Detection of Cloud Cover Average Percentage based on Adaptive Threshold Method for Energy Forecasting in Photovoltaic System

Abstract

Over the past decade, photovoltaic (PV) technology has become one of the fastest-growing renewable energy technologies. However, due to the different energy outputs according to the weather and solar occlusion by the clouds, this will be a challenge for utility companies to use solar energy effectively. Solar Power Producers (SPP) are also asked to submit solar PV generation forecasts. In this context, our study aims to develop a method of analyzing the percentage of cloud cover based on satellite images of peninsular Malaysia. The Adaptive Threshold method is chosen to be used in this study to classify satellite images. The Adaptive Threshold algorithm is coded in Python to detect clouds and analyze the average percentage of cloud cover in Peninsular Malaysia from 8 am (0800) to 7 pm (1900). The analysis of the average percentage of cloud cover in Peninsular Malaysia is done based on the Frequency Cloud Index which has been identified by the maximum average percentage of cloud cover in Peninsular Malaysia which is from 0% to 60%. The average percentage of cloud cover data in Peninsular Malaysia was used to analyze two (2) case studies. The first case study is an analysis for each month to see the trends, maximum and minimum cloud cover in Peninsular Malaysia while the second analysis is to analyze the trend, maximum and minimum cloud cover in Peninsular Malaysia throughout the year from 5 April 2018 to 16 December 2018. Through analysis made, such an approach is seen can be used as an initial process in making predictions about solar irradiance towards solar energy efficiency.