Cooling performances time series of CSP plants: Calculation and analysis using regression and ANN models. (September 2020)

• Record Type: Journal Article
• Authors: Boukelia, T.E.; Ghellab, A.; Laouafi, A.; Bouraoui, A.; Kabar, Y.
• Is Part Of: Renewable energy. Volume 157(2020)
• Page Start: 809
• Page End: 827
• Publication Date: 2020-09
• Subjects: ANN -- Cooling performance -- CSP -- Regression model -- Statistical analysis; Renewable energy sources -- Periodicals; Power resources -- Periodicals; Énergies renouvelables -- Périodiques; Ressources énergétiques -- Périodiques; 333.794
• DOI: 10.1016/j.renene.2020.05.012
• Languages: English
• ISSNs: 0960-1481

Abstract

Abstract: Concentrating solar power (CSP) plants use large quantities of water, for different processes such as cycle makeup and cooling. Thus, the estimation of cooling performances in such kind of plants is highly required. On the other hand, yearly round simulations of cooling performances of these plants require many calculations, data analysis, and time consuming. In this regard, empirical models and artificial neural networks (ANN) can be good alternatives in this topic. Therefore, the two main aims of this study are: (1) to compare the cooling performances, including water usage and power consumption for cooling of different CSP layouts, and (2) to develop regression and ANN models () to estimate these performances during the whole year, without passing through a detailed modelling. According to the obtained results, the configurations based on molten salt technology show better cooling performances compared to other configurations, while the direct steam generation plant is the worst. Furthermore, the generated database using ANN is more accurate than that generated by different regression models. However, these regression models still the easiest and the simplest methodology for this purpose. Graphical abstract: Image 1 Highlights: Four different CSP layouts have been considered for the study. The cooling performances of these CSP layouts have been compared. Regression and ANN models have been developed to estimate these performances. Statistical analysis has been evaluated to compare the calculative ability of these models.