The performance of an electromagnetic flowmeter head is assessed in terms of a weight vector W such that the output voltage ∝ ∫ v. Wdτ, where v is the velocity and τ the flowmeter volume. The condition curl W = 0 with W → 0 at ∞ is shown to be necessary and sufficient for the velocity to depend only on the flow rate and not on the flow pattern. A class of such ‘ideal’ meters is described. It is shown that meters with point electrodes can never be ideal but may, with considerable complication of the magnetic field, be made immune to asymmetric velocity-profile variations if the flow is rectilinear.

Results are presented from thin (resistive) shell experiments on HBTX and compared with theoretical (linear and non-linear) studies of the plasma stability. Current pulses of 3--5 ms are obtained, compared with the shell time constant for vertical field penetration of 0.5 ms. Theoretically predicted thin shell modes, phase locked to the wall, are prominent experimentally.

The operational limits observed in spherical tokamaks, notably the small tight aspect ratio tokamak ͑START͒ device ͓A. Sykes et al., Nucl. Fusion 32, 694 ͑1992͔͒, are consistent with those found in conventional aspect ratio tokamaks. In particular the highest ␤ achieved ͑ϳ40%͒ is consistent with an ideal magneto-hydro-dynamic ͑MHD͒ Troyon type limit, the upper limit on density is well described by the Greenwald density (a 2 n e /I p ϳ1) and the normalized current (I p /aB t ) is limited such that q 95 տ2. Stability calculations indicate scope for increasing both normalized ␤ and normalized current beyond the values so far achieved, although wall stabilization is generally needed for low-n modes. In double null configurations current terminating disruptions occur at each of the operational boundaries, though the current quench tends to be slow at the density limit and disruptions at high ␤ may be due to the low q. In early limiter START discharges, before the divertor coils were installed, disruptions rarely occurred. Instead internal reconnection events which have all the characteristics of a disruption except the current quench occurred. These various disruptive behaviors are explained in terms of a model in which helicity is conserved during the disruption. Due to the low toroidal field beam ions in START, and ␣ particles in a ST power plant, are super-Alfvénic. This gives the possibility for toroidal Alfvén eigenmodes ͑TAEs͒ to occur and such modes are frequently observed in START neutral beam injection ͑NBI͒ discharges, but seem to be benign. The features of these observed TAEs are shown to be in agreement with MHD calculations.

The computation of magnetic equilibria in the START spherical tokamak is more difficult than those in more conventional large aspect ratio tokamaks. This difficulty arises partly as a result of the use of induction compression to generate high current plasma, as this precludes the positioning of magnetic diagnostics close to the outboard side of the plasma. In addition, the effect of a conducting wall with a high, but finite, conductivity must be included. A method is presented for obtaining plasma equilibrium reconstructions based on the EFIT code. New constraints are used to relate isoflux surface locations deduced from radial profile measurements of electron temperature. A model of flux diffusion through the vessel wall is developed. It is shown that neglecting flux diffusion in the vessel wall can lead to a significant underestimate in the calculation of the plasma βt. Using a relatively sparse set of magnetic signals, βt can be obtained to within a fractional error of ±10%. Using constraints to relate isoflux surface locations, the principle involved in determining the internal q profile is demonstrated.

A method for maintaining toroidal current in toroidal plasmas is discussed. The method requires application of suitably phased oscillating toroidal and poloidal voltages to the plasma resulting in a magnetic field configuration with small oscillations around some mean state. In such quasi-steady states the usual v sec limitation on discharge duration is eliminated. The current drive effect is caused by a nonlinear interaction between the toroidal and poloidal circuits that can be understood in general terms from symmetry considerations. Specific calculations of the effect are made using two models: (1) a zero-dimensional relaxation model, relevant to the reversed-field pinch, and (2) a one-dimensional resistive diffusion model (assuming slab geometry). The results for the relaxation model indicate a useful current drive effect that may be of importance for the reversed-field-pinch program.

No abstract

T i e magnetic fieid profilesin an RFP are calculated on the assumption that the plasma relaxes to a state of minimum Ohmic dissipation (Montgomery and Phillips) subjeci to fixed values of the ulroidal flu and voltage (helicity dissipation). A prescription for the solution of the Euler-Lagxange equation in arbilmy geometry is given using a boundary condition derived from the variational principle rather than assuming zero current density ai the wall. Explicit solutions are found for straight cylinders, first with circular boundaries and then with square boundaries. For the latter the equilibria obtained are not in pressure balance.

The aim is to discuss the effect of the flow pattern on the sensitivity of electromagnetic flowmeters, but not the electronic devices used for interpreting the signal nor the practical details of designing such meters. A more extensive treatment of some of the problems discussed may be found in Reference 1.