Airport runway icing risk early warning method fused with neural network
By integrating neural networks and attention mechanisms, combined with improved Jinzhai optimization algorithm and time convolution network, an efficient airport runway ice-covered risk warning model was built, solving the problems of insufficient prediction accuracy and limited feature extraction capabilities in the existing technology, and achieving more accurate and real-time risk warning.
Patent Information
- Application Number
- CN202411727086.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-28
- Publication Date
- 2025-04-04
- Estimated Expiration
- 2044-11-28
AI Technical Summary
When predicting the thickness of the airport runway ice, the prior art has problems such as insufficient prediction accuracy, limited feature extraction capability, low hyperparameter selection efficiency, lack of dynamic attention capability, and insufficient real-time and adaptability.
The fusion neural network method is adopted, combining attention mechanism, improved Jinjiao Optimization Algorithm (IGJO), time convolution network (TCN) and bidirectional gated cyclic unit (BiGRU), to construct an early warning model for ice-covered airport runways. This model improves prediction accuracy and real-time response performance by optimizing hyperparameters, dynamic feature empowerment, and multi-source data fusion.
It significantly improves the accuracy and real-time warning of airport runway ice-covering risk, can more accurately capture the multi-scale changes in meteorological data, enhances sensitivity to extreme weather conditions and sudden risks, and reduces aviation safety hazards.