Synapse
⌘+K
Synapse
PulseExploreJournal ClubResearchersJournals
Instagram
HomeJournal ClubExplore
March 4, 2026FireOpen Access

Wind Speed Prediction Based on AM-BiLSTM Improved by PSO-VMD for Forest Fire Spread

View Full Paper
Ask AI
Bookmark
Share

Authors

HZHaining ZhuSLShuwen LiuHJHuimin Jia

Discussion

Loading...

Member takes

Overview

This research demonstrates enhanced wind speed prediction accuracy in wildfire scenarios through a novel integrated forecasting method.

Key Points

  • The study aims to improve wind speed forecasting to better simulate wildfire spread dynamics.
  • Utilized variational mode decomposition to process wind speed data.
  • Optimized decomposition parameters using particle swarm optimization.
  • Implemented an attention mechanism in a bidirectional long short-term memory model for wind speed prediction.
  • Developed a multi-branch convolutional neural network for fire spread feature extraction.
  • Achieved high precision and recall values over fire spread prediction experiments.
  • Convergence of the MCNN model loss with optimal predictions at a combustion threshold of 0.7.
  • Demonstrated improvement in evaluation metrics compared to CNN, DCIGN, and DNN models.
  • Stable prediction errors remained below 6% across various experimental scenarios.

Cite This Study

Zhu et al. (2026) studied this question.

synapsesocial.com/papers/69a7cc9fd48f933b5eed84a4https://doi.org/10.3390/fire9030110
View Full Paper
Ask AI
Bookmark
Share