Scholarlink

This paper presents a novel LSTM-based model for autonomous temperature control in superconducting linear accelerators (linacs). Traditional PID controllers struggle to maintain stable temperature due to unpredictable external factors. Our approach addresses this challenge by developing a recurrent neural network trained on data from a Python-programmed TCLab PID framework. An in-house embedded HTTP server facilitates remote control of heaters based on LSTM predictions. The system achieved a temperature control accuracy of ± 1% over extended periods, demonstrating the effectiveness of our model. Successful implementation in the cryogenic control network of the Inter University Accelerator Centre, New Delhi, India, validates the practical application of this technology.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An LSTM Based Temperature Control

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Joby Antony,
  • Rajesh Nirdoshi

摘要

This paper presents a novel LSTM-based model for autonomous temperature control in superconducting linear accelerators (linacs). Traditional PID controllers struggle to maintain stable temperature due to unpredictable external factors. Our approach addresses this challenge by developing a recurrent neural network trained on data from a Python-programmed TCLab PID framework. An in-house embedded HTTP server facilitates remote control of heaters based on LSTM predictions. The system achieved a temperature control accuracy of ± 1% over extended periods, demonstrating the effectiveness of our model. Successful implementation in the cryogenic control network of the Inter University Accelerator Centre, New Delhi, India, validates the practical application of this technology.