NBI optimization on SMART and implications for scenario development

Plasma Physics and Controlled Fusion PAPER • OPEN ACCESS NBI optimization on SMART and implications for scenario development To cite this article: M Podestà et al 2024 Plasma Phys. Control. Fusion 66 045021 View the article online for updates and enhancements. You may also like Hydraulic Study of the Water Supply to the City of Seville through Its Aqueduct between the 17th and 19th Centuries Candela Bandrés, María Dolores Robador and Antonio Albardonedo - Single and double null equilibria in the SMART Tokamak S J Doyle, A Mancini, M Agredano-Torres et al. - Recent advances in positron emission particle tracking: a comparative review C R K Windows-Yule, M T Herald, A L Nicuan et al. - This content was downloaded from IP address 150.214.182.235 on 12/07/2024 at 12:08 Plasma Physics and Controlled Fusion Plasma Phys. Control. Fusion 66 (2024) 045021 (17pp) https://doi.org/10.1088/1361-6587/ad2edc NBI optimization on SMART and implications for scenario development M Podest` a1,5,∗, D J Cruz-Zabala2, F M Poli1,6, J Dominguez-Palacios2, J W Berkery1, M Garcia-Mu˜ noz2, E Viezzer2, A Mancini2, J Segado3, L Velarde4 and S M Kaye1 1Princeton Plasma Physics Laboratory, Princeton, NJ 08543, United States of America 2Department of Atomic, Molecular and Nuclear Physics, University of Seville, Seville, Spain 3Department of Mechanical Engineering and Manufacturing, University of Seville, Seville, Spain 4Department of Energy Engineering, University of Seville, Seville, Spain E-mail: [email protected] Received 27 September 2023, revised 26 February 2024 Accepted for publication 29 February 2024 Published 11 March 2024 Abstract The SMall Aspect Ratio Tokamak (SMART) under commissioning at the University of Seville, Spain, aims to explore confinement properties and possible advantages in confinement for compact/spherical tokamaks operating at negative vs. positive triangularity. This work explores the benefits of auxiliary heating through Neutral Beam Injection (NBI) for SMART scenarios beyond the initial Ohmic phase of operations, in support of the device’s mission. Expected values of electron and ion temperature achievable with NBI heating are first predicted for the current flat-top phase, including modeling to optimize the NBI injection geometry to maximize NBI absorption and minimize losses for a given equilibrium. Simulations are then extended for a selected case to cover the current ramp-up phase. Differences with results obtained for the flat-top phase indicate the importance of determining the plasma evolution over time, as well as self-consistently determining the edge plasma parameters for reliable time-dependent simulations. Initial simulation results indicate the advantage of auxiliary NBI heating to achieve nearly double values of pressure and stored energy compared to Ohmic discharges, thus significantly increasing the device’s performance. The scenarios developed in this work will also contribute to diagnostic development and optimization for SMART, as well as providing test cases for initial predictions of macroand micro-instabilities. Keywords: SMART tokamak, Scenario development, neutral beam injection 1. Introduction Future fusion power plants based on the tokamak concept will arguably operate within a limited range of plasma and 5Present address: Ecole Polytechnique Fédérale de Lausanne, Swiss Plasma Center, CH-1015 Lausanne, Switzerland. 6Present address: ITER Organization, 13067 St Paul Lez Durance Cedex, France. ∗Author to whom any correspondence should be addressed. Original Content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. magnetic configuration parameters. Those parameters, optimized for safe steady-state operations, need to be reliably predicted during the design phase of the reactor, since they pose engineering constraints for designing the device. Optimization of future reactors is still undergoing design efforts to assess the merits and prospects for performance (see e.g. [1] and references therein). Two considerations defining the optimal configuration are arising as key indicators for the optimization process: the choice of the so-called aspect ratio (ratio of major to minor plasma radius), and the target scenario (or configuration) for steady-state operations. The latter includes negative triangularity regimes [2–4], for which plasma shaping is reversed from what is commonly used in most experiments. 1 © 2024 The Author(s). Published by IOP Publishing Ltd Plasma Phys. Control. Fusion 66 (2024) 045021 M Podest` aet al Experiments in present tokamaks are expected to provide the basis for identifying the optimum regime(s)in which future reactors will operate. Given that present experimental devices are typically smaller than future reactors, and operate at considerably lower fusion performance, extensive validation of models is required for reliable, quantitative predictions. For example, negative triangularity scenarios have been studied in conventional aspect ratio devices, but the potential advantages for compact Spherical Tokamak (ST) devices operating at aspect ratio ⩽2 has not been extensively explored yet. Transport models, typically developed based on positive triangularity plasmas, also require more extensive validation— and possibly development—when transferred to negative triangularity scenarios [5]. To that end, the SMall Aspect Ratio Tokamak (SMART [6]) device is being commissioned at the University of Seville (Spain). The main goal of SMART is to test experimentally the potential advantages of negative triangularity scenarios in a ST configuration [6–10], and contribute to the overall understanding of the physics and transport scaling of ST-based fusion devices in terms of plasma confinement [11]. Experimental data will also provide valuable information for further development and validation of predictive transport models. The SMART Program includes three phases. After the initial commissioning of the device with Ohmic plasmas (Phase I), additional heating by Neutral Beam Injection (NBI) and enhanced capabilities for controlling the magnetic configuration are planned for Phase II. Phase III further expands the device capabilities in terms of achievable plasma current, magnetic field, auxiliary power and pulse duration. This work explores the optimization of NBI injection in SMART through the TRANSP and Orbit codes [12–14], in view of its commissioning in Phase II. The main goal is to estimate the range of plasma parameters that would be achievable in Phase II, and test a simulation workflow for further scenario development. These are necessary steps in the early phase of SMART, for example to provide estimates of plasma parameters for the design and optimization of diagnostics [15], as well as for an initial assessment of macroand micro-instability properties for Phase II [16,17]. Results are compared to those from previous optimization studies [10,18] conducted with different codes than those used in this work. This initial exercise allows to verify the consistency of results, especially on NBI optimization, across codes that employ different approaches to solve for NB ion dynamics, as briefly discussed in section 3. The target parameters for most simulations are conservative choices that do not require achieving the maximum values planned for Phase II, to demonstrate the device capabilities at the beginning of Phase II. Also, scenario predictions focus on the current ramp and flat-top phases. Little effort has been spent to optimize the ramp-down phase, which is typically much more difficult to predict since its evolution heavily depends on the previous plasma trajectory. After an introduction to the design parameters of SMART in section 2, the main tools used for simulations are briefly summarized in section 3. Optimization of the NBI parameters is discussed in section 4. Section 5contains the main results of this work. Section 5.1 presents the results for NBI optimization in the current flat-top, based on two target configurations. Section 5.2 expands the analysis to an entire discharge, including the current ramp-up phase. The main results from this work are finally summarized in section 6, which concludes the paper. 2. The SMART device The main parameters of the SMART device are summarized in table 1for the three planned phases of operations. Transition from one phase to the next will require achievement of engineering and physics milestones, as well as upgrades of some of the subsystems such as power supplies. In terms of physics, Phase I aims at commissioning the machine with a limited set of diagnostics and modest plasma performance. Installation and commissioning of a NBI system for auxiliary heating and developing a feedback control system are the main physics upgrades foreseen for Phase II, which will extend the range of flexibility of achievable scenarios in terms of plasma parameters. Further upgrades of power supplies and of the NBI system are foreseen during Phase II, in parallel with physics studies. Phase III is expected to achieve the full machine capability and scenario flexibility with extended pulse duration and the addition of electron-cyclotron injection as an additional auxiliary heating system. Engineering upgrades mostly include power supplies, plasma-facing components (from bare stainless steel to C tiles), and diagnostics. A central solenoid is used to drive inductive current [19]. The main (toroidal) magnetic field is generated through 12 rectangular toroidal field (TF) coils [9]. Initial operation (Phase I) will be limited to on-axis field B0⩽0.1T. Upgrades of the TF power supplies in the subsequent phases II and III will enable operations up to B0=1T. Three pairs of in-vessel poloidal field (PF) coils complement two out-vessel coils to provide the required flexibility in the achievable magnetic configurations [7,8,20]. According to initial modeling, a triangularity in the range δ=−0.6→ +0.6 can be achieved. The same PF coils are also used for vertical stability and divertor control. The vertical stability coils are the only ones planned to be controlled by a real-time feedback system, whereas all other coils will be controlled in feed-forward. The cylindrical vacuum vessel has a radius of 0.8 m and height of 1.6 m (figure 1) [7,19]. The vessel and coils geometry are expected to accommodate plasmas with major radius R0=0.45 m and minor radius at the midplane a=0.25 m. Several ports on the top/bottom of the vessel, as well as on the outboard wall, have been designed to host diagnostics, feedthroughs for the PF coils and access for the NBI system. 3. Simulation methodology Previous simulations in preparation for SMART operations used a combination of codes for magnetic configuration 2 Plasma Phys. Control. Fusion 66 (2024) 045021 M Podest` aet al Table 1. SMART parameters for the three phases. Phase I Phase II Phase III B0(T) 0.1 0.4 1.0 Ip(kA) 100 200–500 >500 R0(m) 0.45 a(m) 0.25 δ−0.6→+0.6 flat −top (ms) 150 500 1000 PNBI (MW) 0 0.3–1 1 PECH (kW) 6 6 200 Figure 1. The SMART device under commissioning at University of Seville, Spain. Reproduced with permission from [19]. CC BY-NC-ND 4.0. (FIESTA [21]) and transport analysis (ASTRA [22]). The first studies focused on Ohmic L-mode discharges [20]. Next, NBIheated scenarios were considered [10], including scenarios for Phase II for assessment of transition from L-mode to H-mode. Separate works also explored start-up scenarios to assess the requirements on the PF coils’ currents and power supplies [23], and the optimization of the NBI system starting from Phase II [10,18]. This work builds upon previous work by the SMART Team to expand the analysis to include NBI for timedependent scenario development, and to predict the impact of NBI on the achievable plasma parameters within a common framework provided by the TRANSP code [12]. As the simulations described here use different numerical tools and boundary conditions with respect to previous work, differences are expected in the results, especially in the case of predictive simulations where the magnetic equilibrium is evolved selfconsistently with transport models and during transient phases, like the plasma current ramp-up and ramp-down. These differences are indeed very valuable to identify uncertainties in the models and which boundary conditions have larger effect on the achieved global plasma parameters. 3.1. TRANSP code for scenario development TRANSP [12] is an equilibrium and transport code for tokamak experiments, currently used for interpretive and predictive transport simulations. In predictive mode, turbulence transport models are used to derive density, temperature and rotation profiles for given diffusion coefficients or fluxes. TRANSP includes a dedicated module, NUBEAM [13], for fast ion physics based on neoclassical physics. This work is limited to fixed-boundary simulations using LCFS information from previous work and the TEQ equilibrium solver [24] in TRANSP to re-compute the full equilibrium. For the predictive runs discussed in this work, the MultiMode Model (MMM [25]) for thermal transport is used as it has been shown to reproduce profiles in Spherical Tokamaks [11,26,27]. Comparing the results from multiple thermal transport models will be the subject of future work. MMM contains physics modules to describe thermal transport induced by several classes of instabilities including the effects of sheared flows [25]. As detailed in [25], MMM includes physics models for instabilities driven by electron/ion temperature gradients, trapped electrons, kinetic ballooning and peeling effects, collisionless and collision dominated magnetohydrodynamics drive, as well a model for drift resistive inertial ballooning modes. (The investigation of the mechanisms that dominate thermal transport by micro-instabilities in predicted SMART plasmas is beyond the scope of this work, and will be discussed in future publications once initial parameter ranges are assessed.) Of note, results from previous predictive analysis, e.g. reported in [10], were based on a-priori assumptions on transport coefficients inferred from scaling laws and/or other similar devices. This aspect is important to understand differences between the results previously reported and what is discussed in the following sections. Simulations discussed in section 4use analytic profiles for density, temperature and rotation, based on previous simulations or inferred from similar experiments. This is done to benchmark the NBI calculations. Next, when optimizing the NBI configuration, the density profile is prescribed and the electron and ion temperature profiles are predicted inside ρ=0.9 (see section 5), where ρis the normalized toroidal flux used as radial variable. Finally, simulations that predict both the ramp-up and the flattop phase (section 5.2) evolve the values of density and temperature at the separatrix, based on a two-point ScrapeOff-Layer (SOL) model [28]. The two-point SOL model extends previous work [29,30] to calculate the electron/ion temperature and plasma density in the SOL, given the particle and power fluxes from the core through the LCFS calculated by TRANSP and given recycling coefficients. This approach is used for time-dependent predictions in section 5.2. 3.2. Orbit code for NBI optimization studies Orbit [14] is a guiding-center, particle-following code mostly used to simulate the dynamics of high-energy particles (EPs) for a given magnetic field configuration. The code evolves 3 Plasma Phys. Control. Fusion 66 (2024) 045021 M Podest` aet al trajectories of a selected number of markers based on the Hamiltonian formulation for particles’ orbits in axi-symmetric configurations [31]. The study of the effects of perturbation fields, e.g. induced by plasma instabilities, is not considered in this work. The Orbit code is well-suited for a quick assessment of the EP confinement in a given configuration, which enables scans over a broad range of plasma conditions starting from a single scenario identified e.g. through TRANSP analysis. Markers for Orbit simulations can be extracted from a fast ion distribution obtained through TRANSP/NUBEAM. The distribution can be representative of either confined fast ions resulting from slowing-down and scattering, or of the deposited NB ions including possibly lost particles. Recent code developments [32] include the possibility to follow particles in the vacuum region between the last-closed flux surface (LCFS) and a limiter. This feature allows to estimate the poloidal distribution of lost energetic particles, which is required to assess the needs for protective structures preventing localized heat loads on the wall or diagnostic windows. (Incidentally, TRANSP/NUBEAM flags a marker as ‘lost’ if its position corrected for finite Larmor radius (FLR) effects travels farther than one Larmor radius from the LCFS. This implies that losses may be overestimated if the LCFS is located far from the wall, see additional discussion in section 5). Of note, a guiding-center approach has been often questioned when dealing with fast ion simulations for compact, low-field devices such as STs (see [33] and references therein). Under those conditions, the NB ion gyro-radius can be comparable to—or even greater than—typical scale length of the background fields, thus breaking one of the main assumptions behind the guiding-center formulation. As shown in the following Sections, a comparison between Orbit/NUBEAM and the full-orbit code ASCOT [34,35] gives confidence that sufficiently accurate results in terms of overall trends and global quantities can be achieved for the cases discussed in this work. More extensive discussion on this topic is left for future publications. 4. Simulations for NBI optimization Neutral Beam injection, or NBI, is widely used as an additional heating and current drive system on present tokamaks. Its optimization in terms of geometrical and beam parameters is required to minimize losses of NB ions, in order to maximize absorption and resulting plasma performance, see [36–38] among many others. In this section, scenarios from previous work are first used to cross-validate the NBI optimization results reported in [10, 18] with results from TRANSP/NUBEAM and Orbit analysis. As discussed later in this section, good agreement is found among different codes about expected NBI absorption and losses. The analysis is then extended to explore more NBI optimization options based on the scenarios discussed in section 5, especially for time-dependent simulations. The NBI optimization study is based on parameters from a NBI system whose parameters are comparable to those of the system in use on the COMPASS device [39,40]. The system will be available on SMART for Phase II. For the initial operations in Phase II of SMART, the expected injection energies are 15–40 keV and currents in the range 8–12 A, resulting in an injected power up to ≈400 kW once transmission and neutralizer losses are accounted for. Depending on the timing of further upgrades of the NBI system, the available power may increase to 1 MW, although a conservative upper value of 0.4 MW is assumed for most of the following analysis. The working gas is hydrogen. Note that in [18] the NBI injection voltage for most cases was limited to Vinj =15 kV. Experimentally, however, there can be interest in maximizing the amount of NB power delivered to the plasma to favor access to high confinement operations (H-mode) and to improve NB current drive. Therefore, an increased Vinj =25 kV has been typically used for this initial study, although it results in increased NB ion losses, as discussed later. A higher injection voltage is also consistent with the more recent study discussed in [10], although in that work it was assumed that the NB system would have been capable to deliver up to 1 MW of power at Vinj =25 kV. (No specific NBI system was targeted in [10], and 1 MW is likely to be achievable only by installing a second NBI line). The initial target configurations and plasma parameters used in this work from ASTRA simulations [10] are shown in figure 2for a pair of matched positive vs negative triangularity discharges. Density and electron temperature are comparable for the two cases, although the negative δcase is predicted to achieve slightly higher electron temperature with larger values of the outer gap, where the the outer gap is the distance between the last-closed flux surface (LCFS) on the outboard mid-plane and the vessel wall. Figure 3illustrates the geometry for scans of NB injection parameters, namely the tangency radius inside the plasma and the elevation of the NBI source outside the vessel to enable off-axis NBI. TRANSP simulations are first used to identify the optimum range of tangency radii that minimizes first-orbit losses and shine-through, thus maximizing the NB power effectively coupled to the plasma. Figure 4shows TRANSP/NUBEAM results of NBI losses for a tangency radius scan covering the entire radial range from highto low-field side, Rtan =19 cm−64 cm. The NBI is injected horizontally at the midplane. The injection voltage is set to Vinj =25 kV. The power lost through first-orbit losses remains negligible for Rtan ⩾40 cm, which roughly corresponds to the magnetic axis (figure 4(a)). For more perpendicular injection, 19 cm ⩽Rtan ⩽40 cm, first-orbit losses increase rapidly and up to 20%of the injected power is lost. Shine-through losses have a different dependence on Rtan. Except for the smaller Rtan =19 cm, the lost power remains ≈5%–10% up to Rtan ≈50 cm, then rapidly increases to >20%as the path of NB injection is reduced and plasma density and temperature decrease towards the edge. For the configurations shown in figure 2, shine-through losses are generally higher for the negative 4 Plasma Phys. Control. Fusion 66 (2024) 045021 M Podest` aet al Figure 2. Plasma scenarios for NBI tangency radius scan. (a) Plasma boundary in a poloidal cross-section for positive (blue) and negative (red) triangularity, in addition to the wall boundaries and in-vessel PF coils shown as thick black lines (PF coils outside the vessel on the low-field side are not shown). Electron density (b) and temperature (c) are shown for the target positive and negative triangularity scenarios from panel (a). Solid (dashed) lines in panel (c) refer to electron (ion) temperature. The abscissa in (b) and (c) is the toroidal flux normalized to its value at the LCFS. Figure 3. (a) Schematic of SMART at the midplane as seen from above, with lines repesenting the vessel wall at R=0.8 m (black), the magnetic axis at R=0.45 m (dashed), and three radii corresponding to Rtan =20,44,64 cm. Solid lines show the NBI path corresponding to the three tangency radii. (b) Vertical cross section showing the NBI path outside the vessel (solid lines) and the corresponding deposition locations at guiding center for three elevations of the NBI source, Ztilt =−50,0,+50 cm. Figure 4. NBI losses vs. tangency radius for midplane NB injection. (a) Fraction of injected NBI power lost through first-orbit losses to the wall (solid) and through shine-through (dashed) calculated by TRANSP/NUBEAM. (b) Fraction of lost particles calculated by TRANSP/NUBEAM (black, green) and through Orbit for the positive triangularity case. As in panel (a), the dashed line correspond to shine-through losses from TRANSP. 5 Plasma Phys. Control. Fusion 66 (2024) 045021 M Podest` aet al Figure 5. (a) Contour of the nominal TF ripple field amplitude in a poloidal cross-section for a negative triangularity configuration. (b) Particle lost fraction as a function of the TF ripple amplitude, rescaled with respect to the nominal amplitude. triangularity case, most likely due to the smaller crosssection, which results in a slightly reduced NB path inside the plasma. Prompt losses calculated by NUBEAM are compared to those from Orbit in figure 4(b). The comparison is shown in terms of fraction of markers that are lost, rather than power as in figure 4(a), as that information is more readily available from Orbit. The two codes use different conditions to classify markers as ‘lost’. NUBEAM considers a marker lost when it either gets closer than one Larmor radius to a wall/limiter structure, or when its guiding center position is farther out than one Larmor radius from the LCFS. Therefore, some particles that exit the main plasma due to drifts but would otherwise reenter the plasma may contribute to enhance first orbit losses. By default, Orbit marks particles as lost once their guiding center crosses the LCFS. Similar to NUBEAM, this could lead to over-estimated losses. An option to follow particles in the vacuum region between the LCFS and the wall has been recently implemented [32]. In this case, particles are lost once their FLR-corrected position hits the wall. Results of the NUBEAM/Orbit comparison shown in figure 4(b) confirm the expectations. First orbit losses to the LCFS from Orbit exceed those from NUBEAM by 10%–15%. When particles are allowed to re-enter the plasma in Orbit, losses decrease by ≈10%with respect to results from NUBEAM. In general, however, the two codes agree on the main trend vs. Rtan. (Note that losses due to charge-exchange with edge neutrals, included in NUBEAM, are not implemented in Orbit. According to TRANSP, their contribution is ⩽3%for the two cases analyzed here). Another possible mechanism for NB ions loss are ripple perturbations to the static magnetic field caused by the finite number of TF coils. Based on the geometry and number of the coils, the TF ripple field amplitude is up to 1% at the plasma edge on the low-field side, and rapidly decreases to ≪1% inside the plasma, cf figures 7 and 8 in [9]. Orbit simulations are performed to assess the NB ion losses associated with TF ripple for a worst-case scenario with negative triangularity, in which the plasma region on the low-field side is maximized. As shown in figure 5, losses remain limited to ≈0.5%for the nominal value of the TF ripple field. Only by assuming a ripple field 4–10 times larger than the nominal value we find that the losses become comparable to those caused by other mechanisms, such as charge-exchange and shine-through. Rippleinduced losses are therefore neglected in the following analysis. The results shown in figure 4are consistent with previous work, which indicated an optimum tangency radius Rtan = 44 cm [18]. The consistency of the results across several codes (ASCOT, Orbit and TRANSP/NUBEAM) gives confidence in their validity, as well as in the validity of using TRANSP to extend the study to time-dependent simulations. As discussed previously, in [18] the NBI injection voltage for most cases was limited to Vinj =15 kV as opposed to the Vinj =25 kV used in figure 4. Figure 6shows the variation of NB ion losses for Rtan =44 cm and positive triangularity as the injection voltage is varied between 20 kV and 40 kV, overlapping with the Vinj values explored in [10]. Shine-through losses steadily increase from 3%up to 10%for the maximum injection voltage, while first-orbit losses to the wall remain negligible. For the maximum available NB current of 12 A, the injected power increases from 204 kW to 408 kW for Vinj = 20 →40 kV. Once losses are taken into account, that translates into an increase in coupled NBI power from ≈200 kW to ≈370 kW, which provides significant flexibility in terms of additional heating capability. Additional flexibility in the profiles achievable by NBI heating in Phases II and III on SMART can be achieved by considering a more elaborate implementation of the NBI system, in which the NBI injector can be tilted vertically. Such flexibility clearly involves complexity in terms of engineering, but it also enables opportunities in terms of physics research and scenario control as demonstrated on other facilities such as DIII-D [36]. Examples of deposited NB ions and their profile after a 1 ms orbiting time are shown in figure 7. It is assumed that the NBI source (about 2 m away from the vacuum vessel) can 6 Plasma Phys. Control. Fusion 66 (2024) 045021 M Podest` aet al Figure 6. Fraction of injected NBI power lost through first-orbit losses to the wall (solid) and through shine-through (dashed) as a function of the NBI injection voltage. Losses to the wall are multiplied by 100 for clarity. First-orbit losses are negligible for Vinj <40 kV, hence the results are noisy due to poor statistics of the lost markers. Figure 7. Examples of NBI deposited markers (green) and their relaxed profile after 1 ms orbiting (black). The NBI is tilted from Ztilt =−50 cm to Ztilt = +50 cm at fixed nominal tangency radius Rtan =44 cm. The yellow markers represent deposited NB ions lost to the wall. be displaced vertically in the range Ztilt =±50 cm keeping the same cross-over point at the vacuum vessel. The nominal Rtan =44 cm for midplane injection is assumed for all cases. A sample of NB deposited ions is first obtained through NUBEAM, then Orbit is used to evolve particles’ trajectories. First-orbit losses to the wall are recorded as discussed above. From the distribution of markers shown in figure 7, it is clear that the possibility of tilting the NBI source would provide a great advantage in terms of achievable NB ion profiles. That is further demonstrated in figure 8, which shows the NB ion density and resulting NB driven current profiles for Ztilt =−50 →+50 cm. In fact, the possibility of varying the radial gradient of the NB ion profile would provide an excellent tool for controlled studies of the role of NB ions in driving deleterious instabilities such as Alfvénic modes. In addition to modifying the radial NB ion density profile, a tiltable NBI system would also provide good control over the NB ion distribution in velocity space. Figure 9 shows examples of the NB ion distribution vs energy and pitch (ratio of parallel to total velocity) for Ztilt =−50,0,+50 cm and Vinj =40 kV. As the tilt is increased from more negative to more positive, thus leading to a better alignment of the NBI injection path with the local magnetic field, the fraction of NB ions shifts from more trapped particles (pitch near zero) to more co-passing particles (pitch near 1). This fine control of the EP distribution resulting from NBI would also provide great flexibility for fast ion studies on a device such as SMART, enabling studies that are not easily achievable/controllable on larger fusion devices. 7 Plasma Phys. Control. Fusion 66 (2024) 045021 M Podest` aet al Figure 8. (a) NB ions density and (b) current density as the NBI is tilted from Ztilt =−50 cm to Ztilt = +50 cm for fixed nominal tangency radius Rtan =44 cm. Figure 9. NB ion distributions at ρ=0.3 for Ztilt =−50,0,+50 cm. All distributions are normalized by a common factor. 5. NBI-heated scenarios for Phase II of SMART This section discusses initial results on scenario development for NBI-heated plasmas in Phase II of SMART operations. Based on the results on NBI optimization from section 4, simulations are first limited to predicting expected values of electron and ion temperature for a prescribed plasma rotation in the current flat-top phase. To reduce uncertainties in the predictive runs, electron density and plasma rotation are prescribed, as their predictions are presently unreliable when multiple transport channels are simultaneously activated in TRANSP predictions [41]. Density and rotation values are based on available results from other compact (or spherical) tokamaks with parameters similar to SMART, such as LTX [42,43] and Globus-M [44]. Once reliable profiles are obtained for the current flat-top phase, predictions of the discharge are then extended backwards in time for an initial assessment of the plasma evolution during the current ramp, cf section 5.2. 5.1. Profile prediction in the current flat-top As a starting point, the same magnetic configuration shown in figure 2for the positive triangularity case is selected. Figure 10. (a) Density profiles at the beginning of the simulation, assuming the plasma current has already achieved its flat-top value Ip=200 kA. (b) Initial guesses for Te(solid) and Ti(dashed) profiles used in TRANSP predictions during the current flat-top phase. The plasma is assumed to remain in low-confinement mode (L-mode) for the entire simulation, thus the achieved performance can be interpreted as a lower limit. The initial guesses of plasma profiles for the transport solver in TRANSP are shown in figure 10. As mentioned above, density and rotation profiles are prescribed. Based on ASTRA [22] simulations, cases 8 Plasma Phys. Control. Fusion 66 (2024) 045021 M Podest` aet al Figure 20. Example of the effects of large MHD events such as sawteeth on the predicted Te,Tiprofiles. (a) Evolution of the on-axis q-profile withouth sawteeth (black), with infrequent sawteeth (blue) and frequent sawteeth (red). (b) Safety factor radial profile for the case without sawteeth (black) and for the two other cases. Solid (dashed) lines refer to profiles just before (after) a crash. (c) and (d) Evolution of the on-axis Teand Tivalues withouth sawteeth (black), with infrequent sawteeth (blue) and frequent sawteeth (red). (e) and (f) Radial profiles of Teand Tifor the case without sawteeth (black) and for the two other cases. Solid (dashed) lines refer to profiles just before (after) a crash. Profiles are taken from the time range 100 ms ⩽t⩽120 ms, corresponding to the current flat-top. 6. Conclusions The SMART tokamak under commissioning at the University of Seville (Spain) will explore confinement properties of negative triangularity configurations in a compact (spherical) tokamak. After an initial phase limited to Ohmic heating, installation of a NBI system for auxiliary heating is planned to extend the achievable plasma regimes and performance. Numerical optimization of NBI-heated scenarios indicates an optimal value for the NBI tangency radius, Rtan =44 cm, that minimizes first-orbit losses and shine-through for the plasma parameters expected in Phase II. Simulations that assume a vertically steerable NB injection show the potential advantage of such flexible system in terms of scenario and fast ion distribution control, which would compensate the increased complexity (and cost) from the engineering point of 15 Plasma Phys. Control. Fusion 66 (2024) 045021 M Podest` aet al view. Predictions based on the increased flexibility in NBI will inform possible upgrades of the NBI system from Phase II to Phase III of SMART operations. Based on initial simulations for the current flat-top phase with prescribed density and rotation, electron and ion temperatures are predicted to increase by 35%–50% for the nominal NBI parameters of Rtan =44 cm and Vinj ≈25 kV compared to purely Ohmic discharges. This confirms the added value of a NBI system capable to deliver 0.25–1 MW power to SMART plasmas to extend the range of achievable parameters for thermal and fast ion transport studies. The extension of predictive simulations for Teand Tito the entire discharge, including the current ramp-up and a self-consistent estimate of pedestal parameters, provides more insight on the predicted parameters. It is clear that adding selfconsistent boundary conditions through a reduced scrape-off layer model [28] largely modifies the simulations’ outcome, for instance by suggesting more peaked profiles with lower edge temperature than what initially predicted in [7]. Future work will focus on this issue, as well as on the comparison across different thermal transport models to assess the reliability of present numerical tools for scenario predictions. Data availability statement The data cannot be made publicly available upon publication because they are not available in a format that is sufficiently accessible or reusable by other researchers. The data that support the findings of this study are available upon reasonable request from the authors. Acknowledgments This manuscript has been authored by Princeton University and is based upon work supported by the U.S. Department of Energy, Office of Science, Office of Fusion Energy Sciences, under Contract Number DE-AC02-09CH11466 with the U.S. Department of Energy. This work has been carried out within the framework of the EUROfusion Consortium, funded by the European Union via the Euratom Research and Training Programme (Grant Agreement No 101052200— EUROfusion). Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Commission. Neither the European Union nor the European Commission can be held responsible for them. ORCID iDs M Podest` ahttps://orcid.org/0000-0003-4975-0585 D J Cruz-Zabala https://orcid.org/0000-0001-5925-5153 F M Poli https://orcid.org/0000-0003-3959-4371 J W Berkery https://orcid.org/0000-0002-8062-3210 M Garcia-Mu˜ noz https://orcid.org/0000-0002-3241-502X E Viezzer https://orcid.org/0000-0001-6419-6848 S M Kaye https://orcid.org/0000-0002-2514-1163 References [1] Menard J E, Grierson B A, Brown T, Rana C, Zhai Y, Poli F M, Maingi R, Guttenfelder W and Snyder P B 2022 Fusion pilot plant performance and the role of a sustained high power density tokamak Nucl. Fusion 62 036026 [2] Camenen Y, Pochelon A, Behn R , Bottino A, Bortolon A, Coda S, Karpushov A, Sauter O and Zhuang G (TCV team) 2007 Impact of plasma triangularity and collisionality on electron heat transport in TCV L-mode plasmas Nucl. Fusion 47 510 [3] Austin M E et al 2019 Achievement of reactor-relevant performance in negative triangularity shape in the DIII-D tokamak Phys. Rev. Lett. 122 115001 [4] Marinoni A et al 2021 Diverted negative triangularity plasmas on DIII-D: the benefit of high confinement without the liability of an edge pedestal Nucl. Fusion 61 116010 [5] Lyons B C et al 2023 Flexible, integrated modeling of tokamak stability, transport, equilibrium and pedestal physics Phys. Plasmas 30 092510 [6] Garcia-Mu˜ noz M et al 2023 Physics basis, design and commissioning of the SMall Aspect Ratio Tokamak (SMART) 29th IAEA Fusion Energy Conf. (London, UK) p 8.2155 [7] Mancini A et al 2021 Mechanical and electromagnetic design of the vacuum vessel of the SMART tokamak Fusion Eng. Des. 171 112542 [8] Doyle S J et al 2021 Magnetic equilibrium design for the SMART tokamak Fusion Eng. Des. 171 112706 [9] Agredano-Torres M et al 2021 Coils and power supplies design for the SMART tokamak Fusion Eng. Des. 168 112683 [10] Mancini A et al 2023 Predictive simulations for plasma scenarios in the SMART tokamak Fusion Eng. Des. 192 113833 [11] Kaye S M, Connor J W and Roach C M 2021 Plasma Phys. Control. Fusion 63 123001 [12] Breslau J, Gorelenkova M, Poli F, Sachdev J, Pankin A and Perumpilly G 2018 TRANSP, US-DoE computer software (available at: https://doi.org/10.11578/dc.20180627.4) [13] Pankin A, McCune D, Andre R, Bateman G and Kritz A 2004 The tokamak Monte Carlo fast ion module NUBEAM in the National Transport Code Collaboration library Comput. Phys. Commun. 159 157 [14] White R B and Chance M S 1984 Phys. Fluids 27 2455 [15] Kaur M et al 2023 Design of a multi-pulse thomson scattering system for the SMART tokamak 65th Annual Meeting of the APS Division of Plasma Physics (Denver, Colorado) [16] Dominguez-Palacios J et al 2023 First MHD stability analysis of the SMall Aspect Ratio Tokamak (SMART) 7th Asia-Pacific Conf. on Plasma Physics (Port Messe, Nagoya, 12–17 November 2023) [17] Labbate J A et al 2023 Analyzing mode stability to identify stable operating regions for the novel SMall Aspect Ratio Tokamak (SMART) 65th Annual Meeting of the APS Division of Plasma Physics (Denver, Colorado) [18] Velarde-Gallardo L 2021 Neutral beam injector design for the SMall Aspect Ratio Tokamak (SMART) of the University of Seville Master Thesis University of Seville [19] Segado-Fernandez J et al 2023 Analysis and design of the central stack for the SMART tokamak Fus. Eng. Des. 193 113832 [20] Doyle S J et al 2021 Single and double null equilibria in the SMART tokamak Plasma Res. Express 3044001 [21] Buxton P F and Gryaznevich M P 2017 Merging compression start-up predictions for ST40 Fusion Eng. Des. 123 551 [22] Pereverzev G V and Yushmanov P N 2002 Astra; automated system for transport analysis IPP Report 5/98 (Max-Planck-Institut Für Plasmaphysik) 16 Plasma Phys. Control. Fusion 66 (2024) 045021 M Podest` aet al [23] Sanjuán J P 2020 Optimization of the plasma start-up of the Seville spherical tokamak (SMART) Master Thesis University of Seville [24] LoDestro L L and Pearlstein L D 1994 On the grad-shafranov equation as an eigenvalue problem, with implications for q solvers Phys. Plasmas 190 [25] Rafiq T, Kritz A H, Weiland J , Pankin A Y and Luo L 2013 Physics basis of multi-mode anomalous transport module Phys. Plasmas 20 032506 [26] Berkery J (NSTX-U Team) 2023 NSTX-U research advancing the physics of spherical tokamaks 29th IAEA Fusion Energy Conf. (London, UK) [27] Rafiq T 2023 Plasma profile prediction in NSTX discharges using the updated multi-mode anomalous transport module 29th IAEA Fusion Energy Conf. (London, UK, October 2023) [28] Zhang X et al 2023 Reduced physics model of the tokamak scrape-off-layer for pulse design Nucl. Mater. Energy 34 101354 [29] Stangeby P C 2000 The Plasma Boundary of Magnetic Fusion Devices (Taylor and Francis) [30] Goldston R J 2011 Heuristic drift-based model of the power scrape-off width in low-gas-puff H-mode tokamaks Nucl. Fusion 52 013009 [31] White R B 2006 The Theory of Toroidally Confined Plasmas 2nd edn (Imperial College Press) [32] Bonofiglo P, Podesta` a M, White R, Kiptily V, Goloborodko V, Gorelenkov N, Vallar M, Cecil F, Giroud C and Brezinsek S (JET Contributors) 2022 Nucl. Fusion 62 026026 [33] Cecconello M et al 2019 Discrepancy between estimated and measured fusion product rates on MAST using TRANSP/NUBEAM Plasma Phys. Control. Fusion 59 016006 [34] Hirvijoki E, Asunta O, Koskela T, Kurki-Suonio T, Miettunen J, Sipilä S, Snicker A and Äkäslompolo S 2014 ASCOT: Solving the kinetic equation of minority particle species in tokamak plasmas Comput. Phys. Commun. 185 1310 [35] Akaslompolo S, Drevlak M, Turkin Y , Bozhenkov S, Jesche T, Kontula J, Kurki-Suonio T and Wolf R C 2018 Modelling of NBI ion wall loads in the W7-X stellarator Nucl. Fusion 58 082010 [36] Scoville J, Boyer M, Crowley B , Eidietis N W, Pawley C J and Rauch J M 2019 New capabilities and upgrade path for the DIII-D neutral beam heating system Fusion Eng. Des. 146 6 [37] Kim J Y, Rhee T, Kim J , Yoon S W, Park B H, Isobe M, Ogawa K and Ko W H 2016 Prompt loss of beam ions in KSTAR plasmas AIP Adv. 6105013 [38] Wang J, Wu B and Hu C 2010 Simulation on heating and current drive using neutral beam injection with variable injection angle on EAST Plasma Sci. Technol. 12 289 [39] Deichuli P, Davydenko V and Belov V 2012 Commissioning of heating neutral beams for COMPASS-D tokamak Rev. Sci. Instrum. 83 02B114 [40] Bogar K, Varju J, Havlicek J , Stockel J, Podolnik A, Hron M and Panek R 2023 Plasma Phys. Control. Fusion 65 025004 [41] Teplukhina A A et al 2021 Sensitivity studies on the TRANSP particle balance equation for JET deuterium and hydrogen-deuterium plasma discharges 47th EPS Conf. on Plasma Phys. (Sitges, Spain) vol 45A p 4.1065 [42] Boyle D P, Majeski R, Schmitt J C, Hansen C, Kaita R, Kubota S, Lucia M and Rognlien T D 2017 Observation of flat electron temperature profiles in the Lithium Tokamak eXperiment Phys. Rev. Lett. 119 015001 [43] Boyle D et al 2023 Extending the low-recycling, flat temperature profile regime in the Lithium Tokamak eXperiment-β(LTX-β) with ohmic and neutral beam heating Nucl. Fusion 63 056020 [44] Gusev V et al 2015 Review of GLOBUS-M spherical tokamak results Nucl. Fusion 55 104016 [45] Taroni A, Erbai M, Springmann E and Tibone F 1994 Global and local energy confinement properties of simple transport coefficients of the Bohm type Plasma Phys. Control. Fusion 36 1629 17