Artificial Intelligence for Concentrated Solar Plant Maintenance Management

Concentrated Solar Power (CSP) is an alternative to the conventional energy sources which has had significant advances nowadays. A proper predictive maintenance program for the absorber pipes is required to detect defects in the tubes at an early stage, in order to reduce corrective maintenance costs and increase the reliability, availability, and safety of the concentrator solar plant. This paper presents a novel approach based on signal processing employing neuronal network to determine effectively the temperature of pipe, using only ultrasonic transducers. The main novelty presented in this paper is to determine the temperature of CSP without requiring additional sensors. This is achieved by using existing ultrasonic transducers which is mainly designed for inspection of the absorber tubes. It can also identify suddenly changes in the temperature of the CSP, e.g. due to faults such as corrosion, which generate hot spots close to welds.