Impact of the New TCV Baffled Divertor Upgrade on Pedestal Structure and Performance

A new set of carbon tiles, neutral beam heating optics and gas baffles were installed on TCV during the baffled divertor upgrade in early 2019. The installation of the baffles allows a deconvolution of the roles of main chamber and divertor neutral pressure on the H-mode pedestal structure. This physical barrier allows relatively high neutral pressures to be constrained to the divertor, thus preventing neutrals from entering the main chamber and potentially degrading core confinement. This study presents the experimentally measured and modelled pedestal heights and structure for a series of H-mode discharges prior to and after this upgrade. Increased pedestal performance at high divertor neutral pressure was observed after the baffled divertor upgrade. This was consistent across all triangularities and outer target locations investigated and is attributed to higher pedestal top temperatures being maintained at high gas injection rates. ASTRA simulations indicated beam heating power coupled to the plasma did not significantly vary after the baffled divertor upgrade or as a function of divertor neutral gas pressure. Analysis of the pedestal structure exposed a strong correlation between pedestal performance and the density pedestal position prior to and after the baffled divertor upgrade. The baffled divertor upgrade limited the outward shift of the density pedestal, thus maintaining higher pedestal performance at high divertor neutral pressures. Stability analysis indicated the majority of discharges studied were within 25% of the stability boundary. No correlation was found between the distance from the stability boundary and pedestal performance or structure. Comparison with the EPED1 model indicated that TCV discharges do not have a fixed dependence between pedestal βθ and pedestal width. A large variation in the EPED1 relating parameter was observed and found to vary with the density pedestal position.


Introduction 1
The high confinement plasma mode (H-mode) is de-2 fined by an edge transport barrier that produces strong 3 temperature and density gradients termed the pedestal [1]. 4 This operational mode produces the highest performance 5 discharges and it is currently foreseen that next step fusion 6 devices will operate in H-mode with a detached divertor. 7 Access to detachment is primarily achieved through addi- 8 tional fuelling and puffing of impurities and this can have a 9 significant influence on the pedestal, which has been shown 10 to be strongly linked to fusion yield [2]. 11 Experiments on ASDEX Upgrade (AUG), JET and un-12 baffled TCV have shown that increased fuelling can reduce 13 pedestal performance. Analysis of AUG discharges showed 14 a change in the high field side high density (HFSHD) re- degrading the pedestal stability [3,4,5]. No definitive ex- 17 planation has yet been produced for JET and TCV but 18 a correlation with the relative shift in temperature and 19 density pedestal positions has been reported [6, 7, 8,9]. 20 An understanding of the structure of the pedestal is 21 given in terms of the EPED framework [10]: the pedestal 22 gradient is set by a transport limit (often taken to be a 23 kinetic ballooning mode (KBM) or similar proxy) while 24 the combination of the gradient and pedestal width (ul- 25 timately, the pedestal height) is determined by the onset 26 of a global peeling-ballooning mode (triggering an ELM 27 crash). A characteristic pedestal cycle begins with a steep-28 ening pressure gradient until a maximum limit is reached. 29 This limit is dictated by transport and modelled using 30 KBM stability as a proxy. The pedestal width then in-31 creases at the maximum gradient until the PB stability 32 boundary. The gradient limit can be inverted to a relation 33 between the pedestal height and width (w) to give the 34 well known dependence with pedestal poloidal β (β P ed. which can significantly alter pedestal performance [8,10]. and fitted with an mtanh function as described in [16,17].

77
The profiles are radially shifted such that the seperatirx 78 temperature is set to 50 eV, a value obtained from previous 79 TCV database scaling [18]. few discharges with ECRH are omitted from this analysis.

105
Fuelling rates of up to 10 21 e/s and nitrogen seeding rates 106 up to 10 20 e/s were applied during the discharges. For 107 simplicity, fuelling and seeding rates will be categorised as 108 outlined in Table 1.  in ITERH-98(y,2) with increasing divertor pressure was 119 measured for both divertor configurations but a weaker 120 decrease was measured after the baffled divertor upgrade.

121
Discharges after the upgrade were also able to produce 2-  tral indicates that core energy transport is stiff but particle 153 transport is not.

154
The influence of δ T and R t on stored energy and P ped 155 is presented in Figure 5. Discharges after the divertor up-156 grade produced generally higher ITERH-98(y,2) and P ped 157 across the range of δ T . No trends between ITERH-98(y,2) 158 and δ T , P ped or R t were observed. A weak declining trend 159 in maximum achievable P ped was found with increasing δ T 160 Figure 5 (middle) and its cause has not yet been identified.

161
It can be concluded from the large variation in P ped for a 162 given δ T or R t that the influence of shaping is significantly 163 lower than that of the divertor pressure or baffled divertor 164 upgrade.