Assessment of the Wind Energy Potential on High Buildings in Bangkok Using Artificial Neural Network Model Case Study: Wisawawat Buildings Pathumwan Institute of Technology
Abstract:
The purpose of this thesis is to study, analyze and assess the potential of wind energy on the Witsawawat Building, Pathumwan Institute of Technology, Bangkok. The data, which are included; wind speed, wind direction, temperature and relative humidity were collected in one season from January 2018 to December 2018. The experiment are at a rooftop of the Witsawawat Building, which has a coordinate at a latitude of 13.748901 and longitude of 100.5260016, and average wind speed at 1.86 meters per second at an altitude of 25 meters, the wind direction is 183 degrees (The southwest, or from Charoen Phon Intersection). In this research, Artificial Neural Network (ANN) model is used to estimate the potential short-term wind power. The training algorithms of ANN, Gradient Descent algorithm (GD), LevenbergMarquardt algorithm (LM), Extend Kalman Filter (EKF) and Ensemble Kalman Filter (EnKF) are applied to optimize the
weights adjusting. The performance of ANN models were measured by Mean Square Error (MSE). The results show that the ANN structure is six input nodes, ten hidden nodes and an one output node ( 6-10-1), trained with Levenberg-Marquardt algorithm obtains a high correlation coefficient, 0.8867 and the lowest MSE, 0.0076.