Analysing Effect of Solar Photovoltaic Production on Load Curves and their Forecasting

Abstract

Increasing share of intermittent renewable energy sources has generated several issues for power system operators. One aspect of these is the unpredictability of volatile production, affecting day-ahead load forecasting on system level, which is a major challenge for transmission system operators to solve. The literature widely discusses the short- and medium-horizon forecasting methods for weather dependent renewable energy sources, and proposals have also been raised to solve the issue. The aim of present paper is to estimate the effect of solar photovoltaic generation on the daily load curve of a national power system. To achieve this, current forecasting methods of transmission system operators are reviewed and evaluated. Then an own forecasting method is designed and implemented, based on historical load data from years where share of solar photovoltaics was neglectable. A learning algorithm is used for future predictions, using installed capacity and weather data. Finally, forecasted and actual load curves are compared, and effects of solar photovoltaic generation are estimated.