Short-term Solar Irradiance and Power Prediction for Large Scale Grid Integration of Photovoltaic Electricity

Global use of alternative energy sources like solar photovoltaic is increasing rapidly due to a continuous decline in the cost as well as the environmental concerns associated with the consumption of fossil fuels. However, the output power from a PV plant is stochastic in nature due to meteorological conditions, which make PV an unreliable power source. Several studies have reported the use of long and medium term solar irradiance prediction methods to make PV power more reliable and at the same time decrease the energy storage systems. But there are no reports addressing the use of shortterm solar irradiance and PV power prediction methods for managing the ramping requirements of PV power integration into the grid. The objective of this thesis was to develop a system to make PV power predictable, more reliable and dispatchable to utilities in long, medium and short-term time horizons without using energy storage systems. The output power of a PV plant can fluctuate by over 50% in 30 to 90 seconds. This imposes high ramp rate in the grid, which a traditional grid cannot withstand. To minimize this ramping effect, a novel short term solar irradiance and PV power prediction system was implemented. This method utilized an array of four sensors capable of measuring the cloud shadow speed and direction. The speed and direction of cloud shadow was used for the prediction of solar irradiance at a location within the PV plant and the output power of PV plant. The ramp down rate was calculated and a linear power curtailment technique was applied to minimize the ramp rate in PV power generation. The power from the PV plant was injected into IEEE 34-node test feeder in EMTP-RV to analyze the effect of PV power in grid simulation. The results indicated a significant decrease in the ramp down rate of PV power generation using short-term prediction system. This show that short term solar irradiance and PV power prediction can be used for the large scale grid integration of PV electricity without using energy storage systems