Elsevier

Fusion Engineering and Design

Volume 146, Part B, September 2019, Pages 2338-2342
Fusion Engineering and Design

Comparison of approaches to the electromagnetic analysis of COMPASS-U vacuum vessel during fast transients

https://doi.org/10.1016/j.fusengdes.2019.03.185Get rights and content

Abstract

The poloidal distribution of electromagnetic loads during fast transients in the vacuum vessel of COMPASS-U tokamak is calculated analytically. The estimates are then compared with CarMa0NL numerical simulations. The results show that the poloidal eddy currents in the vacuum vessel must be taken into account for the proper evaluation of disruption forces in the COMPASS-U tokamak.

Introduction

This paper contributes to R&D activities for the vacuum vessel (‘wall’) of COMPASS-U tokamak [1], which is a medium-size device presently in the conceptual design phase. It will operate at high plasma current (2 MA) and strong magnetic field (5 T), therefore, large electromagnetic (EM) forces on conducting structures near the plasma are expected during disruptions.

The forces appear due to the currents induced or injected into the wall. Traditionally the toroidal current in the wall is considered as the main attribute of disruptions [2,3], for more references see [4]. Recently it was demonstrated that the poloidal currents in the wall can contribute significantly into the disruption forces [[5], [6], [7]]. This study is aimed at evaluation of this effect in COMPASS-U tokamak.

The wall currents are driven by the electric field E induced by the magnetic field variations: ×E=B/t. The force balance equation p=j×B requires B/t0 in the plasma at changes of its pressure p (here, at thermal quenches, TQs) and current density j (here, at current quenches, CQs). This is accompanied by the changes in the plasma shape and position, which also contributes into E in the wall and greatly complicates the description of plasma-wall electromagnetic coupling [4,8,9]. Here both the toroidal and poloidal wall currents are accounted for assuming their purely inductive generation (no halo currents).

Calculation of the poloidal currents in the wall is an additional complication that should be justified. For this purpose, we start from the estimates based on the available theory [4,6,7] and make CarMa0NL [8,10] computations to test the validity of its predictions.

To facilitate the task, we approximate the COMPASS-U tokamak with 3D vacuum vessel (VV) shown in Fig. 1 by a circular tokamak, as illustrated in Fig. 2. Moreover, we do not consider VDEs and related halo currents, but only TQs and CQs, which are assumed to be instantaneous, so that we can treat the wall as ideal.

The motivation of this basic 2D research is to provide a necessary basis for the 3D EM modelling of axisymmetric disruptions on COMPASS-U, which is currently being carried out with ANSYS Maxwell and CarMa0NL. Assuming that mechanical and electromagnetic problems are decoupled, the results of EM modelling are being used for static and transient structural analysis in ANSYS Mechanical. Coupled electromechanical formulation may be considered at later stages of the design process [11].

The mechanical stress on the vessel depends on the time history of the EM loads and their durations. The model described below provides estimates for the EM loads during the worst scenario when all the current is transferred instantaneously to the wall. In this case the duration of the EM force is determined by the resistive decay time of the wall only.

Analytical expressions for local and integral forces on the circular tokamak [7] wall have been derived for the case when the plasma-wall gap on high-field-side (HFS) is smaller than the gap on low-field-side (LFS), see Fig. 2a. Their predictions for a high-aspect-ratio tokamak have been recently verified by CarMa0NL code [12]. In the present study, we investigate the opposite setup when the plasma column is shifted towards HFS (Fig. 2b), which corresponds to several plasma scenarios for COMPASS-U. As will be demonstrated in the following sections, in this case the analytical predictions and numerical results are also in good quantitative agreement.

As was found in [13] for COMPASS-U with pre-disruption poloidal beta βJ=0.5 and net plasma current J=2MA, the integral radial force on the wall during TQ is four times smaller than that during CQ. Therefore, in the present study we compare analytical and numerical results primarily for the CQ and adjust poloidal beta for numerical calculations to be as low as βJ=0.13. Note for completeness that similar comparisons for TQ forces have been done in [3,14] with greatly contradicting numerical results.

Section snippets

Results for the case with account of the poloidal current in the wall (Aw)

To estimate the EM forces during disruptions, we approximate the COMPASS-U tokamak with 3D VV, shown in Fig. 1, by a circular tokamak with major radius R0=1m and minor radius bw=0.5m, as illustrated in Fig. 2. With plasma minor radius b=0.3m, this choice of wall parameters keeps the same the ratio between plasma and vessel cross-section areas for circular 2D and elongated 3D geometries.

For internal inductance li=0.94 and poloidal beta βJ=0.13, the Shafranov shift Δiw calculated with

Numerical results obtained with CarMa0NL

In this section we compare analytical predictions (Fig. 3) with CarMa0NL numerical results, presented in Fig. 4, Fig. 5. The CarMa0NL [8,10] solves 2D nonlinear evolutionary equilibrium MHD equations, self-consistently coupled to eddy currents equations, describing 3D volumetric conductors. Despite the 2D nature of the problem under consideration, the choice of CarMa0NL is suggested by the need to include poloidal eddy currents generated in the conducting structures, a rare feature for

Discussion and conclusions

Comparing Figs. 3a and b with 5a and b, respectively, one can see good qualitative agreement between analytical and numerical results for the poloidal distribution of the normal component of EM force. For the case without poloidal eddies, also quantitative agreement for m = 0 harmonic is excellent, within only 6% difference. The difference for other harmonics may be caused by two factors. First, the formulas in [6,7,16] are derived for a circular vessel in the high aspect ratio approximation,

Acknowledgments

This work has been carried out within the framework of the project COMPASS-U: Tokamak for cutting-edge fusion research No. CZ.02.1.01/0.0/0.0/16_019/0000768 and co-funded from European structural and investment funds.

References (16)

  • R. Panek

    Fusion Eng. Des.

    (2017)
  • V.D. Pustovitov

    Fusion Eng. Des.

    (2017)
  • F. Villone et al.

    Fusion Eng. Des.

    (2015)
  • P. Testoni

    Fusion Eng. Des.

    (2018)
  • S. Miyamoto

    Nucl. Fusion

    (2014)
  • R.R. Khayrutdinov et al.

    Plasma Phys. Control. Fusion

    (2016)
  • V.D. Pustovitov

    Plasma Phys. Control. Fusion

    (2017)
  • V.D. Pustovitov et al.

    Plasma Phys. Control. Fusion

    (2018)
There are more references available in the full text version of this article.

Cited by (20)

  • Analytical solution of tokamak vibrations during axisymmetric plasma disruptions

    2022, Fusion Engineering and Design
    Citation Excerpt :

    In the recent decades the CAD and CAE tools developed considerably, allowing complex 3D multi-physical simulations informing the design stage. Numerous electromagnetic, thermal, CFD and mechanical models have been created during the design stage of the COMPASS-U tokamak [10–17]. Especially important was the identification of the worst-case loads for the machine – which were found to be the Vertical Displacement Events (VDE).

  • Preliminary design of the COMPASS upgrade tokamak

    2021, Fusion Engineering and Design
    Citation Excerpt :

    Electromagnetic forces from disruption-induced currents within the insulation are reduced by using low-conductivity metals (silver coated stainless steel or titanium) as a reflector and having additional insulating cuts in the reflector layers. An extensive electromagnetic modelling of plasma disruption events expected in the COMPASS-U tokamak was conducted [4,5]. List of disruption sequences which are considered as the most severe for the mechanical design of the vacuum vessel was proposed [4].

  • Effects of structural and electromagnetic loads on vacuum vessel of Pakistan spherical Tokamak

    2021, Fusion Engineering and Design
    Citation Excerpt :

    The SFD increases because of the thickness of the inner cylinder where the eddy current and EM force on the inner cylinder increases. It depends upon the wall characteristic of the VV and current in the vessel wall [25]. It also increases with area discontinuity at the top and bottom of the cylinder where the thickness is reduced from 8 mm.

  • Electromagnetic model for finite element analyses of plasma disruption events used in the design phase of the COMPASS-U tokamak

    2021, Fusion Engineering and Design
    Citation Excerpt :

    The codes were extensively benchmarked against available experimental data and were also used to predict the electromagnetic loads resulting from disruptions in the design phase of different tokamaks, e.g. EAST [6], JT60-SA [7], ITER [8,9]. Similar self-consistent predictions were previously performed for the COMPASS-U tokamak by means of the CarMa0NL code [10]. In the presented paper, a non-self-consistent engineering approach, which focuses on finding the maximal possible envelope of the disruption induced mechanical loads on the vacuum vessel assembly, is followed.

View all citing articles on Scopus
View full text