DOI QR코드

DOI QR Code

Analyzing the Impact of Multivariate Inputs on Deep Learning-Based Reservoir Level Prediction and Approaches for Mid to Long-Term Forecasting

다변량 입력이 딥러닝 기반 저수율 예측에 미치는 영향 분석과 중장기 예측 방안

  • 박혜승 (협성대학교 소프트웨어공학과) ;
  • 윤종욱 (협성대학교 경영학과) ;
  • 이호준 ((주)수리이엔씨 SI사업부) ;
  • 양현호 (국립군산대학교 컴퓨터소프트웨어학부)
  • Received : 2024.02.19
  • Accepted : 2024.03.22
  • Published : 2024.04.30

Abstract

Local reservoirs are crucial sources for agricultural water supply, necessitating stable water level management to prepare for extreme climate conditions such as droughts. Water level prediction is significantly influenced by local climate characteristics, such as localized rainfall, as well as seasonal factors including cropping times, making it essential to understand the correlation between input and output data as much as selecting an appropriate prediction model. In this study, extensive multivariate data from over 400 reservoirs in Jeollabuk-do from 1991 to 2022 was utilized to train and validate a water level prediction model that comprehensively reflects the complex hydrological and climatological environmental factors of each reservoir, and to analyze the impact of each input feature on the prediction performance of water levels. Instead of focusing on improvements in water level performance through neural network structures, the study adopts a basic Feedforward Neural Network composed of fully connected layers, batch normalization, dropout, and activation functions, focusing on the correlation between multivariate input data and prediction performance. Additionally, most existing studies only present short-term prediction performance on a daily basis, which is not suitable for practical environments that require medium to long-term predictions, such as 10 days or a month. Therefore, this study measured the water level prediction performance up to one month ahead through a recursive method that uses daily prediction values as the next input. The experiment identified performance changes according to the prediction period and analyzed the impact of each input feature on the overall performance based on an Ablation study.

지역 저수지들은 농업용수 공급의 중요한 수원공으로 가뭄과 같은 극단적 기후 조건을 대비하여 안정적인 저수율 관리가 필수적이다. 저수율 예측은 국지적 강우와 같은 지역적 기후 특성뿐만 아니라 작부시기를 포함하는 계절적 요인 등에 크게 영향을 받기 때문에 적절한 예측 모델을 선정하는 것만큼 입/출력 데이터 간 상관관계 파악이 무엇보다 중요하다. 이에 본 연구에서는 1991년부터 2022년까지의 전라북도 400여 개 저수지의 광범위한 다변량 데이터를 활용하여 각 저수지의 복잡한 수문학·기후학적 환경요인을 포괄적으로 반영한 저수율 예측 모델을 학습 및 검증하고, 각 입력 특성이 저수율 예측 성능에 미치는 영향력을 분석하고자 한다. 신경망 구조에 따른 저수율 예측 성능 개선이 아닌 다변량의 입력 데이터와 예측 성능 간의 상관관계에 초점을 맞추기 위하여 실험에 사용된 예측 모델로 합성곱신경망 또는 순환신경망과 같은 복잡한 형태가 아닌 완전연결계층, 배치정규화, 드롭아웃, 활성화 함수 등의 조합으로 구성된 기본적인 순방향 신경망을 채택하였다. 추가적으로 대부분의 기존 연구에서는 하루 단위의 단기 예측 성능만을 제시하고 있으며 이러한 단기 예측 방식은 10일, 한 달 단위 등 중장기적 예측이 필요한 실무환경에 적합하지 않기 때문에, 본 연구에서는 하루 단위 예측값을 다음 입력으로 사용하는 재귀적 방식을 통해 최대 한 달 뒤 저수율 예측 성능을 측정하였다. 실험을 통해 예측 기간에 따른 성능 변화 양상을 파악하였으며, Ablation study를 바탕으로 예측 모델의 각 입력 특성이 전체 성능에 끼치는 영향을 분석하였다.

Keywords

Acknowledgement

이 논문은 한국국토정보공사 공간정보연구원 산학협력R&D사업의 지원을 받아 수행된 연구임(과제명: 공간정보기반 인공지능분석을 활용한 농업용저수지의 가뭄대비 저수율 예측, 과제번호: 2023-501).

References

  1. H. Lim, H. An, G. Choi, J. Lee, and J. Do, "Prediction of pollution loads in agricultural reservoirs using LSTM algorithm: Case study of reservoirs in Nonsan City," Korean Journal of Agricultural Science, Vol.49, No.2, pp.193-202, 2022. https://doi.org/10.7744/KJOAS.20220016
  2. J. Jung, S. Kim, D. Kim, and S. Yang, "A study on the use of geospatial information-based simulation for preemptive response to water disasters in agricultural land," Smart Media Journal, Vol.11, No.7, pp.52-60, 2022. https://doi.org/10.30693/SMJ.2022.11.7.52
  3. Y. Seo, E. Choi, and W. Yeo, "Reservoir water level forecasting using machine learning models," Journal of the Korean Society of Agricultural Engineers, Vol.59, No.3, pp.97-110, 2017. https://doi.org/10.5389/KSAE.2017.59.3.097
  4. S. Jung, D. Lee, and K. Lee, "Prediction of river water level using deep-learning open library," Journal of the Korean Society of Hazard Mitigation, Vol.18, No.1, pp.1-11, 2018. https://doi.org/10.9798/KOSHAM.2018.18.1.1
  5. H. Sak, A. W. Senior, and F. Beaufays, "Long short-term memory recurrent neural network architectures for large scale acoustic modeling," 2014.
  6. H. Han, C. Choi, J. Jung, and H. Kim, "Application of sequence to sequence learning based LSTM model (LSTMs2s) for forecasting dam inflow," Journal of Korea Water Resources Association, Vol.54, No.3, pp.157-166, 2021.
  7. M. Yang, W. Nam, H. Kim, T. Kim, A. Shin, and M. Kang, "Anomaly detection in reservoir water level data using the LSTM model based on deep learning," Journal of the Korean Society of Hazard Mitigation, Vol.21, No.1, pp.71-81, 2021. https://doi.org/10.9798/KOSHAM.2021.21.1.71
  8. Y. Seong, K. Park, and Y. Jung, "Flow rate prediction at Paldang Bridge using deep learning models," Journal of Korea Water Resources Association, Vol.55, No.8, pp.565-575, 2022.
  9. K. Cho, B. V. Merrienboer, C. Gulcehre, D.Bahdanau, F. Bougares, H. Schwenk, and Y. Bengio, "Learning phrase representations using RNN encoder-decoder for statistical machine translation," arXiv preprint arXiv:1406.1078, 2014.
  10. Joh. Sunguk, and Lee, Yangwon. "Prediction of Water Storage Rate for Agricultural Reservoirs Using Univariate and Multivariate LSTM Models," Korean Journal of Remote Sensing 39.5 (2023): 1125-1134.
  11. M. Das, S. K. Ghosh, V. M. Chowdary, A.Saikrishnaveni, and R. K. Sharma, "A probabilistic nonlinear model for forecasting daily water level in reservoir," Water Resources Management, Vol.30, pp.3107-3122, 2016.
  12. RAWRIS [Internet], https://rawris.ekr.or.kr
  13. KMA Weather Data Service [Internet], https://data.kma.go.kr
  14. H. L. Penman, "Natural evaporation from open water, bare soil and grass," Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences, Vol.193, No.1032, pp.120-145, 1948.
  15. H. Salehinejad, S. Sankar, J. Barfett, E.Colak, and S. Valaee, "Recent advances in recurrent neural networks," arXiv preprint arXiv:1801.01078, 2017.
  16. D. P. Kingma and J. Ba, "Adam: A method for stochastic optimization," arXiv preprint arXiv:1412.6980, 2014.
  17. A. Paszke et al., "Pytorch: An imperative style, highperformance deep learning library," Advances in Neural Information Processing Systems, Vol.32, 2019.