Thus, it appears that for the present accelerator the decelerating eddy currents are small in comparison to the total transverse accelerating current of nearly 500 amp, and no large velocity decreases in the end regions would be expected. Because eddy currents also may decelerate the flow by heating it, several attempts were made to determine their effect on the plasma velocity at the accelerator entrance and exit. In each instance there was no indication that the velocity was changed significantly. Unfortunately, it was not possible to measure the impact pressure at the exit of the acceleration section. However, a careful accelerator performance analysis by Wilson (of the type described in Ref. 7) indicated that the impact pressure measured just aft of the channel exit (3 cm aft of electrode pair 117) was reasonably well predicted by a modified one-dimensional MHD channelflow theory, which neglects the possible existence of eddy currents. This, together with the previously noted constancy of impact pressures during the open circuit tests, suggests that eddy currents did not cause a significant velocity decrease in the downstream unpowered section. In addition, wall static pressures, which were measured throughout the length of the channel, gave no evidence of velocity changes in either of the unpowered sections. Thus, eddy currents, which served to decelerate the flow, were present, but to a much lesser extent than was reported by Leonard and Fay. More generally, it may be concluded from the present tests that accelerator exit-velocity losses may be minimized by keeping the exit magnetic field gradients small as has been pointed out by Leonard and Fay. Further, one could attempt to quench the plasma before allowing it to pass through a strong magnetic field gradient if such gradients are unavoidable.
Read full abstract