Transient performance analysis of a heliostat field: Using artificial neural network to predict the net radiation

Abstract

The dynamic analysis of the relative position of the sun ray from the heliostat mirrors in a heliostat solar field is performed. Solar angles, such as solar azimuth and solar elevation angles, are carried out in hourly basis for selected 4 days chosen as representative days to investigate the solar angles for the four seasons of the year. The dynamic position of each heliostat mirror relative to the south and ground is also simulated in this study. Surface azimuth and surface tilt angles are the two variables, which are modeled to determine the position of the mirrors. Furthermore, the hourly variation of the incident angle of the sun ray on each surface of the mirror is determined using well‐established set of equations. By utilizing artificial neural network (ANN) and the calculated hourly incident angles of each mirror, the hourly direct normal irradiance (DNI) data of the considered site, and surface area of the mirrors, the total amount of radiation power reflected by the heliostat field is determined in hourly basis for a whole year. The total daily reflected energy of the field was calculated for 365 days. The daily energy reflected by the heliostat field was maximum during January 20 and June 28 with the approximated value of 100 kWh/day. Finally, monthly mean energy output of the field was compared for the 12 months of the year. The average energy output of the heliostat field ranged from 65.00 to 65.27 kWh/day.