Scholarlink

This study enhances underground mine temperature prediction using Long Short-Term Memory (LSTM) neural networks, achieving a 34% improvement over traditional methods. Extensive training with ample temperature datasets shows the LSTM model's robust real-world performance. To tackle underground environment challenges, an Embedded System RK3588 is introduced, optimizing and extending performance for scenarios with only temperature data. The integration of RK3588 with the LSTM model enables continuous, accurate, and robust temperature prediction, improving adaptability to complex conditions and supporting comprehensive safety assessments. This synergy enhances real-time inference, significantly advancing mine safety management by improving fire prediction and prevention.

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

Underground Temperature Prediction Based on LSTM Neural Network and Embedded System Reasoning

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Jie Li,
  • Mei Wang,
  • ZhiBo Gong,
  • LiZhi Li

摘要

This study enhances underground mine temperature prediction using Long Short-Term Memory (LSTM) neural networks, achieving a 34% improvement over traditional methods. Extensive training with ample temperature datasets shows the LSTM model's robust real-world performance. To tackle underground environment challenges, an Embedded System RK3588 is introduced, optimizing and extending performance for scenarios with only temperature data. The integration of RK3588 with the LSTM model enables continuous, accurate, and robust temperature prediction, improving adaptability to complex conditions and supporting comprehensive safety assessments. This synergy enhances real-time inference, significantly advancing mine safety management by improving fire prediction and prevention.