Artificial neural network model for wind energy on urban building in Bangkok

Abstract

Renewable energy is clean and effectively infinite. Wind is as sources of sustainable energy. Accessing wind power, it can reduce electric cost for urban building. Wind power generates electricity by converting kinetic energy in wind to generate electricity. In this paper, wind energy is measured on urban building in Bangkok at Pathumwan Institute of Technology (PIT) which is the height from the ground at 25 meters. Approximating a wind power, the weather data consists of wind speed, wind direction, temperature and humidity. The datasets are collected in a minute and converted into database system (PITWeatherDB). Artificial Neural Network (ANN) has been applied to estimate the potential wind energy in short-term. The performances of ANN are considered by mean square error (MSE) and the correlation between ANN output and observed data from PITWeatherDB are measured. The experimental results show that the topology of six neuron nodes in input layer, ten neuron nodes in hidden layer and a neuron node output is trained by Levenberg–Marquardt algorithm. It has high correlation and minimum MSE.