Optimizing Performance of Evacuated Tube Solar Air Heater with Perforated Twisted Tape

Abstract This paper introduces a new optimization strategy for the thermal performance parameters of an Evacuated Tube with Perforated Twisted Tape Solar Air Heater (ETPTTSAH) using an integrated Taguchi, ANOVA, and X Gradient Boosting (XGB)-Artificial Neural Network (ANN) technique, the first such integration in the research on ETPTTSAHs. The parameters used in experiments with ETPTTSAH included the air mass flow rate (100, 300 and 500 kg/h), pitch length (PL) (57, 72, and 86 mm), and hole diameter (HD) (2, 3, and 4 mm) of the perforated twisted tape (PTT). An orthogonal array L27 was drawn up with thermal efficiency, thermo-hydraulic efficiency, outlet air temperature, and useful heat gain as outcome responses. Findings show that the thermal performance of the evacuated tube solar air heater (ETSAH) improves considerably with the addition of PTT relative to a plain tube setup. ANOVA and Taguchi optimization were performed in MINITAB, and the ANN was trained in Anaconda using the XGB regression algorithm. Optimization showed that mass flow rate is the key (>94% contribution to all the outputs), whereas PL and HD have minimal contribution (<7%). The consistency of the results is proved by the fact that both methods, Taguchi statistical analysis and XGB-ANN machine learning model (R2 = 0.999), measure the same optimal parameters: mass flow 300 kg/h, PL 57 mm, HD 2 mm. This combined approach provides a robust optimization framework for solar thermal systems that can be applied to related systems.