scieee Open visual document viewer

Predictive simulations for plasma scenarios in the SMART tokamak

Mancini, Alessio; Velarde Gallardo, Lina; Viezzer, Eleonora; Cruz Zabala, Diego José; Rivero Rodríguez, Juan Francisco; García Muñoz, Manuel; Sanchís Sánchez, Lucía; Snicker, Antti; García Domínguez, J.; Segado Fernández, Jorge; Hidalgo Salaverri, Javier

Abstract

The SMall Aspect Ratio Tokamak (SMART) is a new spherical machine that is currently being constructed at the University of Seville (Mancini et al., 2021; Agredano-Torres et al., 2021). The operation of SMART will cover three different phases reaching an inductive plasma current ( Iₚ )of more than 500 kA, a toroidal magnetic field ( Bₜ ) of 1 T and a pulse length of 500 ms (Mancini et al., 2021; Agredano-Torres et al., 2021). The main goal of the SMART tokamak is to study high plasma confinement regimes in a broad triangularity range (-0.5 ≤ δ ≤ 0.5) (Doyle et al., 2021; Doyle et al., 2021). While in phase 1 the ohmic heating alone is expected to provide enough power to access the H-mode, in phase 2 and phase 3 the access to the H-mode will be ensured by applying Neutral Beam Injection (NBI) as external heating system. The NBI will consist of one injector at 25 keV and 1 MW of power. The overall design of the NBI, including injection geometry, energy and power have been optimized using the ASCOT5 code (Hirvijoki et al., 2021). The SMART scenarios have been developed with the help of the free boundary equilibrium solver code FIESTA (Cunningham, 2013) coupled to the linear time independent, rigid plasma model RZIP (Lazarus et al., 1990) to calculate the target equilibria for all the different operational phases. To assess the feasibility of those scenarios, predictive modelling needs to be included to evaluate properly the evolution of the temperatures, density profiles for both electrons and ions. To this extent, the 1.5D transport code ASTRA (Pereverzev and Yushmanov, 2002) has been used including models for the ohmic current, bootstrap current and current driven by NBI. This contribution discusses the electron and ion density and temperature profiles obtained for various scenarios for phase 1 and 2 and presents the design study of the NBI.

Full text

Fusion Enginee ing and Design 192 (2023) 113833 A ailable online 28 May 2023 0920-3796/© 2023 The Au ho (s). Published by Else ie B.V. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by- nc-nd/4.0/). Con en s lis s a ailable a ScienceDi ec Fusion Enginee ing and Design jou nal homepage: www.else ie .com/loca e/ usengdes P edic i e simula ions o plasma scena ios in he SMART okamak A. Mancini a,b,∗, L. Vela de b,c, E. Viezze a, D.J. C uz-Zabala a, J.F. Ri e o-Rod iguez d, M. Ga cia-Muñoz a, L. Sanchis e, A. Snicke e, J. Segado-Fe nandez b, J. Ga cia-Dominguez b, J. Hidalgo-Sala e i b,c, P. Cano-Megias b,c, M. Toscano-Jimenez , he PSFT G oup aDepa men o A omic, Molecula and Nuclea Physics, Uni e si y o Se ille, Se ille, Spain bCen o Nacional de Acele ado es (U. Se illa, CSIC, J. de Andalucia), Se illa, Spain cDepa men o Ene gy Enginee ing, Uni e si y o Se ille, 41092 Se ille, Spain dUni ed Kingdom A omic Ene gy Au ho i y, Culham Cen e o Fusion Ene gy, Culham Science Cen e, Abingdon, Oxon, OX14 3DB, UK eDepa men o Applied Physics, Aal o Uni e si y, Espoo, Finland Depa men o Applied Physics III, Uni e si y o Se ille, Spain ARTICLE INFO Keywo ds: SMART ASTRA ASCOT5 FIESTA Plasma scena ios ABSTRACT The SMall Aspec Ra io Tokamak (SMART) is a new sphe ical machine ha is cu en ly being cons uc ed a he Uni e si y o Se ille (Mancini e al., 2021; Ag edano-To es e al., 2021). The ope a ion o SMART will co e h ee di e en phases eaching an induc i e plasma cu en (𝐼𝑃) o mo e han 500 kA, a o oidal magne ic ield (𝐵𝑇) o 1 T and a pulse leng h o 500 ms (Mancini e al., 2021; Ag edano-To es e al., 2021). The main goal o he SMART okamak is o s udy high plasma con inemen egimes in a b oad iangula i y ange (-0.5≤𝛿≤0.5) (Doyle e al., 2021; Doyle e al., 2021). While in phase 1 he ohmic hea ing alone is expec ed o p o ide enough powe o access he H-mode, in phase 2 and phase 3 he access o he H-mode will be ensu ed by applying Neu al Beam Injec ion (NBI) as ex e nal hea ing sys em. The NBI will consis o one injec o a 25 keV and 1 MW o powe . The o e all design o he NBI, including injec ion geome y, ene gy and powe ha e been op imized using he ASCOT5 code (Hi ijoki e al., 2021). The SMART scena ios ha e been de eloped wi h he help o he ee bounda y equilib ium sol e code FIESTA (Cunningham, 2013) coupled o he linea ime independen , igid plasma model RZIP (Laza us e al., 1990) o calcula e he a ge equilib ia o all he di e en ope a ional phases. To assess he easibili y o hose scena ios, p edic i e modelling needs o be included o e alua e p ope ly he e olu ion o he empe a u es, densi y p o iles o bo h elec ons and ions. To his ex en , he 1.5D anspo code ASTRA (Pe e e ze and Yushmano , 2002) has been used including models o he ohmic cu en , boo s ap cu en and cu en d i en by NBI. This con ibu ion discusses he elec on and ion densi y and empe a u e p o iles ob ained o a ious scena ios o phase 1 and 2 and p esen s he design s udy o he NBI. 1. In oduc ion The SMall Aspec Ra io Tokamak (SMART) is a new sphe ical oka- mak cu en ly being cons uc ed and assembled a he Uni e si y o Se ille [1,2]. SMART will ope a e h ough h ee phases ha di e in alue o he plasma cu en 𝐼𝑝up o mo e han 500 kA, o oidal magne ic ield 𝐵𝑇up o 1 T and pulse leng h 𝜏up o 500 ms, see Table 1. I s e sa ili y is he ope a ion in single and double null con igu a ion, and wi h posi i e and nega i e iangula i y (−0.5≤𝛿≤ 0.5[3,4]). By phase 2, SMART will also be equipped wi h a Neu al Beam Injec o (NBI) wi h a maximum powe o 1 MW, o s udy he e ec s o as -ion physics in high posi i e and nega i e iangula i y scena ios. In his en i onmen , he p edic ion o plasma pe o mance ∗Co esponding au ho a : Depa men o A omic, Molecula and Nuclea Physics, Uni e si y o Se ille, Se ille, Spain. E-mail add ess: [email p o ec ed] (A. Mancini). o he machine is essen ial and can be achie ed by using p edic i e modelling codes in o de o compu e he e olu ion o plasma p o iles, such as empe a u e and densi y, o he h ee di e en phases. The aim o he pape is he desc ip ion o he SMART pe o mances in he i s wo phases and i is o ganized as ollows. Sec ion 2will p o ide a gene al desc ip ion o he SMART okamak, while he equilib ium ob ained wi h FIESTA [3–5] and he anspo model used in ASTRA [6] will be de ailed in Sec ion 3. Phase 1 in i s baseline and posi i e iangula i y scena ios will be add essed in Sec ion 4, while Phase 2 will be discussed in Sec ion 5wi h densi y and empe a u e p o iles in 5.1 and he desc ip ion o he op imiza ion p ocess o he NBI in 5.2. This wo k ocuses solely in phase 1 and phase 2 in posi i e iangula i y, h ps://doi.o g/10.1016/j. usengdes.2023.113833 Recei ed 28 Oc obe 2022; Recei ed in e ised o m 31 Ma ch 2023; Accep ed 17 May 2023 Fusion Enginee ing and Design 192 (2023) 113833 2 A. Mancini e al. Fig. 1. O e iew o he SMART okamak oge he wi h he Neu al Beam Injec o (NBI). Fig. 2. Phase 1 equilib ium om FIESTA. Baseline is depic ed on he le while he posi i e iangula i y case is on he igh . while nega i e iangula i y will be assessed in a sepa a e wo k. The posi i e iangula i y scena ios ha e been compu ed in o de o assess he pe o mance o he machine in e ms o o oidal be a 𝛽𝑡, poloidal be a 𝛽𝑝, no malized be a 𝛽𝑁=𝛽𝑡∕(𝐼𝑝∕(𝑎𝐵𝑇)) and con inemen ime 𝜏𝐸. In addi ion, he e alua ion o hose scena ios will cons i u e a solid base o he design o many diagnos ics, and o MHD and gy okine ics analysis, and hei alidi y will be p o en once he machine will come in o ope a ion. 2. SMART okamak The SMall Aspec Ra io Tokamak (SMART) is a compac sphe ical machine cha ac e ized by an o e all heigh o 3 m and an o e all diame e o 2 m. I is composed by an AISI 316 L s ainless s eel acuum essel ha ing an inne wall diame e o 300 mm, ou e wall diame e o 1600 mm and an inne heigh o 1600 mm. The acuum essel has an o e all numbe o 44 ci cula po s and 2 ec angula po s mainly aimed o diagnos ics and main enance pu poses. A se o wel e coppe o oidal ield coils (4 numbe o u ns each) a e capable o p oducing he desi ed o oidal magne ic ield (𝐵𝑇) a he majo adius (𝑅0) o he plasma o all he ope a i e phases (see Table 1). Fou poloidal ield coils (PF1 and PF2), wi h 23 numbe o u ns each, a e needed o he e ical con ol and shaping o he plasma. Table 1 Ope a i e s ages o he SMART okamak. Phase 1 Phase 2 Phase 3 𝑅0[m] 0.4 0.4 0.4 𝑎[m] 0.25 0.25 0.25 𝜅≤1.95 ≤2.00 ≤2.30 𝛿±0.4 ±0.5 ±0.5 𝐼𝑃[kA] 100 200 > 500 𝐵𝑇[T] 0.1 0.4 1 𝜏[ms] 0.100 0.150 0.500 𝑃𝐸𝐶𝑅𝐻 [kW] 6 (2.45 GHz) 6 (7.5 GHz) 200 𝑃𝑁𝐵𝐼 [kW] – 1000 1000 Table 2 ASTRA inpu model da a. Phase 1 Phase 2 base @𝛿𝑚𝑎𝑥 base @𝛿𝑚𝑎𝑥 𝛥𝑠[m] 0.059 0.067 0.035 0.032 𝐼𝑃[kA] 30 30 200 200 𝐵𝑇[T] 0.1 0.1 0.4 0.4 𝜅1.83 1.48 1.95 1.79 𝛿0.23 0.4 0.35 0.4 𝑆𝑝[m2] 5.081 5.12 6.80 6.31 𝑉𝑝[m3] 0.68 0.70 1.03 0.90 𝛽𝑡1.74 2.46 2.92 3.3 𝛽𝑝0.337 0.395 0.325 0.322 𝛽𝑛1.25 1.77 1.46 1.61 The PF1 coils a e placed ou side he essel while he PF2 a e placed inside (see Fig. 1). Two pai s o di e o ield coils (DIV1 and DIV2 in Fig. 1), all placed inside he machine, a e needed o each he desi ed elonga ion 𝜅and o ope a e he machine in single o double null con ig- u a ion and in posi i e and nega i e iangula i y [3,4]. The numbe o u ns o hese coils di e s, being 35 o DIV1 and 23 o DIV2. A coppe solenoid wi h 230 numbe o u ns comple es he magne ic sys em o he machine. SMART will be equipped wi h ou acuum pumps: wo d y pumps o 80 m3h−1 and wo u bomolecula pumps o 2300 l s−1 designed o ob ain he equi ed le el o acuum o 10−8 o wi h a maximum leakage o 10−8mba ⋅l s−1. Addi ional hea ing sys ems such as Elec on Cyclo on Resonance Hea ing (ECRH) and Neu al Beam Injec o (NBI) will also be ins alled du ing he ope a ional phases o he machine. The op imiza ion p ocess and he pa ame e s o he NBI a e discussed in Sec ion 5.2. Addi ional in o ma ion a e de ailed in [1,2]. 3. Plasma scena ios modelling The de elopmen o he ope a i e scena ios assume an impo an ole in he design o okamak de ices, as hey in luence he ope a ion o he machine, he assessmen o he pe o mance bu also he design o se e al diagnos ic sys ems which need an es ima ion o he p o iles o elec on and ion densi y (𝑛𝑒,𝑛𝑖) oge he wi h hei empe a u e p o iles (𝑇𝑒,𝑇𝑖). In he SMART design, he modelling o he plasma scena ios ha e been compu ed using he ASTRA code, a 1.5 D anspo code which combines 2D equilib ium equa ions wi h a se o 1D anspo equa ions [6]. The inpu pa ame e s o he modelling a e he majo adius 𝑅0, mino adius 𝑎, elonga ion 𝜅, iangula i y 𝛿, Sha ano shi 𝛥𝑠, plasma olume 𝑉𝑝and plasma su ace 𝑆𝑝, which ha e been com- pu ed wi h he FIESTA code [5,7]. The equilib ium p o iles o phase 1 and phase 2 a e shown in Figs. 2 and 3wi h he main equilib ium pa ame e s summa ized in Table 2. The di e en be a (𝛽𝑁, 𝛽𝑡and 𝛽𝑝) included in Table 2 ha e been compu ed wi h FIESTA and e i ied wi h ASTRA. The anspo simula ions we e ca ied ou wi h ASTRA by consid- e ing a hyd ogen plasma wi h 𝑍𝑒𝑓𝑓 equal o 2, a G eenwald ac ion (𝑓𝐺𝑊 ) o 0.4 and a Gy oBohm model [8] o he anspo coe icien s: 𝜒𝑖=𝐶1 𝑎2𝑇3∕2 𝑒 𝐵2 𝑡𝑜𝑟𝜇2 ∇𝑇𝑒 𝑇𝑒 (1) Fusion Enginee ing and Design 192 (2023) 113833 3 A. Mancini e al. Fig. 3. Phase 2 equilib ium om FIESTA. Baseline is depic ed on he le while he posi i e iangula i y case is on he igh . 𝜒𝑒= 2𝜒𝑖(2) 𝐷=𝐶2𝜒𝑒(3) 𝑣=𝐶3𝑣𝑝𝑉′−𝐷 𝑅0(0.2𝑅0 𝐿𝑇 𝑒 + 0.1𝑠𝐶4)(4) being 𝜒𝑖 he ion hea anspo coe icien , 𝜒𝑒 he elec on hea anspo coe icien (assuming pu e Ion Tempe a u e G adien , ITG), 𝐷 he di usion coe icien , 𝑣 he con ec i e eloci y, 𝑣𝑝=𝐸||∕𝐵𝑝 he pinch eloci y indica ing wi h 𝐸|| he pa allel componen o he elec ic ield and 𝐵𝑝 he poloidal componen o he magne ic ield. Finally 𝐿𝑇 𝑒 =𝑇𝑒∕∇𝑇𝑒is he elec on empe a u e g adien leng h scale and 𝑠 he shea . The coe icien s 𝐶1,𝐶2,𝐶3and 𝐶4ha e been chosen equal o [0.4, 0.2, 0.1, 1] o ha e compa able anspo coe icien s o machines wi h simila size/aspec a io o SMART (i.e. GLOBUS-M [9] and Pegasus [10]) and o ake in o accoun also neoclassical e ec s. Fo he simula ions in H-mode, 𝜒𝑖has been chosen equal o 4 m2s−1 wi h he edge anspo ba ie (ETB) placed a 𝜌𝑝𝑜𝑙 = 0.95 [9–11] and lea ing he same app oxima ion o he Gy oBohm model o he o he coe icien s (see Sec ion 5.1). 4. Phase 1 scena ios In his sec ion he phase 1 scena ios will be discussed bo h in he baseline and posi i e iangula i y case. Simula ions o phase 1 equilib ia ha e been un wi h he pa ame e s included in Table 2. Figs. 4 and 5show he densi y and empe a u e p o iles as a unc ion o he no malized adius 𝜌𝑝𝑜𝑙 =√(𝜓−𝜓𝑏)∕(𝜓𝑏−𝜓0), being 𝜓𝑏and 𝜓0 he alue o he poloidal lux a he sepa a ix and in he cen e espec i ely. The densi y inc eases wi h he iangula i y bo h o ions and elec- ons. The co e elec on densi y 𝑛𝑒0inc eases om 0.865 × 1019 m−3 o 1.01 × 1019 m−3, while he co e ion densi y 𝑛𝑖0inc eases om o 0.7 × 1019 m−3 o 0.8 × 1019 m−3. Tempe a u e p o iles show he same end, wi h a maximum o 0.171 keV and 0.08 keV espec i ely o 𝑇𝑒 and 𝑇𝑖in he co e. The bene i s o inc easing he iangula i y is also mani es ed in he educ ion o he anspo coe icien s 𝜒𝑖,𝜒𝑒and 𝐷 as shown in Fig. 6. In phase 1, no addi ional sou ce o ene gy is used o hea he plasma apa om he in insically ohmic hea ing. The ohmic powe , 𝑃𝑂𝐻 , can be es ima ed as [12]: 𝑃𝑂𝐻 [MW] = 7 × 10−2 𝑍𝑒𝑓𝑓 𝐵2 𝑇𝑉𝑝 𝑞2 𝑎𝑅2 0𝑇3∕2 𝑒 (5) Fig. 4. 𝑛𝑒and 𝑛𝑖p o iles o phase 1 in he baseline scena io and in he posi i e iangula i y case. Fig. 5. 𝑇𝑒and 𝑇𝑖p o iles o phase 1 in he baseline scena io and in he posi i e iangula i y case. Fig. 6. 𝜒𝑖,𝜒𝑒and 𝐷p o iles o phase 1 in he baseline scena io and in he posi i e iangula i y case. whe e 𝑞𝑎= 5𝜅𝑎2𝐵𝑇∕𝑅𝐼𝑝is he sa e y ac o a 𝑟=𝑎and 𝑇𝑒 he a e age elec on empe a u e in keV. Using his simple 0-D app oxima ion, wi h he pa ame e s included in Table 2, he ohmic powe 𝑃𝑂𝐻 is 47 kW, simila o 50 kW es ima ed wi h ASTRA. The h eshold powe o he ansi ion om L- o he H-mode (𝑃𝐿𝐻 ) can be e alua ed wi h he ollowing exp ession [13,14]: 𝑃𝐿𝐻 [MW] = 0.0488𝑛0.717 20 𝐵0.803 𝑇𝑆0.941 (6) whe e 𝑆= 4𝜋2𝑎𝑅√0.5(1 + 𝑘2)is he plasma su ace a ea and 𝑛20 he elec on a e age densi y exp essed in 1020 uni . Fo phase 1, he Fusion Enginee ing and Design 192 (2023) 113833 4 A. Mancini e al. Fig. 7. Densi y p o iles o phase 2 wi h and wi hou NBI. h eshold powe is app oxima ely 9 kW. The adia i e powe 𝑃𝑟𝑎𝑑 om he plasma, app oxima ed by he sum o he B emss ahlung 𝑃𝐵𝑟 and cyclo on 𝑃𝑐𝑦𝑐𝑙 losses, can be neglec ed as app oxima ely equal o 60 W. 𝑃𝐵𝑟 and 𝑃𝑐𝑦𝑐𝑙 ha e been es ima ed wi h he ollowingexp essions [15]: 𝑃𝐵𝑟[W] = 𝑍2 𝑒𝑓𝑓 𝑛𝑖𝑛𝑒𝑇0.5 𝑒 [7.69 × 1018]2𝑉𝑝(7) 𝑃𝑐𝑦𝑐𝑙[W] = 6.21 × 10−22𝐵2 𝑇𝑛𝑒𝑇𝑒𝑉𝑝(8) which gi es esul s in acco dance wi h ASTRA simula ions. The e o e, neglec ing he adia i e powe 𝑃𝑟𝑎𝑑 , i is expec ed ha in phase 1 SMART will achie e he ohmic H-mode (𝑃𝑂𝐻 ≫ 𝑃𝐿𝐻 ). The ene gy con inemen ime has been app oxima ed wi h he ollowing scaling law alid a low densi ies and o ohmic hea ing [16]: 𝜏𝐸[s] = 0.07𝑛20𝑎𝑅2 0𝑞𝑐𝑦𝑙 (9) whe e 𝑞𝑐𝑦𝑙 = 5𝑎2𝐵𝑇∕𝐼𝑝𝑅 he cylind ical sa e y ac o . Fo phase 1, an ene gy con inemen ime o 0.33 ms is expec ed while ASTRA p edic s 0.22 ms, sligh ly lowe bu o he same o de o magni ude. The boo s ap cu en ac ion 𝑓𝑏𝑠 is expec ed o inc ease wi h he iangula i y 𝛿 om 14.3% o 15.1% as he e is a small inc ease in 𝛽𝑝, see Table 2. 5. Phase 2 scena ios 5.1. ASTRA esul s ASTRA simula ions ha e been ca ied ou in o de o de e mine he p o iles o densi y and empe a u e in phase 2 wi h and wi hou he NBI. The pa ame e s o he NBI used o he ASTRA simula ions a e he injec ed powe (𝑃𝑁𝐵𝐼 ) o 1 MW and ene gy (𝐸0) o 25 keV as desc ibed in Sec ion 5.2. Only he NBI hea ing has been included in he simula ions. No ECRH hea ing has been conside ed as i will only be used o p e-ioniza ion du ing he plasma s a -up. The baseline case has been compa ed wi h he ex eme posi i e iangula i y, e en hough he wo con igu a ions di e sligh ly (see Table 2). Figs. 7 and 8show he densi y and empe a u e p o iles in phase 2 whe e he blue cu es ep esen s he scena ios wi hou NBI, and he ed ones wi h he NBI. Do ed cu es in bo h cases ep esen he scena ios in he highes posi i e iangula i y case. Wi h he NBI a ise in bo h densi y and empe a u e is expec ed. 𝑛𝑒0and 𝑛𝑖0will inc ease up o 4.7× 1019 m−3 and 3.8 × 1019 m−3. No big di e ence is expec ed be ween he baseline and he ex eme posi i e iangula i y case, as he di e ence in iangula i y o hose wo cases is 0.1. A la ge inc ease is expec ed in he elec on and ion empe a u e. 𝑇𝑒and 𝑇𝑖will inc ease up o 0.56 keV and 0.7 keV espec i ely in Fig. 8. Tempe a u e p o iles o phase 2 wi h and wi hou NBI. Fig. 9. T anspo coe icien s wi h he ETB loca ed a 𝜌𝑝𝑜𝑙 = 0.95. he baseline case. In he case o ex eme posi i e iangula i y, he ion empe a u e will be sligh ly highe up o 0.75 keV and he elec on empe a u e up o 0.6 keV. Scena ios wi h NBI will enhance he 𝛽 pa ame e s. In he baseline case, he 𝛽𝑡will ise up o 8.23%, 𝛽𝑝up o 0.7 and 𝛽𝑁up o 4.1, compa ed o he alues eached in phase 1 (see Table 2). In he ex eme posi i e case 𝛽𝑡will ise up o 9.1%, 𝛽𝑝o 0.69 and 𝛽𝑁up o 4.44. Mo eo e , ASTRA esul s o esee a boo s ap cu en ac ion ha inc ease om 𝑓𝑏𝑠 ≈ 15% in phase 1 o 𝑓𝑏𝑠 ≈ 37% in phase 2 (𝐼𝑏𝑠 ≈66 kA), wi h a cu en ac ion induced om he NBI (𝑓𝐶𝐷) o almos 15% which co esponds o an induced cu en 𝐼𝐶𝐷 ≈ 30 kA, in line wi h he esul s o GLOBUS-M [17]. ASTRA simula ions p edic ha he o al beam powe 𝑃𝐶𝐷 is app ox- ima ely 0.7 MW high enough o o e come 𝑃𝐿𝐻 , see o mula (6), which is app oxima ely 90 kW. The adia i e powe can be neglec ed as in phase 1. In ac , using he o mulas (7) and (8), he adia i e powe can be es ima ed o almos 2.2 kW. The e o e i is expec ed ha in phase 2 wi h he NBI, SMART will achie e he H-mode as 𝑃𝐶𝐷 ≫ 𝑃𝐿𝐻 . In addi ion wi hou he NBI, he ohmic powe (𝑃𝑂𝐻 ), e alua ed wi h he o mula (5), is app oxima ely 0.22 MW simila o he 0.27 MW ob ained wi h ASTRA. As he ohmic powe is highe han he h eshold powe i is likely o expec ha SMART will achie e in a H-mode egime also wi hou he NBI. This jus i ies he app oxima ion o using he anspo coe icien s wi h he p o iles shown in Fig. 9 and desc ibed in Sec ion 3. Fusion Enginee ing and Design 192 (2023) 113833 5 A. Mancini e al. As he ope a ion will be in H-mode, he con inemen ime 𝜏𝐸can be es ima ed wi h he ITER IPB(y,2) scaling law [18]: 𝜏𝐸[s] = 0.145 𝐼0.93 𝑃𝑅1.39 0𝑎0.58𝑘0.78𝑛0.41 20 𝐵0.15 𝑇𝑀0.19 𝑃0.69 𝐻 (10) whe e 𝑃𝐻=𝑃𝑁𝐵𝐼 is he ex e nal hea ing powe . Fo phase 2 we ge a con inemen ime o 6.5 ms sligh ly lowe bu o he same o de o magni ude o he one compu ed by ASTRA, 10 ms. A be e es ima ion o he con inemen ime 𝜏𝐸is ob ained when conside ing a NSTX-Gy oBohm scaling law alid o sphe ical okamaks [19]: 𝜏𝐸[s] = 0.21 𝐼0.054 𝑝𝐵0.91 𝑇𝑅2.14 0 𝑛0.05 𝑒𝑃0.38 𝐻 (11) which gi es a 𝜏𝐸o almos 9 ms, in line wi h he ASTRA esul s. 5.2. NBI design and op imiza ion To ensu e H-mode in phase 2, 1 MW o Neu al Beam Injec ion (NBI) will be applied o posi i e- iangula i y SMART plasmas. Beam heigh and wid h ha e been de e mined scaling NBIs o simila machines in size o SMART, such us Globus-M. The p elimina y dimensions o he g id beams conside ed in he simula ions a e heigh and wid h o 30 cm and 33 cm espec i ely. The main pa ame e s o he NBI ha e been op imized h ough he Mon e-Ca lo o bi - ollowing code ASCOT5 [20], using he wall geome y and he magne ic equilib ium calcula ed wi h FIESTA (see Sec ion 3) and he empe a u e and densi y p o iles ob ained wi h ASTRA wi hou ex e nal hea ing, see Sec ion 5.1. The op imiza ion o he main NBI pa ame e s has been pe o med wi h an i e a i e p ocess, analysing he con inemen o di e en pa icles injec ed a he sepa a ix (𝑟= 0.75 m and 𝑧= 0 m) and ionized inside he plasma. He e, he op imiza ion o he injec ion geome y is p esen ed, pe o med se ing he NBI main beam ene gy, 𝐸0, o 25 keV. This has been conside ed as he bes choice o phase 2 acco ding o p e ious analysis, and ma ching he ene gy obse ed in o he simila machines [9]. The ene gy o he beam is limi ed on one hand by he gy o adius o he esul ing pa icles, ha inc eases wi h ene gy (𝑟𝐿= 𝑚𝑣⟂∕𝑞𝐵 ∝𝐸0∕𝐵), and by he o bi d i s, hus a ec ing as -ion losses. On he o he hand, he highe he ene gy, he lowe he c oss-sec ion o he global ioniza ion p ocesses be ween he injec ed neu als and he he mal plasma [21], inc easing he shine- h ough (i.e., he pe cen age o pa icles ha a el h ough he plasma wi hou ge ing ionized), and hus, dec easing he o al e iciency. The e o e, he wo main pa ame e s employed o op imize he NBI geome y and ene gy a e he p omp losses (i.e., pa icles ha a e los be o e comple ing a o oidal pe iod) and he shine- h ough. Mo eo e , a 25 keV injec ion ene gy p o ides a supe -Al énic as -ion dis ibu ion, as he Al én eloci y a 𝜌𝑝𝑜𝑙 = 0.5is 𝑣𝐴∼ 1.6 × 106m∕s, while he eloci y co esponding o he main injec ion ene gy is 𝑣= 2.2 × 106m∕s, making i possible o d i e uns able a wide a ie y o Al én Eigenmodes, enabling o s udy hei beha iou as a unc ion o plasma shape and iangula i y. Highe injec ion ene gies, up o 35 keV, will also be conside ed du ing he op imiza ion p ocess. In o de o op imize he injec ion con igu a ions, he as -ion bi h dis ibu ion has been modelled wi h a se o 106ma ke s (high enough o achie e good s a is ics, acco ding o p e ious analysis), whose o bi s a e ollowed o 1 ms. The di e en se s o ma ke s ha e been gene a ed wi h he BBNBI code [22], by a ying he angency adius o he beam cen e line, as shown in Fig. 10. Then, by means o he ASCOT5 code, he ionized pa icles ha e been ollowed in a collisionless and MHD- quiescen en i onmen , in o de o calcula e he pe cen ages o p omp losses (as men ioned, pa icles los be o e comple ing a o oidal pe iod, and hus, no a ec ed by Coulomb collisions). The selec ed simula ion ime is su icien o accoun o he p omp losses. The shine- h ough has also been calcula ed o each o he 9 cases s udied, wi h he aim o minimize he o al amoun o pa icles ha a e los . Fig. 10. Ske ch o he o oidal iew o he SMART NBI con igu a ion. The black ci cles ep esen he essel walls and he ed ones he o oidal sec ion o he sepa a ix. The yellow line ep esen s he beamline. The pe pendicula dis ance om he beam o he cen e o he de ice is ep esen ed by he blue ci cle and line, indica ing he angency adius, 𝑅𝑡. Fig. 11. NBI bi h dis ibu ion in he o oidal p ojec ion o he mos ex eme con igu a ions, #1 (up) and #9 (down). The esul s o he di e en con igu a ions s udied by a ying he chosen op imiza ion pa ame e , 𝑅𝑡, a e shown in Fig. 12. As he plasma inne and ou e adius is loca ed a 0.243 mand 0.747 m,𝑅𝑡has been modi ied om 0.280 m o 0.735 m. The o oidal iew o he bi h dis ibu ion o he pa icles gene a ed wi h hese wo ex eme con igu a ions a e shown in Fig. 11. The minimum o al amoun o losses is p o ided by con igu a ions #3 and #4, see Fig. 12. The ise in he shine- h ough wi h 𝑅𝑡is explained conside ing ha , as he angency adius is inc eased, he beam a els a sho e pa h h ough he plasma, in a egion wi h lowe densi y. On he con a y, inc easing he angency adius, he beam aligns wi h he magne ic ield lines, dec easing he p omp losses o Fusion Enginee ing and Design 192 (2023) 113833 6 A. Mancini e al. Fig. 12. E olu ion o he esul ing losses o he di e en con igu a ions ( ed ci cles) s udied o he NBI sys em. Fig. 13. Slowing down dis ibu ion o he op imum con igu a ion, #4, in pi ch-ene gy. a negligible le el. Hence, con igu a ions om #5 o #9 should be disca ded, as he shine- h ough is oo high, as well as con igu a ions #1 and #2, due o hei highe esul in p omp losses in compa ison o he o he s. As con igu a ions #3 and #4 show simila esul s, he shine- h ough and he p omp losses ha e been compu ed a highe injec ion ene gies, up o 35 keV, o bo h geome ies. This c i e ia has been chosen in o de o assess hei pe o mance in he possible case o an inc ease in he ene gy o he beam in he u u e [23], which would be pe o med wi hou changing he injec ion geome y o he NBI. 35 keV is se as he maximum pe mi ed injec ion ene gy, as i allows main aining he shine- h ough below 10% wi h he cu en plasma p o iles. The esul s in Table 3 show ha , as he ene gy inc eases, he p omp losses in con igu a ion #3 ise apidly. The shine- h ough, on he con a y, inc eases o bo h con igu a ions, and i s absolu e alue is compa able. This e olu ion o he p omp losses wi h he ene gy has led o choose con igu a ion #4 as he op imum one, showing ha he in- jec ion ene gy o he NBI sys em could be inc eased in he u u e i desi ed, and he p omp losses would s ay in an admissible ange, while p ese ing an accep able e olu ion o he shine- h ough ac ion, always kep below 10%. The slowing down dis ibu ion o con igu a ion #4 and a 25 keV in- jec ion ene gy has also been analysed. Fo his simula ion, he injec ed pa icles a e allowed o in e ac wi h he bulk plasma ia Coulomb collisions o 1 s, which is ound su icien o mos o he as -ion dis ibu ion o he malize. The ma ke s a e s opped when hey each ei he he selec ed he mal limi o 1 keV, o he maximum simula ion Fig. 14. Bi h dis ibu ion o he pa icles gene a ed wi h he op imized con igu a ion (#4). To oidal (up) and poloidal (down) p ojec ions. The la e includes a ypical con ined o bi . Table 3 Resul ing losses o con igu a ions #3 and #4, and di e en injec ion ene gies. ID Ene gy Shine- h ough P omp losses To al losses 25 keV 4.9% 0.1% 5.0% 3 30 keV 6.9% 0.4% 7.3% 35 keV 9.0% 1.1% 10.1% 25 keV 5.3% 0.0 % 5.3% 4 30 keV 7.4% 0.1 % 7.5% 35 keV 9.6% 0.5 % 10.1% ime. As shown in Fig. 13, mos o he pa icles ha e a pi ch angle abo e 0.9, meaning ha hey will desc ibe passing o bi s. This is desi able in o de o educe possible losses due o he d i o apped o bi s, and esul s in a be e con inemen o he beam pa icles. Hence, he bi h dis ibu ion o he selec ed con igu a ion (#4) is shown in Fig. 14, whe e an example o a passing o bi has been plo ed in he poloidal p ojec ion o he dis ibu ion. 6. Conclusions Re e ence scena ios o he SMART okamak ha e been compu ed wi h he 1.5D anspo code ASTRA coupled wi h FIESTA o p edic he densi y and empe a u e p o iles needed p ima ily o he assessmen o he pe o mances o he machine. Only he baseline and maximum posi i e iangula i y scena ios ha e been conside ed o phase 1 and phase 2. In phase 1 he pa ame e s chosen he e sugges ha he ma- chine will access he H-mode wi h a maximum 𝑛𝑒0and 𝑛𝑖0o 1×1019 m−3 and 0.8×1019 m−3. Maximum empe a u e 𝑇𝑒0and 𝑇𝑖0o 0.171 keV and 0.080 keV will be achie ed wi h a maximum con inemen ene gy ime 𝜏𝐸o 0.2 ms, lowe han he one es ima ed wi h he con inemen scaling law o ohmic hea ing bu o he same o de o magni ude. Phase 2 will be equipped wi h a Neu al Beam Injec o whose cha ac e is ics ha e been p esen ed oge he wi h he op imiza ion o i s pa ame e s, Fusion Enginee ing and Design 192 (2023) 113833 7 A. Mancini e al. in pa icula ene gy and injec ion geome y. The bes con igu a ion has been de e mined, opening he possibili y o inc easing he NBI ene gy wi hou changing he injec ion geome y in he u u e. ASTRA simula ions o phase 2 p edic ha SMART will access H-mode wi h and wi hou NBI as bo h he Ohmic Powe (𝑃𝑂𝐻 ) and he NBI coupled powe (𝑃𝐶𝐷) a e su icien ly highe han he h eshold powe (𝑃𝐿𝐻 ) as long as he expe imen s will con i m ha he adia i e powe losses will be negligible. Co e densi ies up o 4.65 × 1019 m−3 will be eached wi h a co e empe a u e o 0.75 keV. Be a no malized 𝛽𝑁will inc ease up o 4.4 wi h a o oidal 𝛽𝑡up o 9%. A boo s ap cu en ac ion o 𝑓𝑏𝑠 o 37% will be eached and a NBI cu en d i e ac ion 𝑓𝐶𝐷 o 15% is expec ed. Nega i e iangula i y p o iles will be assessed in a sepa a e wo k, explo ing also he phase 3 scena io. Decla a ion o compe ing in e es The au ho s decla e ha hey ha e no known compe ing inan- cial in e es s o pe sonal ela ionships ha could ha e appea ed o in luence he wo k epo ed in his pape . Da a a ailabili y Da a will be made a ailable on eques Acknowledgemen s This wo k ecei ed unding om he Fondo Eu opeo de Desa ollo Regional (FEDER) by he Eu opean Commission unde g an ag eemen numbe s IE17-5670 and US-15570. The au ho s g a e ully acknowledge he inancial suppo o he Eu opean Resea ch Council (ERC) unde he Eu opean Union’s Ho izon 2020 esea ch and inno a ion p og amme (g an ag eemen No. 805162). Re e ences [1] A. Mancini, e al., Mechanical and elec omagne ic design o he acuum essel o he SMART okamak, Fusion Eng. Des. 171 (2021). [2] M. Ag edano-To es, J.L. Ga cia-Sanchez, A. Mancini, e al., Coils and powe supplies design o he SMART okamak, Fusion Eng. Des. 168 (2021). [3] S.J. Doyle, D. Lopez-Ai es, A. Mancini, e al., Magne ic equilib ium design o he SMART okamak, Fusion Eng. Des. 171 (2021). [4] S.J. Doyle, A. Mancini, e al., Single and double null equilib ia in he SMART okamak, Plasma Res. Exp ess 3 (2021). [5] G. Cunningham, High pe o mance plasma e ical posi ion con ol sys em o upg aded MAST, Fusion Eng. Des. 88 (2013) 3238–3247. [6] G.V. Pe e e ze , P.N. Yushmano , ASTRA au oma ed sys em o anspo analysis, IPP Rep. (2002). [7] E.A. Laza us, J.B. Lis e , G.H. Neilson, Con ol o he e ical ins abili y in okamaks, Nucl. Fusion 30 (1990). [8] A. Ta oni, e al., Global and local ene gy con inemen p ope ies o simple anspo coe icien s o he bohm ype, Plasma Phys. Con ol. 36 (1994) 1629. [9] I.Y. Senichenko , e al., Simula ion o L and H egimes o sphe ical okamak Globus-M wi h ASTRA anspo code, in: 35 h EPS Con e ence on Plasma Physics, 2008. [10] K.E. Thome, e al., High con inemen mode and edge localized mode cha ac e is ics in a nea -uni y aspec a io okamak, Phys. Re . Le . 116 (2016). [11] E. Viezze , e al., Ion hea anspo dynamics du ing edge localized mode cycles a ASDEX upg ade, Nucl. Fusion 58 (2018). [12] W.M. S acey, Fusion: An In oduc ion o he Physics and Technology o Magne ic Con inemen Fusion, 2010. [13] K.E. Thome, e al., H-mode plasmas a e y low aspec a io on he pegasus o oidal expe imen , Nucl. Fusion 57 (2016) 022018. [14] Y.R. Ma in, e al., Powe equi emen o accessing he H-mode in ITER, J. Phys. Con . Se . 123 (2008) 012033. [15] NRL, NRL plasma o mula y, NRL/PU/6790–18-640, 2018. [16] W.M. S acey, Fusion Plasma Physics, 2012. [17] P.B. Shchegole , e al., NBI o hea ing and cu en d i e in sphe ical okamak Globus-M,-M2, in: 43 d EPS Con e ence on Plasma Physics, 2016. [18] ITER-Physics Expe G oup on Con inemen and T anspo , e al., Chap e 2, Nucl. Fusion 39 (1999). [19] P.F. Bux on, e al., On he ene gy con inemen ime in sphe ical okamaks: Implica ions o he design o pilo plan s and usion eac o s, Plasma Phys. Con ol. Fusion 61 (2019). [20] E. Hi ijoki, e al., ASCOT: Sol ing he kine ic equa ion o mino i y pa icle species in okamak plasmas, Compu . Phys. Comm. 185 (2014) 1310–1321. [21] A.C. Ri ie e, Pene a ion o as hyd ogen a oms in o a usion eac o plasma, Nucl. Fusion 11 (1971) 363–369. [22] O. Asun a, e al., Modelling neu al beams in usion de ices: Beamle -based model o as pa icle simula ions, Compu . Phys. Comm. 188 (2015) 33–46. [23] A.Y. Telno a, e al., New 50 keV neu al beam injec o o he Globus-M2 sphe ical okamak, J. Phys. Con . Se . 1400 (2019).