An interpretable hybrid deep learning model for flood forecasting based on Transformer and LSTM

Journal of Hydrology Regional Studies, Journal Year: 2024, Volume and Issue: 54, P. 101873 - 101873

Published: June 27, 2024

Language: Английский

---

Understanding GANs: fundamentals, variants, training challenges, applications, and open problems

Multimedia Tools and Applications, Journal Year: 2024, Volume and Issue: unknown

Published: May 14, 2024

Language: Английский

---

Short-term stochastic multi-objective optimization scheduling of wind-solar-hydro hybrid system considering source-load uncertainties

Applied Energy, Journal Year: 2024, Volume and Issue: 372, P. 123781 - 123781

Published: June 27, 2024

Language: Английский

---

Research Progress and Prospects of Urban Flooding Simulation: From Traditional Numerical Models to Deep Learning Approaches

Environmental Modelling & Software, Journal Year: 2024, Volume and Issue: unknown, P. 106213 - 106213

Published: Sept. 1, 2024

Language: Английский

---

Applying Machine Learning Methods to Improve Rainfall–Runoff Modeling in Subtropical River Basins

Water, Journal Year: 2024, Volume and Issue: 16(15), P. 2199 - 2199

Published: Aug. 2, 2024

Machine learning models’ performance in simulating monthly rainfall–runoff subtropical regions has not been sufficiently investigated. In this study, we evaluate the of six widely used machine models, including Long Short-Term Memory Networks (LSTMs), Support Vector Machines (SVMs), Gaussian Process Regression (GPR), LASSO (LR), Extreme Gradient Boosting (XGB), and Light (LGBM), against a model (WAPABA model) streamflow across three sub-basins Pearl River Basin (PRB). The results indicate that LSTM generally demonstrates superior capability than other five models. Using previous month as an input variable improves all When compared with WAPABA model, better two sub-basins. For simulations wet seasons, shows slightly model. Overall, study confirms suitability methods modeling at scale basins proposes effective strategy for improving their performance.

Language: Английский

---

Generative Design in the Built Environment

Automation in Construction, Journal Year: 2024, Volume and Issue: 166, P. 105638 - 105638

Published: July 27, 2024

Language: Английский

---

Monthly Runoff Prediction for Xijiang River via Gated Recurrent Unit, Discrete Wavelet Transform, and Variational Modal Decomposition

Water, Journal Year: 2024, Volume and Issue: 16(11), P. 1552 - 1552

Published: May 28, 2024

Neural networks have become widely employed in streamflow forecasting due to their ability capture complex hydrological processes and provide accurate predictions. In this study, we propose a framework for monthly runoff prediction using antecedent runoff, water level, precipitation. This integrates the discrete wavelet transform (DWT) denoising, variational modal decomposition (VMD) sub-sequence extraction, gated recurrent unit (GRU) modeling individual sub-sequences. Our findings demonstrate that DWT–VMD–GRU model, utilizing rainfall time series as inputs, outperforms other models such GRU, long short-term memory (LSTM), DWT–GRU, DWT–LSTM, consistently exhibiting superior performance across various evaluation metrics. During testing phase, model yielded RMSE, MAE, MAPE, NSE, KGE values of 245.5 m3/s, 200.5 0.033, 0.997, 0.978, respectively. Additionally, optimal sliding window durations different input combinations typically range from 1 3 months, with (using rainfall) achieving one-month window. The model’s accuracy enhances resource management, flood control, reservoir operation, supporting better-informed decisions efficient allocation.

Language: Английский

---

Flood prediction with time series data mining: Systematic review

Natural Hazards Research, Journal Year: 2023, Volume and Issue: 4(2), P. 194 - 220

Published: Oct. 11, 2023

The global community is continuously working to minimize the impact of disasters through various actions, including earth surveying. For example, flood-prone areas must be identified appropriately, predicted, understood, and socialized. In that case, it will increase risk disaster impacts on affected population in form death, property damage, socio-economic losses. data mining approach has had a significant influence research related flood prediction recent years, namely its researchers forecast, classification, clustering. Floods can also predicted using time series used predict future, type data-driven been developed widely applied predictions hydrology. A review identify, evaluate, interpret all relevant results carried out so far for with approach. method this study PRISMA as tool guide evaluating systematic reviews meta-analyses. Some things discussed are types data, floods their parameters, approaches combinations, evaluation methods studies. This found although univariate dominates studies, multivariate Analysis (53 papers or 48.62%) strengthen long term short term, t; an opportunity further research. opportunities combining team Estimation Classification approaches. contrast, optimization 11% total study. next opportunity. chosen find gap; less response kind flood, easier be. four floods: River Flood (76.1%), Urban (11.9%), Coastal (6.4%), Flash (5.5%). dominant use RMSE, absolute measure same scale target (depending data). Methods produce percentages, such MAPE, which understand by end users, need more frequently future amount determines whether resulting model good, especially choice approach, long-term short-term. Whether short-term long-term, forecasting essential mitigation, based series. Short-term early warning system, while support infrastructure planning government.

Language: Английский

---

A Lake-Flood Forecasting Method Coupling the Ce-Qual-W2 and Pinn Models

Published: Jan. 1, 2025

The limited availability and low accuracy of hydrological data severely influence the flood forecasting. To address this issue, paper proposes a new way to predict floods that combines CE-QUAL-W2 model for lakes' hydrodynamics with PINN physical information. is employed verify dynamic process water level volume in Lake during season. We input lake, verified by model, into model. Utilizing we can learn nonlinear patterns reservoir discharge from historical directly transform problem solving differential equations an optimization loss functions regular equations. real-time simulated also incorporated Xin-An-Jiang (XAJ) Long Short-Term Memory (LSTM) was compared results prediction performance obtained CE-QUAL-W2&PINN This study selects Luoma as research subject, choosing 35 representative events occurred between 1960 2022. show that, (1) events, relative errors observed values were all within 20%, indicating good simulation accuracy. (2) Compared LSTM XAJ models, demonstrates higher faster forecasting capabilities 3-hour period, achieving improvement approximately 30% both training testing. (3) overall determination coefficient CE-QUAL-W2&PIN stands at 0.919. error less than 10% flow periods.

Language: Английский

---

Deep learning model for flood probabilistic forecasting considering spatiotemporal rainfall distribution and hydrologic uncertainty

Journal of Hydrology, Journal Year: 2025, Volume and Issue: unknown, P. 132879 - 132879

Published: Feb. 1, 2025

Language: Английский

---