Engineering topics

Seiya, K.

Publications and source records attributed to Seiya, K..

5 records

DOE OSTI2023

Beyond PID Controllers: PPO with Neuralized PID Policy for Proton Beam Intensity Control in Mu2e

We introduce a novel Proximal Policy Optimization (PPO) algorithm aimed at addressing the challenge of maintaining a uniform proton beam intensity delivery in the Muon to Electron Conversion Experiment (Mu2e) at Fermi National Accelerator Laboratory (Fermilab). Our primary objective is to regulate the spill process to ensure a consistent intensity profile, with the ultimate goal of creating an automated controller capable of providing real-time feedback and calibration of the Spill Regulation System (SRS) parameters on a millisecond timescale. We treat the Mu2e accelerator system as a Markov Decision Process suitable for Reinforcement Learning (RL), utilizing PPO to reduce bias and enhance training stability. A key innovation in our approach is the integration of a neuralized Proportional-Integral-Derivative (PID) controller into the policy function, resulting in a significant improvement in the Spill Duty Factor (SDF) by 13.6%, surpassing the performance of the current PID controller baseline by an additional 1.6%. This paper presents the preliminary offline results based on a differentiable simulator of the Mu2e accelerator. It paves the groundwork for real-time implementations and applications, representing a crucial step towards automated proton beam intensity control for the Mu2e experiment.

43 PARTICLE ACCELERATORS↗

DOE OSTI2023

ML-based Real-Time Control at the Edge: An Approach Using hls4ml

This study focuses on implementing a real-time control system for a particle accelerator facility that performs high energy physics experiments. A critical operating parameter in this facility is beam loss, which is the fraction of particles deviating from the accelerated proton beam into a cascade of secondary particles. Accelerators employ a large number of sensors to monitor beam loss. The data from these sensors is monitored by human operators who predict the relative contribution of different sub-systems to the beam loss. Using this information, they engage control interventions. In this paper, we present a controller to track this phenomenon in real-time using edge-Machine Learning (ML) and support control with low latency and high accuracy. We implemented this system on an Intel Arria 10 SoC. Optimizations at the algorithm, high-level synthesis, and interface levels to improve latency and resource usage are presented. Our design implements a neural network, which can predict the main source of beam loss (between two possible causes) at speeds up to 575 frames per second (fps) (average latency of 1.74 ms). The practical deployed system is required to operate at 320 fps, with a 3ms latency requirement, which has been met by our design successfully.

Shi, R.↗

DOE OSTI2023

FPGA Architectures for Distributed ML Systems for Real-time Beam Loss De-blending

The Accelerator Real-time Edge AI for Distributed Systems (READS) project’s goal is to create a Artificial Intelligence (AI) system for real-time beam loss de-blending within the accelerator enclosure, which houses two accelerators: the Main Injector (MI) and the Recycler Ring (RR). In periods of joint operation, when both machines contain high intensity beam, radioactive beam losses from MI and RR overlap on the enclosure’s beam loss monitoring Beam Loss Monitor (BLM) system, making it difficult to attribute those losses to a single machine. Incorrect diagnoses result in unnecessary downtime that incurs both financialand experimental cost. The ML system will automatically disentangle each machine’s contributions to those measured losses, while not disrupting the existing operations-critical functions of the BLM system. This paper will focus on the evolution of the architectures, which provided the high-frequency, low-latency collection of synchronized data streams to make real-time inferences. The ML models, used for learning both local and global machine signatures and producing high quality inferences based on raw BLM loss measurements, will only be discussed at a high-level.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2023

Time-drift Aware RF Optimization with Machine Learning Techniques

The Fermilab Linac delivers 400 MeV H- beam to the rest of the accelerator chain. Providing stable intensity, energy, and emittance is key since it directly affects downstream machines. To operate high current beam, accelerators must minimize uncontrolled particle loss; this can be accomplished by minimizing beam longitudinal emittance via RF parameter optimization. However, RF tuning is required daily since the resonance frequency of the accelerating cavities is affected by ambient temperature and humidity variations and thus drifts with time. In addition, the energy and phase space distribution of particles emerging from the ion source are subject to fluctuations. Such drift is not unique to Fermilab, but rather affects most laboratories. We are exploring machine learning (ML) algorithms for automated RF tuning for 2 objectives: optimization of Linac output energy and phase oscillation correction, with an emphasis on time-drift aware modeling that can account for conditions changing over time.

43 PARTICLE ACCELERATORS↗

DOE OSTI2023

Beam Loss Monitoring with Fixed and Translating Scintillation Detectors Along the Fermilab Drift-tube Linac

The Fermilab Linac is a roughly 145 meter linear accelerator that accelerates H- beam from 750 keV to 400 MeV and provides beam for the Booster and the rest of the accelerator chain. The first section of the Linac is a Drift-Tube Linac (DTL), which in its current state, suffers from a lack of instrumentation along its length. As a result, operational staff do not have access to the diagnostic information needed to tune the critical components of this accelerator, such as the quadrupole magnets within the drift tubes. This work presents an effort to utilize both fixed and translating scintillation detectors to investigate beam loss along the first two cavities of the DTL.

43 PARTICLE ACCELERATORS↗