Advanced Control Systems for Fast-Orbit Feedback (FOFB) of the Electron Beam in the Diamond Synchrotron
Lead Research Organisation:
University of Oxford
Department Name: Engineering Science
Abstract
The Diamond Light Source (Diamond) is a particle accelerator located at the Harwell Science and Innovation Campus in Oxfordshire. Electrons are propelled near the speed of light and injected into the storage ring. The electron beam is focused and kept in its orbital path using a series of magnets arranged in a repeating pattern around the ring. While the electrons travel around the ring at relativistic speeds, they emit a particular form of radiation called synchrotron light. The emitted radiation is deflected into photon beamlines and guided into several experimental stations where it is used for diffraction, spectroscopy and three-dimensional imaging of molecules.
In order to guarantee the quality of the experiments, it is of utmost importance that the beam follows a stable path. The deviations of the electron beam from its ideal orbit are measured using beam position monitors and must be attenuated using active control of the corrector magnets.
The Fast Orbit Feedback (FOFB) in the Diamond storage ring ensures stability of the beam position for all the experiments on the beamlines. The existing controller is implemented in 24 general purpose Central Processing Units (CPUs) distributed around the ring. In this approach, the Multi-Input-Multi-Output control problem is address using multiplication with a conditioned pseudo inverse orbit response matrix followed by an Internal Model Controller (IMC). The calculations in this case can be distributed to 24 CPUs by slicing the matrix into 24 parts.
This approach is being challenged by existing modifications and future upgrades which may introduce correctors (actuators) with homogeneous dynamics (magnets of different construction or over different vessel material). Furthermore, it would be beneficial to deal with constraints on actuators range or rate, as well as the ability to put particular constrains on certain monitors. These desirable features are beyond the abilities of the current implementation.
This project will investigate a future architecture (both analytically and by implementation) of fast orbit control: more advanced control algorithms like mode-by-mode control and model predictive control, as well as a prototype of the necessary new processing node, which will enable fully centralised calculations.
In summary, the project will establish the methods and an implementation of fast orbit control required to adapt to foreseeable challenges on a Diamond-II machine, while at the same time promising better orbit control already after implementation.
This project falls within the EPSRC Engineering research area and collaborates with Diamond Light Source Ltd.
In order to guarantee the quality of the experiments, it is of utmost importance that the beam follows a stable path. The deviations of the electron beam from its ideal orbit are measured using beam position monitors and must be attenuated using active control of the corrector magnets.
The Fast Orbit Feedback (FOFB) in the Diamond storage ring ensures stability of the beam position for all the experiments on the beamlines. The existing controller is implemented in 24 general purpose Central Processing Units (CPUs) distributed around the ring. In this approach, the Multi-Input-Multi-Output control problem is address using multiplication with a conditioned pseudo inverse orbit response matrix followed by an Internal Model Controller (IMC). The calculations in this case can be distributed to 24 CPUs by slicing the matrix into 24 parts.
This approach is being challenged by existing modifications and future upgrades which may introduce correctors (actuators) with homogeneous dynamics (magnets of different construction or over different vessel material). Furthermore, it would be beneficial to deal with constraints on actuators range or rate, as well as the ability to put particular constrains on certain monitors. These desirable features are beyond the abilities of the current implementation.
This project will investigate a future architecture (both analytically and by implementation) of fast orbit control: more advanced control algorithms like mode-by-mode control and model predictive control, as well as a prototype of the necessary new processing node, which will enable fully centralised calculations.
In summary, the project will establish the methods and an implementation of fast orbit control required to adapt to foreseeable challenges on a Diamond-II machine, while at the same time promising better orbit control already after implementation.
This project falls within the EPSRC Engineering research area and collaborates with Diamond Light Source Ltd.
