ระบบการพยากรณ์ปริมาณน้ำไหลเข้าเขื่อนบางลาง

Abstract

Forecasting water inflow into the Bang Lang Dam is important for the management of the Pattani River Basin, which serves as a multi-purpose irrigation source for electricity generation and agriculture in the area. Currently, the information system provides various reports but lacks predictive information regarding the amount of water flowing into the dam, which is crucial for effective water management. The prediction of dam inflow needs to be studied in order to understand the factors that affect the amount of water inflow, serving as a key element in accurate forecasting. Therefore, this research aims to study the factors that influence the water inflow to develop a dashboard model for forecasting water inflow in the Bang Lang Dam located in Bannang Sata District, Yala Province. The study utilized H2O's deep learning model, specifically feedforward neural networks, to create a predictive model for water inflows. Data were imported daily from January 1, 2012, to December 31, 2020. The most significant factor influencing the forecast of water flowing into the dam was the amount of water flowing into the Bang Lang Dam from the previous day, followed by daily rainfall, daily average temperature, daily average relative humidity, and average daily air pressure from STH031 station. The stations BTGH, BLD1, and VLGE35 followed with weight values of 0.136, 0.134, and 0.128, respectively. The model's accuracy was measured using MAE (Mean Absolute Error): 1.300, RMSE (Root Mean Square Error): 3.111, R2 (Coefficient of Determination): 0.767, and R (Correlation Coefficient): 0.876, which indicates good reliability. The model can be displayed as a dashboard tailored to the user's needs. The dashboard is divided into two parts. The first part shows the forecasting results of the amount of water flowing into the dam, presenting 22 variables and displaying the results in millions of cubic meters per day. The second part presents the necessary information to facilitate the operation of the staff involved in water management.