17–22 May 2026
C.I.D
Europe/Zurich timezone

Research on Indirect Measurement Algorithms for Accelerator Tubes Based on Convolutional Neural Networks

TUP7620
19 May 2026, 16:00
2h
C.I.D

C.I.D

Deauville, France
Poster Presentation MC7.T06: Normal Conducting RF Poster session

Speaker

Mr Qingzhu Li (Tsinghua University)

Description

The measurement of accelerator tubes often employs direct probe methods; however, these methods frequently introduce perturbations, leading to inaccurate results, especially at high frequencies. This study presents a novel approach utilizing a Convolutional Neural Network (CNN) combined with a Long Short-Term Memory (LSTM) network to address these challenges. By deeply learning the reflection coefficient curves of cavity chains in microwave networks, our method enables an effective diagnosis of high-frequency accelerator cavities through indirect detuning techniques. The proposed algorithm accurately identifies discrepancies between the actual single-cavity frequency and the design specifications, thereby enhancing the precision of measurements in the high-frequency domain. This research contributes significantly to the field of accelerator tube diagnostics by offering a robust, non-intrusive alternative to traditional direct probing methods.

In which format do you inted to submit your paper? LaTeX

Author

Mr Qingzhu Li (Tsinghua University)

Co-authors

Jiaru Shi (Tsinghua University) Hao Zha (Tsinghua University) Huaibi Chen (Tsinghua University)

Presentation materials

There are no materials yet.