ABSTRACT
The diffusion of solar energy is increasing at the exponential rate at the utility scale. The stochastic nature and its behavior are difficult to predict, thus forcing researchers to use the approach of deep learning. As the world is consuming more electricity and wasting the most vital aspect of the energy, the generation of solar energy is crucial for the balancing of the power grid and its optimization. For efficient operation and stability, it is essential to forecast solar energy. To do so, physical laws are applied for computation of the direction of the sun’s rays as well as its accumulated energy. The computational evaluation for the production of solar energy is quite challenging in the field of AI. There is a need to take into consideration certain factors like the positioning of the sun, fluctuation of the weather condition, and many more. It becomes challenging in cloudy weather because the movement of the clouds is unpredictable. This sometimes introduces computational error in the result, which may be ignored. In this chapter, an attempt is made to use the deep learning networks and models to capture the movement of the cloud pattern and how its impact will be on the generation of solar energy. In the proposed model using a deep learning network, the error rate reduces to 11%. Previous studies have shown an error rate of 21%. The 10% improvement in error rate will have a positive effect on the development of the solar energy industry in a healthier manner and reduce the cost (in USD) along with the reduction in the dependency on the emission of carbon dioxide.
