Presenting new artificial intelligence techniques for finding optimum performance of solar chimney power plant

Abstract

The various applications of solar energy have progressed significantly over recent decades. In this paper, finding the optimum performance criteria of a SCPP is studied. The mathematical modelling of 1-D flow is written and the optimum performance is determined for each group of geometric and environmental variables including the angle of collector roof (β), chimney divergence (α) and ambient wind velocity (Va ). In addition, different artificial intelligence techniques including regression tree, genetic programming and a novel method named trended regression tree are applied to find optimum performance conditions at any arbitrary group of α, β and Va . Results indicate that maximum power output (Pt ) and associated mass flow rate ( m ˙ ) increase with increasing chimney divergent angle and wind velocity at a constant collector roof angle. Comparing the obtained results of artificial intelligence techniques shows that trended regression tree method predicts the optimum conditions of SCPP better than other methods. Genetic programming is used to present fitted algebraic functions on Pt and m ˙ and the ratio of turbine pressure drop to pressure potential (PR). Results demonstrate the algebraic relation of PR estimates the PR accurately at each group of variables.