Search documents

Browse topics

Document details

Forecasting Photovoltaic Energy Using a Fourier Series Based Method
T. Bien, H.-D. Musikowski
Storage, Energy Performance, Forecast
Components for PV Systems
Subtopic: Engineering and System Integration
Event: 23rd European Photovoltaic Solar Energy Conference and Exhibition, 1-5 September 2008, Valencia, Spain
Session: 4BV.1.29
3088 - 3091
ISBN: 3-936338-24-8
Paper DOI: 10.4229/23rdEUPVSEC2008-4BV.1.29
0,00 EUR
Document(s): paper


In this paper a new method of forecasting photovoltaic (PV) energy production is discussed. The significance of PV power production is growing rapidly. This growth could be further accelerated through an efficient method of forecasting such an unpredictable and fluctuating phenomenon like solar irradiance based PV output power. An efficient method of forecasting could limit problems, which mainly occur because of the non-controllable power of Distributed Generation (DG). The forecast is also necessary for efficient management of electricity grids and for investment planning with regard to PV systems. The traditional forecasts of photovoltaic energy are based on the time series of irradiance energy, which are used to calculate the electrical energy of PV systems. In this paper a method of forecasting the energy of PV systems will be shown, which is based on the time series of energy produced by a solar generator. The energy of a PV generator is represented by a time series of daily output energy of the PV generator for three years, which is grouped into 7, 10 and 15 day sums of AC PV energy. The database of the time series was created at the Institute of Electric Power Systems at Otto-von-Guericke University Magdeburg. The forecasts discussed in this paper are made by fitting a time series with a least square fitting method using a Fourier series as a model function. The Fourier series based method of forecasting the photovoltaic energy seems to be very reliable and robust. In this paper several ranks of harmonics of Fourier series were investigated and compared, in order to decide which rank of harmonics could be especially useful for forecasting PV energy.