Search

This research illustrates how time-series forecasting employing recurrent neural networks (RNNs) can be used for anomaly detection in particle accelerators—complex machines that accelerate elementary particles to high speeds for various scientific and industrial applications. Our approach utilizes an RNN to predict temperatures of key components of magnet power supplies (PSs), which can number up to thousands in an accelerator. An anomaly is declared when the predicted temperature deviates significantly from observation. Our method can help identify a PS requiring maintenance before it fails and leads to costly downtime of an entire billion-dollar accelerator facility. We demonstrate that the RNN outperforms a reasonably complex physics-based model at predicting the PS temperatures and at anomaly detection. We conclude that for practical applications it can be beneficial to use RNNs instead of increasing the complexity of the physics-based model. We chose the long short-term memory (LSTM) as opposed to other RNN cell structures due to its widespread use in time-series forecasting and its relative simplicity. However, we demonstrate that the LSTM’s precision of predicting PS temperatures is nearly on par with measurement precision, making more complex or custom architectures unnecessary. Lastly, we dedicate a section of this paper to presenting a proof-of-concept for using infrared cameras for spatially-resolved anomaly detection inside power supplies, which will be a subject of future research.

Generally, turn-to-turn fluctuations of synchrotron radiation power in a storage ring depend on the 6D phase-space distribution of the electron bunch. This effect is related to the interference of fields radiated by different electrons. Changes in the relative electron positions and velocities inside the bunch result in fluctuations in the total emitted energy per pass in a synchrotron radiation source. This effect has been previously described assuming constant and equal electron velocities before entering the synchrotron radiation source. In this paper, we present a generalized formula for the fluctuations with a non-negligible beam divergence. Further, we corroborate this formula in a dedicated experiment with undulator radiation in the Integrable Optics Test Accelerator (IOTA) storage ring at Fermilab. Lastly, possible applications in beam instrumentation are discussed.

In a storage ring, turn-to-turn fluctuations in the intensity of spontaneous synchrotron radiation occur due to two mechanisms. The first mechanism is the quantum uncertainty in the number of emitted photons. The second mechanism is the turn-to-turn variations in the relative positions of classical pointlike electrons in the bunch. We present a unified description of both effects in the framework of quantum optics. We derive an equation for the fluctuations for an arbitrary degree of coherence, which generalizes previously reported results for temporally incoherent radiation. We compare the predictions of our calculation with a previous experiment at Brookhaven National Laboratory, where the latter mechanism was dominant and propose a new dedicated experiment in the Integrable Optics Test Accelerator (IOTA) at Fermilab, where the two mechanisms may have comparable contributions to the fluctuations. Finally, our calculation shows that the magnitude of the fluctuations is rather sensitive to the dimensions and the shape of the electron bunch, thereby indicating possible applications in beam instrumentation. In particular, the small vertical size of the flat beams in IOTA may be estimated via these fluctuations, whereas measurement by a conventional synchrotron radiation monitor is difficult due to the diffraction limit.