Processing sky images collected from sky camera

• Main Author: Ye, Penghao
• Other Authors: Gooi Hoay Beng
• Format: Theses and Dissertations
• Language: English
• Published: 2016
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/68736
• Institution: Nanyang Technological University

Fluctuating and intermittent irradiance interrupts the photovoltaic (PV) power generation. Massive PV grid-connection may bring about the stability problem to running of the power system. The forecast of output power of a PV plant can balance the dispatch of conventional energy (thermal power, hydropower, etc.) and photovoltaic power, and ensure the power supply quality. Irradiance, as the main factor affects the output power of PV power generation. This will lead to the fluctuation of the output power- due to its volatility and mutability, which brings huge difficulty to the PV power prediction. The traditional forecasting method, which is largely based on physical and statistical models, will bring about the lag to the prediction results. Thus it is unable to provide accurate prediction results continuously. Relevant research results have indicated that the primary cause of the sudden change of irradiance is the clouds’ coverage against the sunlight. The prediction of cloud coverage above the solar panel can be achieved as follows: proceed with all-day automatic observation and image acquisition by using ground-based remote sensing cloud detection equipment, then using image processing technique to extract the feature of the clouds’ future movement. The Total Sky Imager™ is designed to capture the cloud coverage data. And a mechanism aiming at parameterization the cloud image serves the purpose. Since the cloud images are captured far from the ground-based camera, the resolution is low and the noisy points are inevitable. This requires elaborate image processing techniques to eliminate the interference factors, making the quantized irradiation and coverage data more valuable.