Drainage basin flood multi-step forecasting method and device based on converter deep learning
CN118939747AActive Publication Date: 2024-11-12WUHAN UNIV
Patent Information
- Application Number
- CN202411048458.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-01
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-08-01
Smart Images

Figure CN118939747A_ABST
Abstract
The invention discloses a river basin flood multi-step forecasting method and device based on converter deep learning, and the method comprises the steps: building a converter deep learning model based on a self-adaptive path in consideration of the runoff and confluence duration of a river basin, and intelligently matching a multi-step flood forecasting architecture, a dual self-attention mechanism is adopted to capture local and global dependency of flood and rainfall; target function regularization is considered, and converter deep learning model parameters are calibrated by adopting a sub-when optimization algorithm; and an interpretability method is applied to analyze the contribution degree of the input factor to the multi-step flood forecasting precision, so that the physical interpretability of the deep learning model is improved. The method comprises the following steps: constructing a converter deep learning model based on a self-adaptive path, and determining effective input factors of rainfall and flood in different flood forecast periods; a dual self-attention mechanism and model parameter calibration based on a sub-when optimization algorithm are considered, so that the accuracy of multi-step forecasting of the watershed flood is improved, and the flood forecasting period is prolonged.
Need to check novelty before this filing date? Find Prior Art
Citation Information
Patent Citations
BiGRU multi-step prediction method and system applied to flood prediction and storage medium
CN111767517A
Early warning method and device for ocean big data collaborative representation based on knowledge distillation
CN114913400A
Sequence-to-sequence multi-step flood forecasting method and device based on space-time attention
CN115630723A
Wind speed prediction method and system and electronic equipment
CN117494573A
Data processing system and method for data collection management and control
CN117952278A
