HTPD 2018

Sensors that measure small 3D magnetic fields in tokamaks are susceptible to both DC and AC vacuum field pickup that must be compensated out. In this paper, we present a novel sensor compensation algorithm that uses a layered low-pass filtering technique to efficiently remove the vacuum field pickup generated by both axisymmetric and non-axisymmetric coils. Given that a single technique is used to compensate the pickup from all coil systems and that the low-pass filtering is conducted in the time domain, the layered low-pass algorithm is uniquely suited for real-time processing. The algorithm was first deployed on the National Spherical Torus Experiment Upgrade (NSTX-U) during the 2016 commissioning campaign where it was successfully used to perform both offline and real-time locked mode identification. The offline version of the algorithm has since been tested on the DIII-D tokamak where high-spectral-content training shots are shown to improve the compensation of poloidal-field-coil-induced pickup. The cross-device portability of the algorithm demonstrates that it can broadly address the challenge of real-time magnetic sensor compensations in tokamaks. This work is supported by DoE contracts DE-AC02-09CH11466 and DE-FC02-04ER54698.