected benefits of a closed divertor geometry, such as target detachment [1]. However, limitations of the overall device, such as inflexible plasma configuration, weak neutral pumping and mechanical trouble from the divertor shaping coils inside the vacuum vessel, restricted the advanced research programme. Now, the new project tries to slightly open the divertor while increasing the neutral particle pumping in the divertor region and decrease the fraction of recycled neutrals escaping to the main chamber. This article introduces the new divertor conceptual design and details the optimization of the divertor geometry, pumping port and dome structure. Providing tolerable power loads on the target plates and sufficient particle exhaust is the basic requirement of the design. The SOLPS5.0 suite of codes, is the main tool employed during the divertor design process [2,3].
The upgraded divertor will have deep, inclined inner and outer vertical plates with tightly fitting dome baffle in the private flux region to minimize its conductance for neutral leakage from the divertor region into the main chamber. The chosen divertor geometry is compatible with a wide range of lower triangularity from 0.3 to 0.6. This is consideration to left more flexible to new conception divertor design in the near future. Fig. 1 shows the separatrix locations for the low and high-triangularity magnetic geometries. The width of the throat, the leg length and the incident angle of target plate marked in Fig. 2 are the three key divertor geometry parameters that are considered here [4]. The width of baffle throat is fixed by the width of SOL layer at mid-plane. Incident angle of target plate is related to the leg length and neutral particle transport through the emitted direction and incident angle. The strength of the neutral exhaust depends most strongly on the location and size of the pumping opening, the pump speed. The magnetic equilibrium, which is the basis of the numerical grid generated for SOLPS5.0, is benchmarked by EFIT, SWEQU and the free boundary equilibrium code TSC [5]. The shape of the baffle follows the magnetic flux surface which is about 3–5 cm away from the separatrix mapped to the outer mid-plane. This distance is anticipated to be a few times of radial power decay length, kp . The targets and baffles will consider CFC as the main material in the initial phase.
⇑ Corresponding author. Tel.: +86 28 82850339; fax: +86 28 82850300. E-mail address: [email protected] (Y.D. Pan). 0022-3115/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jnucmat.2010.11.004
3. Edge plasma simulation The new design has the capability to accommodate lower single null (LSN) and double null (DN) configurations, but LSN will be the preferred option. The x-point location is about 90 cm below
Y.D. Pan et al. / Journal of Nuclear Materials 415 (2011) S952–S956
Fig. 1. Divertor geometry with different lower triangularity 0.46 and 0.60.
S953
the mid-plane. The divertor design is optimized for a triangularity of 0.46. Half of the referenced computational mesh is shown in Fig. 2. The computational domain covers the whole scrape-off layer and lower divertor. A small region of the plasma core periphery and the private flux region are also included. The whole computational domain is resolved into 96 poloidal divisions and 24 radial divisions. Deuterium will be used as the plasma species. For the standard operating case, it is assumed that all of the non-radiated power losses from the main plasma are transported uniformly across the LCFS. A simply estimation suggests that the power entering the SOL is about half of the total auxiliary and ohmic heating power, 5 MW–10 MW. Classical values are assumed for the parallel i e transport coefficient, gijj ; kjj and kjj , and also for the equipartition coefficient k. The anomalous perpendicular transport model used in the present study is constant in space, with the thermal diffusivities xe? ¼ xi? ¼ 1:0m2 =s and the particle diffusivity D = 0.4 m2/s. The wall pumping effect is neglected and the recycling coefficient R is set to 1.0 for the divertor plasma facing components and the vacuum vessel wall. The recycling effect at pumping port is decided by the pumping speed. In this initial study, two possible pumping ports are considered, see Fig. 2, where DIV-1 is in the poloidal port and DIV-2 is in the bottom of vessel. The cryo-pumps are considered as a necessary part of general pumping system, for density control and vessel condition. The pumping speed of the pump is used to set the recycling coefficient at the pump opening surface in the simulation. EIRENE [6,7] is the neutral particle transport code within SOLPS, and the manual states that
Fig. 2. Key parameters of HL-2M SN divertor geometry (unit: mm).
S954
Y.D. Pan et al. / Journal of Nuclear Materials 415 (2011) S952–S956
pffiffiffiffiffiffiffiffiffiffi L ¼ A ð1 RECYCTÞ 3:638 T=m
ð1Þ Exhaust Atomic Flux (s-1)
where A is the surface area (cm2) of pumping port. RECYCT is the recycling coefficient at this pumping port. L (liters/s) is the pumping speed for particles with temperature T(K) and mass m (AMU). If the height of pump opening is 25 cm, then for room temperature, D atom, a pumping speed of 125 m3/s gives RECYCT = 0.9 at the pumping opening.
5.0E+20 4.0E+20 3.0E+20
DIV-2 with Wider Gap DIV-2 with Narrow Gap DIV-1 with Wider Gap
2.0E+20 1.0E+20 0.0E+00
1.0E+19
2.0E+19
3.0E+19
4.0E+19
Separatrix ne (m-3)
4. Results of the B2-EIRENE prediction
Fig. 4. Exhaust atom flux vs. separatrix density at mid-plane when Psol = 5 MW in DIV-1 (poloidal port) and DIV-2 bottom port with different gap distance.
4.1. Power density profile The power surface density profiles at the inner/outer targets, as calculated by SOLPS, are shown in Fig. 3 for Psol = 5,7.5 and 10 MW. Here, the edge ne which 5 cm inside the separatrix surface at midplane is fixed at 2 1019 m3. The power densities at outer or inner target are shown in the opposite side of the Y-axis in order to see it clearly. DIV-1 and DIV-2 indicates two different pumping opening which defined in the Section 2. The power density is strongly peaked near the strike point. The full width at the half height of the profile is some extend of wider with the increase of Psol, but the width is less than 3 cm in each case. When Psol is 5 MW, the peak power density at outer/inner strike point is 2.2/ 5.9 MW/m2 for DIV-1 and 4.3/4.4 MW/m2 for DIV-2. It shows the strong asymmetry should be considered if we use DIV-1 as the pump opening. If we chose DIV-2, the peaked power density at outer strike points are 4.26/7.35/10.12 MW/m2 for the Psol is 5/7.5/ 10 MW. The corresponding peaked power densities at inner strike points are 4.42/7.73/11.0 MW/m2. The in–out asymmetry is much lower when the pump opening is in the bottom of VV. 4.2. Particle exhaust optimization The particle exhaust throughout is decided by many factors, including the position of the pump opening, the pumping speed and the slot of the gap. Fig. 4 shows the rate of particle exhaust at the pump openings. Here, the separatrix electron density at mid-plane is scanned and the power entering the SOL is fixed at 5 MW. The pump surface recycling coefficient is 0.9. With the increase of upstream density,
the exhaust atomic flux in DIV-2 is larger than that in DIV-1. DIV-2 has obviously advantage due to its proximity to the particle source. The total neutral particle flux into the core plasma, from the backflow of divertor recycling, is minimized by optimizing the width of the baffle throat. In addition, DIII-D experiments have shown that significant increases in plenum pressure can be obtained if careful attention is paid to the position of the separatrix with respect to the baffle [8]. The slot width, ‘‘gap’’ here, is defined as the distance between the dome and vertical target. And it is influences the back flow of neutrals. For testing the effect of these parameters, the upstream power and edge electron density were kept fixed at 5 MW and edge ne = 4 1019 m3. Here, The edge ne means the magnetic surface which 5 cm inside the separatrix surface at mid-plane. In this condition, the edge plasma will run far from the lower recycling region. The relation between the exhaust atom flux and the slot width is shown in Fig. 5. As long as the slot width is larger than 5.5 cm, the dependence of the exhaust rate on slot width is weak and the exhausted atom flux is about constant 2.0–2.5 1020 m3. Once the slot width is less than this, the exhaust atom flux peaks and then drops as the width shrinks. The optimal width is 2.0 cm with the pumping efficiency increasing by a factor of 3–4 as the gap narrows from 5 to 2.0 cm. The reason for this enhancement cannot be explained simply. Detailed neutral Monte Carlo modeling EIRENE shows narrow slot will induce enhanced pressure in the baffle and the neutral particle loss. The peak heat deposition on the target was largely independent of the slot width, as was the upstream electron density. 4.3. Divertor operation region To reduce the power load and erosion of the divertor target plates is the main objective in the design of the HL-2M divertor.
Div-1_innerplate 5MW Div-2-innerplate 5MW Div-1_outerplate 5MW Div-2-outerplate 5MW Div-2 innerplate 7.5MW dIV-2 outerplate 7.5MW Div-2 innerplate 10MW Div-2 outerplate 10MW
1.E+07
5.E+06
1.E+21
0.E+00 -0.05
0.00
0.05
0.10
0.15
0.20
0.25
-5.E+06
-1.E+07
-2.E+07
Exhaust Atom Flux
Heat flux at target plates (w/m2)
2.E+07
8.E+20 6.E+20 4.E+20 2.E+20 0.E+00
Position along the target plate(m)
0
20
40
60
80
100
120
Gap Distance (mm) Fig. 3. Power density profiles at target plates with pumping port at DIV-1 (poloidal port) or DIV-2 (bottom port) separately. The total power which enters into SOL is 5/ 7.5/10 MW. The edge ne which 5 cm inside the separatrix surface at mid-plane is 2 1019 m3.
Fig. 5. Exhaust atom flux as a function of dome-to-target distance when Psol = 5 MW, edge density ne = 4 1019 m3 is 5 cm inside the separatrix surface at outer mid-plane.
Y.D. Pan et al. / Journal of Nuclear Materials 415 (2011) S952–S956
1.0
fm
0.6
0.69
1
0.89 0.98
9.50E-01
fm_inner target fm_outer target 0.1 1
10
100
Tt (eV) Fig. 6. The pressure factor vs. Te at the strike point of the inner/outer target plates. When fm is 0.95,1.0 and 0.89. The Te/nesepm at the strike point of outer target plate and ne at outer mid-plane separatrix surface are 49 eV/2.17e19 m3, 16 eV/ 3.6e19 m3 and 4.7 eV/4.57 m3.
It is important to operate in the high recycling or detachment regimes to decrease the heat flux flowing to the target by making the electron temperature low in the target plate. Divertor operation regimes can be very sensitive to the mid-plane separatrix electron density. According to the Greenwald density limitation, HL2M will be able to run safely with the line average densities up to 1.5 1020 m3 in ohmic discharges. As a measure to quantify for R detachment, the pressure fraction
at mid-plane and target, fm ¼ Rtargetplate
midplane
pð1þcM 2 Þdw
S955
plate ‘Tesepa’ and the electron density at outer mid-plane separatrix surface ‘nesepm’ are changed from 49 eV/2.17e19 m3 to 16 eV/3.6e19 m3 and then 4.7 eV/4.57e19 m3. We think fm = 1 is the division of high recycle and detachment. Here, ne at outer mid-plane separatrix surface is 3.6e19 m3. High density and low temperature (10 eV < Tesepa < 50 eV) are shown at target plates in the high-recycling regime. When nesepm increases larger and larger, the detachment becomes more and more. When nesepm is quite lower, only the low recycling regime can be attained. In this regime, due to the high parallel heat conduction, the electron temperature shows little drop along the field lines. In this region, recycling losses can be neglected, so the peak heat flux at the target is related to the power enters the SOL layer. In contrast to the low recycling regime, flow reversal of the D ions, which generally characterizes the high-recycling regime, is observed in the region close to the strike point. Fig. 7 is the mach number profiles when nesepm is 3.37e19 m3 (left) and 5.51e19 m3 (right). When edge plasma runs from high recycling phase (left figure in Fig. 7) to detachment (right figure in Fig. 7), the D ions flow reversal disappears. The negative mach number region near the strike point of target plate completely becomes the positive direction. The backflow is the important source of impurities.
, is an important va-
pð1þcM 2 Þdw
lue. It is calculated from the flux integrated dw pressure profiles
5. Summary
at the target plate and mid-plane. c is adiabatic coefficient and M is mach number. The effect of the acceleration at the sheath entrance appears as the factor cM2. Significant Te and ne gradients along field lines are shown with the increasing of the fm. The pressure factor fm is approach to 1. It means the momentum almost has been lost completely along the mid-plane to target. The relation of fm to the electron temperature Te of the strike points at the target plates is shown in Fig. 6. The pressure factor is decided by different competitive mechanism. The pressure profile at the mid-plane is related to the anomalous radial transport and the pressure profile near the target plate has strong relation with the ionization and CX neutral losses in the divertor region. It is difficulty to divide the divertor operation regions simply. When fm at outer target changes from 0.95 to 1.0 and then decreases to 0.89, The temperature at the strike point of outer target
In this design study, the strictly power deposition profiles on the both target plates are shown as a function of the power entering the SOL, with a focus on the peak power loads and the possible neutral back flow when divertor runs from high recycling phases to detachment. Target detachment was observed in the SOLPS calculations when the separatrix plasma density was high and it has strong effect on the symmetry and stability of the divertor plasma. During the transition, the momentum loss is another key parameter which determines the possible reverse of edge flow. It has strong effect on the symmetry and stability of the divertor plasma and finally on the whole edge region. With an anticipated SOL power flux of 5–10 MW discharge, the peak heat load on the divertor target is evaluated to be 4–10 MW/m2. To reduce neutral leakage from the divertor and to increase the PFR neutral pressure, which enhances divertor pumping, a PFR baffle structure (dome) is introduced in the design. The gaps
Fig. 7. Mach number profiles when nesepm = 3.37e19 m3 (left) and 5.51e19 m3 (right).
S956
Y.D. Pan et al. / Journal of Nuclear Materials 415 (2011) S952–S956
between the vertical targets and dome also are optimized, so that they maximize the neutral flux that the divertor pump. The pumping efficiency is found to increase by a factor of 3–4 when the divertor gap is narrowed from 5.5 to 2.0 cm. Acknowledgements Dr. R. Schneider, D. Coster, S.Z. Zhu and Y.P. Chen have provided many valuable contributions to this work. This work is supported by the National Magnetic Confinement Fusion Science Program, No. 2009GB104008 and National Natural Science Foundations of China and Grand No. 10805016.
References [1] Y.D. Pan, R. Schneider, J. Nucl. Mater. 407 (2007) 363–365. [2] B.J. Braams, A multi-fluid code for simulation of the edge plasma in tokamaks. NET Report No. 68, 1987. [3] R. Schneider, H.S. Bosch, J. Neuhauser, et al., J. Nucl. Mater. 241–243 (1997) 701–706. [4] S. Sakurai, K. Shimizu, K. Masaki, et al., PPCF 44 (2002) 49–760. [5] Y.D. Pan, S.C. Jardin, C. Kessel, The discharge design of HL-2M with the Tokamak Simulation Code (TSC), Princeton University Plasma Physic Laboratory Report PPL-4257, OCT, 2007. [6] D. Reiter, J. Nucl. Mater. 196–198 (1992) 241. [7] D. Reiter. The EIRENE Code, Version Jan. 92, User Manual, March 1992. [8] M.A. Mahdavi, S.L. Allen, D.R. Baker, B. Bastasz, N.H. Brooks, et al., J. Nucl. Mater. 220–222 (April) (1995) 13–24.
Contents lists available at ScienceDirect
Journal of Nuclear Materials journal homepage: www.elsevier.com/locate/jnucmat
Divertor design for HL-2A tokamak modification Y.D. Pan ⇑, J.H. Zhang, W. Li, J.X. Li Southwestern Institute of Physics, P.O. Box 432, Chengdu 610041, China
a r t i c l e
i n f o
Article history: Available online 27 November 2010
a b s t r a c t This article introduces the new divertor design for the planned upgrade to HL-2A and details the optimization of the divertor parameters, including the target geometry, throat width, and the size and location of the pumping chamber entrance. The basic requirements are to provide tolerable power loads on the target plates and sufficient particle exhaust for the Ip = 1.2MA, 10–20 MW auxiliary heated reference discharge. The SOLPS5.0 code package is used to generate a database. A vertical target divertor configuration has been adopted as the first conceptual design, which gives a peak heat load on the divertor target of 4– 10 MW/m2 for an anticipated power flux into the SOL of 5–10 MW discharge. The pumping efficiency is found to increase by a factor of 3–4 when the divertor gap is narrowed from 5.5 to 2.0 cm. Ó 2010 Elsevier B.V. All rights reserved.
1. Introduction
2. Divertor geometry for HL-2A upgraded machine
The upgrade of the HL-2A tokamak to a modified device, HL-2M is a major challenge for the Southwestern Institute of Physics (SWIP). Its key divertor design technology research is quite important for the project. The planned tokamak machine is designed to operate with shaped plasma cross-sections and in the single null (SN) divertor configuration with major radius/minor radius/toroidal at R/a/Bt = 1.85 m/0.5 m/2.5 T. The divertor should be designed to accommodate the ohmic power coming from 1.2 MA plasma current and the auxiliary heating power ECRH + LHCD + NBI at the level of 10–20 MW in 5 s discharges. During the last decade, HL-2A, formerly the ASDEX machine, has some expected benefits of a closed divertor geometry, such as target detachment [1]. However, limitations of the overall device, such as inflexible plasma configuration, weak neutral pumping and mechanical trouble from the divertor shaping coils inside the vacuum vessel, restricted the advanced research programme. Now, the new project tries to slightly open the divertor while increasing the neutral particle pumping in the divertor region and decrease the fraction of recycled neutrals escaping to the main chamber. This article introduces the new divertor conceptual design and details the optimization of the divertor geometry, pumping port and dome structure. Providing tolerable power loads on the target plates and sufficient particle exhaust is the basic requirement of the design. The SOLPS5.0 suite of codes, is the main tool employed during the divertor design process [2,3].
The upgraded divertor will have deep, inclined inner and outer vertical plates with tightly fitting dome baffle in the private flux region to minimize its conductance for neutral leakage from the divertor region into the main chamber. The chosen divertor geometry is compatible with a wide range of lower triangularity from 0.3 to 0.6. This is consideration to left more flexible to new conception divertor design in the near future. Fig. 1 shows the separatrix locations for the low and high-triangularity magnetic geometries. The width of the throat, the leg length and the incident angle of target plate marked in Fig. 2 are the three key divertor geometry parameters that are considered here [4]. The width of baffle throat is fixed by the width of SOL layer at mid-plane. Incident angle of target plate is related to the leg length and neutral particle transport through the emitted direction and incident angle. The strength of the neutral exhaust depends most strongly on the location and size of the pumping opening, the pump speed. The magnetic equilibrium, which is the basis of the numerical grid generated for SOLPS5.0, is benchmarked by EFIT, SWEQU and the free boundary equilibrium code TSC [5]. The shape of the baffle follows the magnetic flux surface which is about 3–5 cm away from the separatrix mapped to the outer mid-plane. This distance is anticipated to be a few times of radial power decay length, kp . The targets and baffles will consider CFC as the main material in the initial phase.
⇑ Corresponding author. Tel.: +86 28 82850339; fax: +86 28 82850300. E-mail address: [email protected] (Y.D. Pan). 0022-3115/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jnucmat.2010.11.004
3. Edge plasma simulation The new design has the capability to accommodate lower single null (LSN) and double null (DN) configurations, but LSN will be the preferred option. The x-point location is about 90 cm below
Y.D. Pan et al. / Journal of Nuclear Materials 415 (2011) S952–S956
Fig. 1. Divertor geometry with different lower triangularity 0.46 and 0.60.
S953
the mid-plane. The divertor design is optimized for a triangularity of 0.46. Half of the referenced computational mesh is shown in Fig. 2. The computational domain covers the whole scrape-off layer and lower divertor. A small region of the plasma core periphery and the private flux region are also included. The whole computational domain is resolved into 96 poloidal divisions and 24 radial divisions. Deuterium will be used as the plasma species. For the standard operating case, it is assumed that all of the non-radiated power losses from the main plasma are transported uniformly across the LCFS. A simply estimation suggests that the power entering the SOL is about half of the total auxiliary and ohmic heating power, 5 MW–10 MW. Classical values are assumed for the parallel i e transport coefficient, gijj ; kjj and kjj , and also for the equipartition coefficient k. The anomalous perpendicular transport model used in the present study is constant in space, with the thermal diffusivities xe? ¼ xi? ¼ 1:0m2 =s and the particle diffusivity D = 0.4 m2/s. The wall pumping effect is neglected and the recycling coefficient R is set to 1.0 for the divertor plasma facing components and the vacuum vessel wall. The recycling effect at pumping port is decided by the pumping speed. In this initial study, two possible pumping ports are considered, see Fig. 2, where DIV-1 is in the poloidal port and DIV-2 is in the bottom of vessel. The cryo-pumps are considered as a necessary part of general pumping system, for density control and vessel condition. The pumping speed of the pump is used to set the recycling coefficient at the pump opening surface in the simulation. EIRENE [6,7] is the neutral particle transport code within SOLPS, and the manual states that
Fig. 2. Key parameters of HL-2M SN divertor geometry (unit: mm).
S954
Y.D. Pan et al. / Journal of Nuclear Materials 415 (2011) S952–S956
pffiffiffiffiffiffiffiffiffiffi L ¼ A ð1 RECYCTÞ 3:638 T=m
ð1Þ Exhaust Atomic Flux (s-1)
where A is the surface area (cm2) of pumping port. RECYCT is the recycling coefficient at this pumping port. L (liters/s) is the pumping speed for particles with temperature T(K) and mass m (AMU). If the height of pump opening is 25 cm, then for room temperature, D atom, a pumping speed of 125 m3/s gives RECYCT = 0.9 at the pumping opening.
5.0E+20 4.0E+20 3.0E+20
DIV-2 with Wider Gap DIV-2 with Narrow Gap DIV-1 with Wider Gap
2.0E+20 1.0E+20 0.0E+00
1.0E+19
2.0E+19
3.0E+19
4.0E+19
Separatrix ne (m-3)
4. Results of the B2-EIRENE prediction
Fig. 4. Exhaust atom flux vs. separatrix density at mid-plane when Psol = 5 MW in DIV-1 (poloidal port) and DIV-2 bottom port with different gap distance.
4.1. Power density profile The power surface density profiles at the inner/outer targets, as calculated by SOLPS, are shown in Fig. 3 for Psol = 5,7.5 and 10 MW. Here, the edge ne which 5 cm inside the separatrix surface at midplane is fixed at 2 1019 m3. The power densities at outer or inner target are shown in the opposite side of the Y-axis in order to see it clearly. DIV-1 and DIV-2 indicates two different pumping opening which defined in the Section 2. The power density is strongly peaked near the strike point. The full width at the half height of the profile is some extend of wider with the increase of Psol, but the width is less than 3 cm in each case. When Psol is 5 MW, the peak power density at outer/inner strike point is 2.2/ 5.9 MW/m2 for DIV-1 and 4.3/4.4 MW/m2 for DIV-2. It shows the strong asymmetry should be considered if we use DIV-1 as the pump opening. If we chose DIV-2, the peaked power density at outer strike points are 4.26/7.35/10.12 MW/m2 for the Psol is 5/7.5/ 10 MW. The corresponding peaked power densities at inner strike points are 4.42/7.73/11.0 MW/m2. The in–out asymmetry is much lower when the pump opening is in the bottom of VV. 4.2. Particle exhaust optimization The particle exhaust throughout is decided by many factors, including the position of the pump opening, the pumping speed and the slot of the gap. Fig. 4 shows the rate of particle exhaust at the pump openings. Here, the separatrix electron density at mid-plane is scanned and the power entering the SOL is fixed at 5 MW. The pump surface recycling coefficient is 0.9. With the increase of upstream density,
the exhaust atomic flux in DIV-2 is larger than that in DIV-1. DIV-2 has obviously advantage due to its proximity to the particle source. The total neutral particle flux into the core plasma, from the backflow of divertor recycling, is minimized by optimizing the width of the baffle throat. In addition, DIII-D experiments have shown that significant increases in plenum pressure can be obtained if careful attention is paid to the position of the separatrix with respect to the baffle [8]. The slot width, ‘‘gap’’ here, is defined as the distance between the dome and vertical target. And it is influences the back flow of neutrals. For testing the effect of these parameters, the upstream power and edge electron density were kept fixed at 5 MW and edge ne = 4 1019 m3. Here, The edge ne means the magnetic surface which 5 cm inside the separatrix surface at mid-plane. In this condition, the edge plasma will run far from the lower recycling region. The relation between the exhaust atom flux and the slot width is shown in Fig. 5. As long as the slot width is larger than 5.5 cm, the dependence of the exhaust rate on slot width is weak and the exhausted atom flux is about constant 2.0–2.5 1020 m3. Once the slot width is less than this, the exhaust atom flux peaks and then drops as the width shrinks. The optimal width is 2.0 cm with the pumping efficiency increasing by a factor of 3–4 as the gap narrows from 5 to 2.0 cm. The reason for this enhancement cannot be explained simply. Detailed neutral Monte Carlo modeling EIRENE shows narrow slot will induce enhanced pressure in the baffle and the neutral particle loss. The peak heat deposition on the target was largely independent of the slot width, as was the upstream electron density. 4.3. Divertor operation region To reduce the power load and erosion of the divertor target plates is the main objective in the design of the HL-2M divertor.
Div-1_innerplate 5MW Div-2-innerplate 5MW Div-1_outerplate 5MW Div-2-outerplate 5MW Div-2 innerplate 7.5MW dIV-2 outerplate 7.5MW Div-2 innerplate 10MW Div-2 outerplate 10MW
1.E+07
5.E+06
1.E+21
0.E+00 -0.05
0.00
0.05
0.10
0.15
0.20
0.25
-5.E+06
-1.E+07
-2.E+07
Exhaust Atom Flux
Heat flux at target plates (w/m2)
2.E+07
8.E+20 6.E+20 4.E+20 2.E+20 0.E+00
Position along the target plate(m)
0
20
40
60
80
100
120
Gap Distance (mm) Fig. 3. Power density profiles at target plates with pumping port at DIV-1 (poloidal port) or DIV-2 (bottom port) separately. The total power which enters into SOL is 5/ 7.5/10 MW. The edge ne which 5 cm inside the separatrix surface at mid-plane is 2 1019 m3.
Fig. 5. Exhaust atom flux as a function of dome-to-target distance when Psol = 5 MW, edge density ne = 4 1019 m3 is 5 cm inside the separatrix surface at outer mid-plane.
Y.D. Pan et al. / Journal of Nuclear Materials 415 (2011) S952–S956
1.0
fm
0.6
0.69
1
0.89 0.98
9.50E-01
fm_inner target fm_outer target 0.1 1
10
100
Tt (eV) Fig. 6. The pressure factor vs. Te at the strike point of the inner/outer target plates. When fm is 0.95,1.0 and 0.89. The Te/nesepm at the strike point of outer target plate and ne at outer mid-plane separatrix surface are 49 eV/2.17e19 m3, 16 eV/ 3.6e19 m3 and 4.7 eV/4.57 m3.
It is important to operate in the high recycling or detachment regimes to decrease the heat flux flowing to the target by making the electron temperature low in the target plate. Divertor operation regimes can be very sensitive to the mid-plane separatrix electron density. According to the Greenwald density limitation, HL2M will be able to run safely with the line average densities up to 1.5 1020 m3 in ohmic discharges. As a measure to quantify for R detachment, the pressure fraction
at mid-plane and target, fm ¼ Rtargetplate
midplane
pð1þcM 2 Þdw
S955
plate ‘Tesepa’ and the electron density at outer mid-plane separatrix surface ‘nesepm’ are changed from 49 eV/2.17e19 m3 to 16 eV/3.6e19 m3 and then 4.7 eV/4.57e19 m3. We think fm = 1 is the division of high recycle and detachment. Here, ne at outer mid-plane separatrix surface is 3.6e19 m3. High density and low temperature (10 eV < Tesepa < 50 eV) are shown at target plates in the high-recycling regime. When nesepm increases larger and larger, the detachment becomes more and more. When nesepm is quite lower, only the low recycling regime can be attained. In this regime, due to the high parallel heat conduction, the electron temperature shows little drop along the field lines. In this region, recycling losses can be neglected, so the peak heat flux at the target is related to the power enters the SOL layer. In contrast to the low recycling regime, flow reversal of the D ions, which generally characterizes the high-recycling regime, is observed in the region close to the strike point. Fig. 7 is the mach number profiles when nesepm is 3.37e19 m3 (left) and 5.51e19 m3 (right). When edge plasma runs from high recycling phase (left figure in Fig. 7) to detachment (right figure in Fig. 7), the D ions flow reversal disappears. The negative mach number region near the strike point of target plate completely becomes the positive direction. The backflow is the important source of impurities.
, is an important va-
pð1þcM 2 Þdw
lue. It is calculated from the flux integrated dw pressure profiles
5. Summary
at the target plate and mid-plane. c is adiabatic coefficient and M is mach number. The effect of the acceleration at the sheath entrance appears as the factor cM2. Significant Te and ne gradients along field lines are shown with the increasing of the fm. The pressure factor fm is approach to 1. It means the momentum almost has been lost completely along the mid-plane to target. The relation of fm to the electron temperature Te of the strike points at the target plates is shown in Fig. 6. The pressure factor is decided by different competitive mechanism. The pressure profile at the mid-plane is related to the anomalous radial transport and the pressure profile near the target plate has strong relation with the ionization and CX neutral losses in the divertor region. It is difficulty to divide the divertor operation regions simply. When fm at outer target changes from 0.95 to 1.0 and then decreases to 0.89, The temperature at the strike point of outer target
In this design study, the strictly power deposition profiles on the both target plates are shown as a function of the power entering the SOL, with a focus on the peak power loads and the possible neutral back flow when divertor runs from high recycling phases to detachment. Target detachment was observed in the SOLPS calculations when the separatrix plasma density was high and it has strong effect on the symmetry and stability of the divertor plasma. During the transition, the momentum loss is another key parameter which determines the possible reverse of edge flow. It has strong effect on the symmetry and stability of the divertor plasma and finally on the whole edge region. With an anticipated SOL power flux of 5–10 MW discharge, the peak heat load on the divertor target is evaluated to be 4–10 MW/m2. To reduce neutral leakage from the divertor and to increase the PFR neutral pressure, which enhances divertor pumping, a PFR baffle structure (dome) is introduced in the design. The gaps
Fig. 7. Mach number profiles when nesepm = 3.37e19 m3 (left) and 5.51e19 m3 (right).
S956
Y.D. Pan et al. / Journal of Nuclear Materials 415 (2011) S952–S956
between the vertical targets and dome also are optimized, so that they maximize the neutral flux that the divertor pump. The pumping efficiency is found to increase by a factor of 3–4 when the divertor gap is narrowed from 5.5 to 2.0 cm. Acknowledgements Dr. R. Schneider, D. Coster, S.Z. Zhu and Y.P. Chen have provided many valuable contributions to this work. This work is supported by the National Magnetic Confinement Fusion Science Program, No. 2009GB104008 and National Natural Science Foundations of China and Grand No. 10805016.
References [1] Y.D. Pan, R. Schneider, J. Nucl. Mater. 407 (2007) 363–365. [2] B.J. Braams, A multi-fluid code for simulation of the edge plasma in tokamaks. NET Report No. 68, 1987. [3] R. Schneider, H.S. Bosch, J. Neuhauser, et al., J. Nucl. Mater. 241–243 (1997) 701–706. [4] S. Sakurai, K. Shimizu, K. Masaki, et al., PPCF 44 (2002) 49–760. [5] Y.D. Pan, S.C. Jardin, C. Kessel, The discharge design of HL-2M with the Tokamak Simulation Code (TSC), Princeton University Plasma Physic Laboratory Report PPL-4257, OCT, 2007. [6] D. Reiter, J. Nucl. Mater. 196–198 (1992) 241. [7] D. Reiter. The EIRENE Code, Version Jan. 92, User Manual, March 1992. [8] M.A. Mahdavi, S.L. Allen, D.R. Baker, B. Bastasz, N.H. Brooks, et al., J. Nucl. Mater. 220–222 (April) (1995) 13–24.