Multi Day Ahead Flood Predictionin South Asian Tropical Zone Using Deep Learning

Research Square (Research Square), Год журнала: 2024, Номер unknown

Опубликована: Март 12, 2024

Abstract A reliable and accurate flood forecasting procedure is a critical need due to the hazardous nature of disaster. Researchers are increasingly favoring innovative approaches with enhanced accuracy, such as machine learning models, over traditional methods for this task. However, lack studies regarding South Asian tropical region, which has its own climate characteristics, was unidentified major issue. This research delves into viability employing ANN, LSTM, BLSTM, ConvLSTM2D Transformer models multi-day ahead simulation. One-day, two-days three-days were selected lead times task considering lower reaches Mahaweli catchment in Sri Lanka, mostly affected by Northeast Monsoon. The prediction capability extreme stream flows also interest. Observed rainfall data from three nearby rain gauges, along historical discharges target river gauge, serve input features models. ANN model showed worst performance, having mean NSE 0.67. An improved performance observed compared LSTM based especially multiple day scenarios. For all water levels drops down drastically when time increased.

Язык: Английский

---

The Technological Assessment of Green Buildings using Artificial Neural Networks

Heliyon, Год журнала: 2024, Номер 10(16), С. e36400 - e36400

Опубликована: Авг. 1, 2024

This study aims to construct a comprehensive evaluation model for efficiently assessing appropriate technologies within green buildings. Initially, an Internet of Things (IoT)-based environmental monitoring system is devised and implemented collect real-time parameters both inside outside the building. To evaluate technical suitability buildings, this employs multifaceted approach encompassing various criteria, including energy efficiency, impact, economic benefits, user comfort, sustainability. Specifically, it involves parameters, analysis consumption data, indoor quality indicators derived from satisfaction surveys. Subsequently, Multi-Layer Perceptron (MLP) selected as conventional artificial neural network (ANN) model, while Long Short-Term Memory (LSTM) chosen advanced recurrent in realm deep learning. These models are utilized process explore collected data assess The dataset comprises physical quantities such temperature, humidity, light intensity, well efficiency building operating costs. Furthermore, assessment considers building's life cycle factors health, safety. By incorporating these holistic achieved, ensuring technologies' effectiveness. prediction results demonstrate that proposed hybrid exhibits high accuracy robust stability predicting parameters. For instance, Root Mean Square Error (RMSE) temperature 1.2 °C, Absolute (MAE) 0.9 determination coefficient (R

Язык: Английский

---

Behavior of LSTM and transformer deep learning models in flood simulation considering South Asian tropical climate

Journal of Hydroinformatics, Год журнала: 2024, Номер 26(9), С. 2216 - 2234

Опубликована: Сен. 1, 2024

ABSTRACT The imperative for a reliable and accurate flood forecasting procedure stems from the hazardous nature of disaster. In response, researchers are increasingly turning to innovative approaches, particularly machine learning models, which offer enhanced accuracy compared traditional methods. However, notable gap exists in literature concerning studies focused on South Asian tropical region, possesses distinct climate characteristics. This study investigates applicability behavior long short-term memory (LSTM) transformer models simulation considering Mahaweli catchment Sri Lanka, is mostly affected by Northeast Monsoon. importance different input variables prediction was also key focus this study. Input features included observed rainfall data collected three nearby rain gauges, as well historical discharge target river gauge. Results showed that use past water level denotes higher impact output other such rainfall, both architectures. All denoted satisfactory performances simulating daily levels, with Nash–Sutcliffe Efficiency (NSE) values greater than 0.77 while encoder model superior performance encoder–decoder models.

Язык: Английский

---

Behavior of LSTM and Transformer Deep Learning Models in Flood Simulation Considering South Asian Tropical Climate

Research Square (Research Square), Год журнала: 2024, Номер unknown

Опубликована: Март 18, 2024

Abstract The imperative for a reliable and accurate flood forecasting procedure stem from the hazardous nature of disaster. In response, researchers are increasingly turning to innovative approaches, particularly machine learning models, which offer enhanced accuracy compared traditional methods. However, notable gap exists in literature concerning studies focused on South Asian tropical region, possesses distinct climate characteristics. This study investigates applicability behavior Long Short-Term Memory (LSTM) Transformer models simulation with one day lead time, at lower reach Mahaweli catchment Sri Lanka, is mostly affected by Northeast Monsoon. importance different input variables prediction was also key focus this study. Input features included observed rainfall data collected three nearby rain gauges, as well historical discharge target river gauge. Results showed that use past water level denotes higher impact output other such rainfall, both architectures. All denoted satisfactory performances simulating daily levels, especially low stream flows, Nash Sutcliffe Efficiency (NSE) values greater than 0.77 while Encoder model superior performance Decoder models.

Язык: Английский

---

Lake Water Level Forecasting Using LSTM and GRU: A Deep Learning Approach

Lecture notes in networks and systems, Год журнала: 2024, Номер unknown, С. 197 - 216

Опубликована: Янв. 1, 2024

Язык: Английский

---