Enhancement of the ASDEX Upgrade Real-Time Plasma Position Reflectometry Diagnostic

Abstract

Microwave O-mode reflectometry is a diagnostic technique that will play an important supplementary role for plasma position control for ITER and forseeably for DEMO. Density profiles from reflectometry will provide, at high temporal resolutions, estimates of the gap between the plasma magnetic separatrix and the tokamak vessel walls. These estimates will be used to detect and correct drifts in the magnetic gap measurements, the primary measurements used for plasma position and shape control. The feasibility of this alternative feedback control approach was demonstrated in 2011 on ASDEX Upgrade (AUG) , where the reflectometry gap estimate actually replaced the corresponding magnetic measurement in the position control loop. Presently, the AUG’s real-time (RT) reflectometry diagnostic is being upgraded to improve not only its density range coverage but also its acquisition and RT data processing performance. The diagnostic is now capable of acquiring a total of 16 channels (previously 8) from which 8, corresponding to microwave bands K, Ka, Q and V from both the high (HFS) and low field side (LFS) reflectometers, are used in the RT density profile and separatrix gap calculations. The modern NUMA hardware architecture of the updated data processing server allows for an efficient and separate handling of the data-flows produced by the hosted acquisition systems. The higher RAM and CPU interconnection bandwidths allow the implementation of new operation modes that exploit the very high data throughput ( $> 1.2~\hbox{GB/s}$ ) of each of the two used acquisition boards. RT reconstruction of the density profiles is a complex algorithm, whose performance will also be improved by the additional processing power (16 cores instead of 8). The system was specified to acquire and store HFS and LFS density profile data every $ whilst simultaneously producing profile measurements for RT control every 1 ms. In this new operation mode, identical to the one planned for the ITER plasma position reflectometer (PPR), the diagnostic can reach an inbound data throughput (ADCs to RT processing host) and a computational load that are closer to the ones expected to be generated by each of the planned 4 ITER PPRs ( $ \approx 3 ~\hbox{GB/s}$ maximum inbound and outbound data stream bandwidth per PPR). Herein we discuss the enhancements introduced in the AUG’s RT reflectometry diagnostic to implement the new operation mode and perform the aimed control experiments. Preliminary RT experimental data obtained in both the HFS and LFS is shown to illustrate the system’s plasma position and shape control capabilities to be demonstrated during AUG’s 2014 experimental campaign.