Research Of Power Forecasting For Photovoltaic Plant And It Application Under Haze Weather Condition

Due to high concentration of greenhouse gases in the atmosphere have increased dependency on photovoltaic(PV)power,but its random nature poses a challenge for system operators to precisely predict and forecast PV power.The conventional forecasting methods were accurate for clean weather.But when the PV plants worked under heavy haze the radiation is negatively impacted and thus reducing PV Power,therefore to deal with haze weather Air Quality Index(AQI)is introduced as parameter to predict PV power.AQI which is indication of how polluted is the air,has been known to have strong correlation with power generated by the PV panels.In this thesis,a review of forecasting solar power,radiation and a hybrid forecasting tool is proposed to forecast PV power in clear and haze weather condition.The data of the proposed power station is analyzed and a relationship between PV power and meteorological parameters is designed.The training data and test data needed for the proposed model are obtained from measured data in the database of the 3KW PV power station at North China Electric Power University(Latitude:40.0896822,Longitude:116.31003484),Beijing with 12 poly crystalline modules inclination angle 410.Each module produces maximum 245 watt and each module has 60 solar cells with cell efficiency of 17.25%.After review of numerous forecasting methods,a new forecasting method is required to predict PV power in haze condition too;therefore an effort is made in this direction.First of all data is normalized and then trained with neural network,after training,the test data is also normalized and applied to the well trained network to predict PV power in two cases 1)PV power for the day on which average AQI level below 150 is predicted with ordinary neural network method considering radiation,humidity,wind-speed,temperature as input parameters.2)PV power for the day on which average AQI level is above 150 then an additional parameter AQI is also added to the input data to forecast PV power.The errors calculated for four days ahead shows that the accuracy is high during haze weather and the hybrid model is appropriate for weather of Beijing where haze factor is considerable.Output of PV power is negatively impacted by AQI which has never been considered as meteorological parameter in previous models of forecasting of PV power.