When dealing with irradiance forecasting artificial intelligence techniques
In this work, global horizontal irradiation in cloudless days was assessed in Singapore via a series of proposed filters in order to gauge the angiotensin ii receptor of irradiation at ground level caused by anthropogenic haze, primarily originating from forest and agricultural land clearing fires in Sumatra (Indonesia), 150–300 km away, where from the majority of the fire “hot spots” take place.
A novel filtering technique was proposed and implemented to differentiate between the impacts from clouds and air pollution caused by haze on the reduction on irradiance levels on ground. Parameters used in the filtering included relative humidity and diffuse irradiance fraction in order to subsequently remove the effects of rain and clouds on solar irradiance. A third filter was proposed in order to eliminate cloud enhancement effects (over-irradiances) or near cloudy conditions which could fall within sensor accuracy levels.
Global horizontal irradiance level reductions reached ∼15% during a strong haze episode of June 2013, both when performing the analysis with the use of a pyranometer as well as with a crystalline silicon sensor that has a narrower spectral range (300–1100 nm).