Applications of artificial neural networks in concentrating solar power systems

Concentrating solar power (CSP) systems are one of the growing solutions to increased demands for renewable electricity generation. This growth implies the global capacity of these systems and, therefore, requires an increase in characterization tasks to ensure availability, design, and reliability. Accurate electric power forecasting contributes to guaranteeing safe dispatch and stable operation of a CSP system. As a great prediction tool, artificial neural network (ANN) methods recently have been used in CSP forecasting. In this chapter, applications of the ANN-based models to predict the key design criteria, and thermal and economical parameters that influence the performance of CSP systems are discussed. The results have shown different types of classical ANN models, in particular: Multilayered Perceptron ANN (MLP-ANN), forward feed ANN (FFANN), Data Handling Group Method (DHGM), and Adaptive Neuro-Fuzzy Inference System (ANFIS) models are used for performance prediction of CSP systems. The ANN-MLP model is the most widely and well-developed model for predicting the performance of CSP devices and offers a powerful tool for the simulation of such a CSP system. However, the prediction results of classical ANN methods are no longer accurate enough to predict in the smart grid. To improve the prediction accuracy, hybrid artificial intelligence models are needed to determine the optimal parameters of basic ANNs in order to maximize the performance prediction accuracy of different CSP systems to become more efficient and low computationally in the future.