Energy performance of a solar cavity receiver with neural output flow regulation under controlled conditions

This article presents the theoretical and experimental performance of a Solar Cavity Receiver (SCR) which operates under controlled conditions of radiation and flow regulation using an Artificial Neural Network (ANN); the objective of this application of neural networks is to obtain the maximum temperature of a fluid in an open hydraulic circuit that presents variations in pressure and temperature of the inlet fluid. The SCR is evaluated for two cases: In the first case, inlet fluid presents a constant temperature and pressure variations of 62–96 kPa, and in the second case, the system presents variations in pressure 55–89 kPa and temperature 29 °C to 31 °C; it was found that an ANN presents a response with a varied range of 2–14%, with average temperature variations lower than 6% respect to the set-point, with the advantage of being self-regulated through ANN training.