Minimization of Betatron Oscillations of Electron Beam Injected Into a Time-Varying Lattice via Extremum Seeking

We report on a beam-based experiment performed at the SPEAR3 storage ring of the Stanford Synchrotron Radiation Lightsource at the SLAC National Accelerator Laboratory, in which a model-independent extremum-seeking optimization algorithm was utilized to minimize betatron oscillations in the presence of a time-varying kicker magnetic field, by automatically tuning the pulsewidth, voltage, and delay of two other kicker magnets, and the current of two skew quadrupole magnets, simultaneously, in order to optimize injection kick matching. Adaptive tuning was performed on eight parameters simultaneously. The scheme was able to continuously maintain the match of a five-magnet lattice while the field strength of a kicker magnet was continuously varied at a rate much higher (±6% sinusoidal voltage change over 1.5 h) than typically experienced in operation. The ability to quickly tune or compensate for time variation of coupled components, as demonstrated here, is very important for the more general, more difficult problem of global accelerator tuning to quickly switch between various experimental setups.