Search

This paper investigated the artificial neural network (ANN) modeling to forecast the one-hour ahead global and diffuse solar radiation using the observed data during 2019-2021 at a meteorological station in Bangkok, Thailand. Results from the statistical analysis portrayed the abundance of the tropical solar radiation and its large variation all year round. As a step to move advance the renewable energy use of the solar radiation at the location, the ANN forecasting models were proposed for all climatic conditions. The model input were the ambient temperature, solar elevation, and an insolation parameter namely normalized global solar radiation. The model performance was assessed by using the two statistic parameters: normalized mean bias error (nMBE), and normalized root mean square error (nRMSE). The results showed that the ANN models could forecast the global radiation with the nMBE of 1.79% and nRMSE of 26.20%. For the diffuse radiation forecast, the nMBE and nRMSE of the models were obtained at 7.78% and 36.21%, respectively. In order to demonstrate a benefit of the two developed ANN models, their forecasted horizontal data were employed to further predict the total solar radiation on vertical plane. Benchmarked with the smart persistence model, the developed ANN models possessed the higher forecasting accuracy for all weather conditions. [ABSTRACT FROM AUTHOR]

• Thermal performance of the radiant ceiling cooling was assessed for intermittent operation in tropical office environment. • Insulated envelope and outdoor air heat recovery improved both indoor comfort and energy saving of the radiant ceiling system. • The local people would achieve their neutral sensation by the radiant ceiling cooling for more than 95% of the office hours. • The radiant system saved the electrical energy up to 40% opposed to the conventional mixed-air system. This paper investigated the radiant ceiling cooling in office buildings in a tropical hot humid climate. Full-scale experiments were conducted in an outdoor laboratory chamber to observe the indoor thermal conditions and the envelope heat gain from the intermittent operation of the installed radiant chilled panels (RCP) coupled with an outdoor air unit (OAU). The radiant system was also modeled by using a simulation software namely Transient System Simulation Software (TRNSYS), and the calculation results from the model were compared to the experimental observations. In the simulation part, the validated TRNSYS model was adopted to assess the radiant system performance in a typical office environment under the four schemes of combinations of the technology options for the OAU and the building envelope. The perceived comfort from the radiant cooling was assessed based on the thermal sensation of the local subjects acclimatized to hot humid climate. The energy consumption of the radiant cooling was also apprised as opposed to the conventional mixed-air system. The simulation results presented an insight into the importance of the exhaust air heat recovery and the envelope insulation on energy saving and human comfort improvement for the radiant cooling. [ABSTRACT FROM AUTHOR]