arXiv subjects

Jinyu Wan

Publications and source records attributed to Jinyu Wan.

8 recordsLinked to original sources

arXiv2026

Square matrix-based six-dimensional convergence map for nonlinear beam dynamics analysis

The square matrix-based convergence map (CM) method has proven effective in characterizing nonlinear dynamics in several 4-D dynamical systems. However, when time-dependent perturbations, such as crabbing kicks in colliders, are present, a comprehensive 6-D analysis becomes essential to accurately capture the coupling between transverse and longitudinal motions. In this work, we extend the CM method to the full 6-D phase space by employing an eigen-decomposition-based formulation of the square matrix combined with iterative procedures. The proposed 6-D CM approach is first validated using a simplified crabbing map. We demonstrate that the 6-D CM preserves computational efficiency by using only one-turn map, while successfully resolving high-order resonance structures that remain unresolved by conventional frequency map analysis (FMA). This method is subsequently applied to the dynamic aperture (DA) study of the future Electron-Ion Collider (EIC). The results obtained from the CM analysis exhibit close agreement with those derived from FMA, demonstrating its potential as a powerful tool for nonlinear beam dynamics analysis and DA evaluation, as well as for broader applications in other nonlinear dynamical systems.

physics.acc-ph↗

arXiv2025

A multi-language auto-differentiation module and its application to a parallel particle-in-cell code on distributed computers

The auto differentiable simulation is a type of simulation that outputs of the simulation include not only the simulation result itself, but also their derivatives with respect to various input parameters. It provides an efficient method to study sensitivity of the simulation results with respect to the input parameters. Furthermore, it can be used in gradient based optimization methods for rapidly optimizing design parameters. In this paper, we present the development of a fast and transparent auto-differentiation module/class designed for easy integration into numerous simulation codes. As an application, this auto-differentiation module is integrated into a parallel particle-in-cell code with message passing interface (MPI) on distributed memory computers.

physics.comp-ph↗

arXiv2025

Fast chaos indicator from auto-differentiation for dynamic aperture optimization

Automatic differentiation provides an efficient means of computing derivatives of complex functions with machine precision, thereby enabling differentiable simulation. In this work, we propose the use of the norm of the tangent map, obtained from differentiable tracking of particle trajectories, as a computationally efficient indicator of chaotic behavior in phase space. In many cases, a one-turn or few-turn tangent map is sufficient for this purpose, significantly reducing the computational cost associated with dynamic aperture optimization. As an illustrative application, the proposed indicator is employed in the dynamic aperture optimization of an ALS-U lattice design.

physics.acc-ph↗

arXiv2024

JuTrack: a Julia package for auto-differentiable accelerator modeling and particle tracking

Efficient accelerator modeling and particle tracking are key for the design and configuration of modern particle accelerators. In this work, we present JuTrack, a nested accelerator modeling package developed in the Julia programming language and enhanced with compiler-level automatic differentiation (AD). With the aid of AD, JuTrack enables rapid derivative calculations in accelerator modeling, facilitating sensitivity analyses and optimization tasks. We demonstrate the effectiveness of AD-derived derivatives through several practical applications, including sensitivity analysis of space-charge-induced emittance growth, nonlinear beam dynamics analysis for a synchrotron light source, and lattice parameter tuning of the future Electron-Ion Collider (EIC). Through the incorporation of automatic differentiation, this package opens up new possibilities for accelerator physicists in beam physics studies and accelerator design optimization.

physics.acc-ph↗

arXiv2024

Time-Delayed Koopman Network-Based Model Predictive Control for the FRIB RFQ

The radio-frequency quadrupole (RFQ) at the Facility for Rare Isotope Beams (FRIB) is a critical device to accelerate heavy ion beams from 12 keV/u to 0.5 MeV/u for state-of-the-art nuclear physics experiments. Efficient control of the RFQ resonance frequency detuning still remains a challenge because the temperature-sensitive frequency is solely control by a cooling water system, exhibiting complicated transport delay and nonlinearity in the heat transfer processes. In this work, we propose a long-short term memory (LSTM)-based Koopman network model that can simultaneously learn the time-delayed and non-delayed correlations hidden in the historical operating data. It is proven that the model can effectively predict the behavior of the RFQ resonance frequency using historical data as inputs. With this model, a model predictive control (MPC) framework based on the Newton-Raphson method is proposed and tested. We demonstrate that the MPC framework utilizing deep learning model is able to provide precise and rapid control for the RFQ frequency detuning, reducing the control time by half compared to the proportional-integral-derivative (PID) controller.

physics.acc-ph↗

arXiv2022

Machine learning-based direct solver for one-to-many problems on temporal shaping of relativistic electron beams

To control the temporal profile of a relativistic electron beam to meet requirements of various advanced scientific applications like free-electron-laser and plasma wakefield acceleration, a widely-used technique is to manipulate the dispersion terms which turns out to be one-to-many problems. Due to their intrinsic one-to-many property, current popular stochastic optimization approaches on temporal shaping may face the problems of long computing time or sometimes suggesting only one solution. Here we propose a real-time solver for one-to-many problems of temporal shaping, with the aid of a semi-supervised machine learning method, the conditional generative adversarial network (CGAN). We demonstrate that the CGAN solver can learn the one-to-many dynamics and is able to accurately and quickly predict the required dispersion terms for different custom temporal profiles. This machine learning-based solver is expected to have the potential for wide applications to one-to-many problems in other scientific fields.

physics.acc-ph↗

arXiv2021

Data-driven Chaos Indicator for Nonlinear Dynamics and Applications on Storage Ring Lattice Design

A data-driven chaos indicator concept is introduced to characterize the degree of chaos for nonlinear dynamical systems. The indicator is represented by the prediction accuracy of surrogate models established purely from data. It provides a metric for the predictability of nonlinear motions in a given system. When using the indicator to implement a tune-scan for a quadratic Henon map, the main resonances and their asymmetric stop-band widths can be identified. When applied to particle transportation in a storage ring, as particle motion becomes more chaotic, its surrogate model prediction accuracy decreases correspondingly. Therefore, the prediction accuracy, acting as a chaos indicator, can be used directly as the objective for nonlinear beam dynamics optimization. This method provides a different perspective on nonlinear beam dynamics and an efficient method for nonlinear lattice optimization. Applications in dynamic aperture optimization are demonstrated as real world examples.

physics.acc-ph↗

arXiv2021

Machine learning enabled fast evaluation of dynamic aperture for storage ring accelerators

For any storage ring-based large-scale scientific facility, one of the most important performance parameters is the dynamic aperture (DA), which measures the motion stability of charged particles in a global manner. To date, long-term tracking-based simulation is regarded as the most reliable method to calculate DA. However, numerical tracking may become a significant issue, especially when lots of candidate designs of a storage ring need to be evaluated. In this paper, we present a novel machine learning-based method, which can reduce the computation cost of DA tracking by approximately one order of magnitude, while keeping sufficiently high evaluation accuracy. Moreover, we demonstrate that this method is independent of concrete physical models of a storage ring. This method has the potential to be applied to similar problems of identifying irregular motions in other complex dynamical systems.

physics.comp-ph↗