Solar energy represents one of most important renewable energy sources. The knowledge of short-time differences in the field of sun radiation is crucial for energy production. Since the ration between direct and diffuse solar radiation is important for the quality of solar energy, a model for diffuse solar radiation based on artificial neural network was developed. The model was developed based on global solar radiation and other meteorological measurements. In the paper the design of empirical model for diffuse solar radiation is presented as well as the importance of input-data selection to the quality of model.
F.18 Transfer of new know-how to direct users (seminars, fora, conferences)
COBISS.SI-ID: 28378919The public warning system of ground level ozone is of high importance and for this we need a reliable modelling method for predicting it. A mobile weather station for on-line prediction of ground level ozone concentration represents a practical approach of monitoring the air quality. A nonparametric and probabilistic model is designed in order to adapt itself to the location of the weather station despite the limited availability of meteorological measurements and pollutant concentration measurements. The model is an evolving system from the Gaussian process modelling framework. It uses a vector of current and past measurements to predict a normally distributed random variable. The model may predict inaccurately in atypical weather and emission conditions, but it expresses its uncertainty measure within the predicted variance during the adaptation to new conditions. The choice of model inputs affects the model accuracy. A systematic approach is provided for selecting the most important measurements. The identified model is validated on measurements.
F.18 Transfer of new know-how to direct users (seminars, fora, conferences)
COBISS.SI-ID: 28523815The team from MEIS company cooperates for three years in the international programme MODARIA from International Atomic Energy Agency IAEA in the section for air-pollution modelling. They have used their comprehensive knowledge from the field, which is also result of the present project, for preparing the MODARIA’s section final report. The results were highly appreciated and resulted in activities proposal for the next three years. The proposal contained enhanced validation of air-pollution forecasting models. The proposal was accepted with great interest and will likely be placed as official activity for the next three years.
B.04 Guest lecture
COBISS.SI-ID: 29164583