LSTM을 이용한 탄천에서의 시간별 하천수위 모의
Hourly Water Level Simulation in Tancheon River Using an LSTM
2
조회
0
다운로드
1
인용
0
참고문헌
키워드
초록
This study was conducted on how to simulate runoff, which was done using existing physical models, using an LSTM (Long Short-Term Memory)model based on deep learning. Tancheon, the first tributary of the Han River, was selected as the target area for the model application. To apply themodel, one water level observatory and four rainfall observatories were selected, and hourly data from 2020 to 2023 were collected to apply the model. River water level of the outlet of the Tancheon basin was simulated by inputting precipitation data from four rainfall observation stations in the basinand average preceding 72-hour precipitation data for each hour. As a result of water level simulation using 2021 to 2023 data for learning and testingwith 2020 data, it was confirmed that reliable simulation results were produced through appropriate learning steps, reaching a certain mean absoluteerror in a short period time. Despite the short data period, it was found that the mean absolute percentage error was 0.5544∼0.6226%, showing anaccuracy of over 99.4%. As a result of comparing the simulated and observed values of the rapidly changing river water level during a specific heavyrain period, the coefficient of determination was found to be 0.9754 and 0.9884. It was determined that the performance of LSTM, which aims tosimulate river water levels, could be improved by including preceding precipitation in the input data and using precipitation data from various rainfallobservation stations within the basin.
Abstract
This study was conducted on how to simulate runoff, which was done using existing physical models, using an LSTM (Long Short-Term Memory)model based on deep learning. Tancheon, the first tributary of the Han River, was selected as the target area for the model application. To apply themodel, one water level observatory and four rainfall observatories were selected, and hourly data from 2020 to 2023 were collected to apply the model. River water level of the outlet of the Tancheon basin was simulated by inputting precipitation data from four rainfall observation stations in the basinand average preceding 72-hour precipitation data for each hour. As a result of water level simulation using 2021 to 2023 data for learning and testingwith 2020 data, it was confirmed that reliable simulation results were produced through appropriate learning steps, reaching a certain mean absoluteerror in a short period time. Despite the short data period, it was found that the mean absolute percentage error was 0.5544∼0.6226%, showing anaccuracy of over 99.4%. As a result of comparing the simulated and observed values of the rapidly changing river water level during a specific heavyrain period, the coefficient of determination was found to be 0.9754 and 0.9884. It was determined that the performance of LSTM, which aims tosimulate river water levels, could be improved by including preceding precipitation in the input data and using precipitation data from various rainfallobservation stations within the basin.
참고문헌 (17)
- Chang, F. J., "Real-time multi-step-ahead water level forecasting by recurrent neural networks for urban flood control", Journal of Hydrology, 517, pp.836~846 (2014)
- Chun, B. S., "Estimation of optimal training period for the deep-learning LSTM model to forecast CMIP5-based streamflow", Journal of the Korean Society of Agricultural Engineers, 64(1), pp.39~50 (2022)
- Coulibaly, P., "Real-time short-term natural water inflows forecasting using recurrent neural networks", Proc. International Joint Conference on IEEE, pp.3802~3805 (1999)
- Dawson, C. W., "An artificial neural network approach to rainfall-runoff modeling", Hydrological Sciences Journal, 43(1), pp.47~66 (1998)
- Hochreiter, S., "Long short-term memory", Neural Computation, 9(8), pp.1735~1780 (1997)
- Huang, X., "Evaluation of short-term streamflow prediction methods in urban river basins", Physics and Chemistry of the Earth, pp.123~ (2021)[DOI]
- Jang, W., "RNN-LSTM based soil moisture estimation using Terra MODIS NDVI and LST", Journal of the Korean Society of Agricultural Engineers, 61(6), pp.123~132 (2019)
- Jeong, Y., "Development o f a s tochas tic s now depth prediction model using a Bayesian deep learning method", Journal of the Korean Society of Agricultural Engineers, 64(6), pp.35~41 (2022)
- Joo, D., "Development of methodology for measuring water level in agricultural water reservoir through deep learning analysis of CCTV images", Journal of the Korean Society of Agricultural Engineers, 65(1), pp.15~26 (2023)
- Jung, J., "Flood stage forecasting at the Gurye-Gyo station in Sumjin River using LSTM-based deep learning models", Journal of the Korean Society of Hazard Mitigation, 21(3), pp.193~201 (2021)
- Jung, S., "Prediction of water level in a tidal river using a deep-learning based LSTM model", Journal of Korea Water Resources Association, 51(12), pp.1207~1216 (2018)
- Kim, H. I., "Applying observed rainfall and deep neural network for urban flood analysis", Journal of the Korean Society of Hazard Mitigation, 20(1), pp.339~350 (2020)
- Kim, K. H., "Application of CCTV image and semantic segmentation model for water level estimation of irrigation channel", Journal of the Korean Society of Agricultural Engineers, 64(3), pp.63~73 (2022)
- Kim, Y. G., "Study of the construction of a coastal disaster prevention system using deep learning", Journal of Ocean Engineering and Technology, 33(6), pp.590~596 (2019)
- Lee, M., "Water level prediction in Taewha River basin using deep learning model based on DNN and LSTM", Journal of Korea Water Resources Association, 54(S-1), pp.1061~1069 (2021)
- Oh, I. S., "Machine learning", pp.417~460 (2017), Hanbit Academy, Inc
- Seo, Y., "Reservoir water level forecasting using machine learning models", Journal of the Korean Society of Agricultural Engineers, 59(3), pp.97~110 (2017)