Power Prediction of Photovoltaic System using Neural Network Models
Master thesis
Permanent lenke
http://hdl.handle.net/11250/2622441Utgivelsesdato
2019Metadata
Vis full innførselSamlinger
Sammendrag
This workconsidersa photovoltaic (PV)system installed on the rooftop of Agder Energi’s headquarters located at Kjøita, Kristiansand. The system includes three different types of solar PV modules; Suntech (multi-Si), Sharp (a-Si/μ-Si) and REC (multi-Si), that have a total installed DC capacity of 45 kWp. The system is grid-connected and instrumented for research and monitoring purposes.Artificial Neural Network(ANN)models were trained to obtain the lowest mean square error(MSE), by testing different configurations using a model-based trial and error approach. The modelconfigurations that gavethe lowest(MSE) wereused to predict the power production from each of the PV modules using forecasted weather parameters obtainedfrom MEPS (MetCoOp Ensemble Prediction System), with a one-day ahead and two-days ahead forecast horizon. The input selection of the models wasbased on both model-free and model-based approach, where the final input selectionresulted inglobal horizontal irradiance, wind speed, air temperature and air mass, with the power (AC) productionas output (target).The results indicated that the modelconfigurations of 20hidden neurons in first hidden layer, and 2 hidden neurons in second hidden layergavethe lowest MSEfor allPV modules. Results from the testsetsshowedthat the best model for Suntech gaveMSE = 0.0454,Sharp gaveMSE = 0.0325andREC gaveMSE = 0.0492. R2-valuesbetween 0.95 and 0.96 wereobtained for all three models, indicating good fitting of the predicted valuesand the targets. Testing the Suntech and REC modelswith a hold-out set provided slightly less precise predictions compared to the results from the testset, while a higher precision was found forSharpmodules. Testing the model configurations with forecasted weather parameters indicated that the forecast accuracy of the weather will influence the power prediction, andthe performance parameters will be accordingly. The one-day ahead forecasts provided MSE equal to 0.2647, 0.2378and 0.2647, and for the two-days aheadforecast horizon an MSE equal to 0.2996, 0.2252and 0.2719for Suntech, Sharp and REC, respectively.An error much higher compared to the test set and hold-out set for the models, which inevitablywas expecteddue to the weather forecast uncertainties.Based on the findings in this work, itcan be concludedthat a further optimization of the models will be necessary before obtaining even more precise predictions. However, the models did show good fitting for several days and a potential for using ANN modelsfor power prediction of PV modules.
Beskrivelse
Master's thesis Renewable Energy ENE500 - University of Agder 2019