Using Neural Network to Simulate and Predict the Power Generation by Thin-film Solar Cell

• Main Authors: Yung-Chih Liu, 劉勇志
• Other Authors: Yung-Hsiang Hung
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/40583230104124750635

碩士 === 國立勤益科技大學 === 工業工程與管理系 === 99 === This study was designed to use the artificial neural network system to simulate and estimate differences in manufacture procedure of different thin-film solar cell production lines, and furthermore analyze differences in power generation capacity arising from key variation factors of solar power generation. This study collected the actual power generation capacity of thin-film solar cells and applied the artificial neural network system in actual simulation. Meanwhile, it analyzed the variance between simulated power generation capacity and the actual measured power generation capacity, and adjusted parameters of the artificial neural network system until the values were in a convergent tendency. The proposed simulation construction system can be used to compare the actual power generation capacity with the simulated power generation capacity after the installation of the solar power system. It can also be used to determine the abnormality or the need for repair and maintenance of the outdoor solar system (panel) to achieve the best power generation efficacy of the solar power system.