Reservoir Flood Forecasting Based on Long-Short-Term Memory Neural Network
Accurate flood forecasting is one of the main means to well perform flood control and drainage,and the long-short-term memory neural network (LSTM) has a strong ability to fit time series relationships,which thus is very suitable for simulating and forecasting the complex time series process of basi...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | zho |
| Published: |
Editorial Office of Pearl River
2022-01-01
|
| Series: | Renmin Zhujiang |
| Subjects: | |
| Online Access: | http://www.renminzhujiang.cn/thesisDetails#10.3969/j.issn.1001-9235.2022.12.018 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Accurate flood forecasting is one of the main means to well perform flood control and drainage,and the long-short-term memory neural network (LSTM) has a strong ability to fit time series relationships,which thus is very suitable for simulating and forecasting the complex time series process of basin runoff generation and confluence.To explore the applicability of LSTM in the field of reservoir flood forecasting,this paper established an LSTM model according to different forecast periods in the Baipenzhu Basin and compared it with Xinanjiang model.The LSTM model uses the rainfall and water level data in the basin as input and adopts the water levels of the reservoir at different forecast periods as output.The calibration period is five years,and the verification period is one year.The results show that LSTM has high forecast accuracy when the forecast period is 1~6 h,and the forecast accuracy is the highest when the forecast period is 1h,reaching 0.991.As the forecast period increases,the accuracy of the LSTM model gradually decreases,but its forecast accuracy is higher than that of Xinanjiang model.In addition,reflecting the complexity of the neural network,the prediction period and the number of neurons in the hidden layer will affect not only the forecast accuracy but also the training speed of the model.It is proven that the LSTM model has high forecast accuracy and is of guiding significance to reservoir flood forecasting. |
|---|---|
| ISSN: | 1001-9235 |