The aim of this thesis is to apply and evaluate potential forecasting models for solar power production, based on data from a photovoltaic facility in Sala, Sweden. The thesis evaluates single step forecasting models as well as multiple step forecasting models, where the three compared models for single step forecasting are persistence, autoregressive integrated moving average (ARIMA) and ARIMAX. ARIMAX is an ARIMA model that also takes exogenous predictors in consideration. In this thesis the evaluated exogenous predictor is wind speed. The two compared multiple step models are multiple step persistence and the Gaussian process (GP). Root mean squared error (RMSE) is used as the measurement of evaluation and thus determining the accuracy of the models. Results show that the ARIMAX models performed most accurate in every simulation of the single step models implementation, which implies that adding the exogenous predictor wind speed increases the accuracy. However, the accuracy only increased by 0.04% at most, which is determined as a minimal amount. Moreover, the results show that the GP model was 3% more accurate than the multiple step persistence; however, the GP model could be further developed by adding more training data or exogenous variables to the model.

År 2019 installerades 732 solpaneler på taket i ett hus i Örebro universitet. Energiproduktionenav anläggningen samlades i en databas i Akademiska Hus med ett antal parametrar från enväderstation i samma hus. Att kunna modellera den här produktionen som en funktion avväderparametrar och historiska värden med hjälp av maskininlärning, och jämföra olikamodeller är målet i detta projekt. Det finns gjorda arbeten med samma mål i olikalaborationsmiljöer och andra platser men inte för denna anläggning. Mätvärden under två årfrån 2019 till 2021 kommer från Akademiska Hus och resultaten blir två modeller: ett NarrowNeural Network samt en Support Vector Machine med 7 procent avvikelse och en NonlinearAutoregressive Neural Network för envariatmodellen. / In 2019, 732 solar panels were installed on the roof of a building at Örebro University. Thesolar power production of the facility has been collected in a database in Akademiska Hus,along with several parameters from a weather station in the same building. The goal of thisproject is to model solar power production as a function of weather parameters and historicalvalues using machine learning techniques. This study investigates various predictive models tofind a suitable model for predicting this production. There have been several studies in theliterature that have performed this goal in various laboratory environments and other places,but not for this facility. The measured data for this study is recorded by Akademiska Hus forover two years from 2019 to 2021. The results of this work lead to two suitable machine learningmodels while using weather parameters: 1) Narrow Neural Network and 2) Support VectorMachine with 7% errors in both models. Moreover, this study has investigated univariatemodels to predict the solar power production as a time series based on its historical data. Forthis aim, a Nonlinear Autoregressive Neural Network has been applied which results inconsiderably low errors in the evaluations.