AI Algorithm Enhances Particle Accelerator Performance, Reduces Downtime

Researchers at the SLAC National Accelerator Laboratory have developed an artificial intelligence (AI) algorithm to monitor the performance of the Linac Coherent Light Source, a mile-long particle accelerator. The AI system alerts operators when performance dips and identifies the specific subsystem causing the issue. This technology could improve reliability in other complex systems such as advanced manufacturing plants, the electric grid, and nuclear power plants. The research was funded by the Department of Energy and Stanford University. Daniel Ratner from SLAC National Accelerator Laboratory can be contacted for more information.

AI Algorithm for Particle Accelerator Health Monitoring

Particle accelerators are complex scientific instruments with millions of sensors and thousands of subsystems. At the Linac Coherent Light Source, a Department of Energy user facility at SLAC National Accelerator Laboratory, human operators have to continuously monitor performance and sift through a multitude of sensors to identify problems. To simplify this process, researchers have developed an artificial intelligence (AI) algorithm that mimics how human operators approach this challenge. The AI system monitors the accelerator, alerts operators when performance dips, and identifies the specific subsystem that is responsible. This can reduce downtime and enhance the scientific data collected.

The Impact of the AI Solution

The AI solution provides SLAC operators with information on which components should be switched off and replaced to keep the accelerator running continuously. Improved reliability also keeps more subsystems online, allowing the accelerator to reach its full operating capability. This AI approach could benefit many complex systems, such as other experimental facilities, advanced manufacturing plants, the electric grid, and nuclear power plants, by improving their reliability.

Modern Accelerators and the Challenge of Data Monitoring

Modern accelerators record millions of data streams, which is too many signals for a small operations team to monitor in real time and reliably avoid subsystem faults that lead to costly downtime. For instance, in the Linac Coherent Light Source, faults in the radiofrequency (RF) stations that accelerate the electrons are a primary cause of downtime and drops in performance. An existing automated algorithm tries to identify RF station problems, but almost 70% of the algorithm’s predictions are false positives, leading operators to resort to manual inspection to identify RF station anomalies.

The AI Method and Its Advantages

Inspired by the operators, the AI method simultaneously runs anomaly detection algorithms on both the RF station diagnostics and shot-to-shot measurements of the final beam quality. A fault is predicted only when both algorithms simultaneously identify anomalies. This approach can be entirely automated and identifies more events with fewer false positives than the RF station diagnostics alone. Recent patent-pending work has extended the coincidence concept to deep-learning algorithms, such as neural networks, which can identify faults on raw, unlabeled data without expert input. Researchers expect these machine learning-driven algorithms to have broad applications in complex systems across science and industry.

Contact and Funding Information

The research was led by Daniel Ratner from the SLAC National Accelerator Laboratory. Funding for this research was provided by the Department of Energy (DOE) Office of Science, Basic Energy Sciences, Scientific User Facilities Division, and Stanford University. The research used resources at the Linac Coherent Light Source, a DOE Office of Science user facility

Summary

“Researchers have developed an artificial intelligence (AI) algorithm that monitors the performance of particle accelerators, alerting operators to performance issues and identifying the specific subsystem at fault. This AI approach could improve reliability in complex systems such as experimental facilities, advanced manufacturing plants, and nuclear power plants.”

• Researchers at the SLAC National Accelerator Laboratory have developed an artificial intelligence (AI) algorithm to monitor the performance of the Linac Coherent Light Source, a particle accelerator.
• The AI system alerts operators when performance dips and identifies the specific subsystem causing the issue, reducing downtime and improving the quality of scientific data collected.
• The AI solution also advises operators on which components need to be replaced to keep the accelerator running continuously.
• The AI approach could potentially improve reliability in other complex systems such as other experimental facilities, advanced manufacturing plants, the electric grid, and nuclear power plants.
• The AI method runs anomaly detection algorithms on both the radiofrequency (RF) station diagnostics and measurements of the final beam quality, predicting a fault only when both algorithms identify anomalies.
• This approach has been incorporated into the control room and can be entirely automated. It identifies more events with fewer false positives than the RF station diagnostics alone.
• The research was funded by the Department of Energy (DOE) Office of Science, Basic Energy Sciences, Scientific User Facilities Division, and Stanford University.
• Daniel Ratner is the contact person for this research at the SLAC National Accelerator Laboratory.