Support vector machine regression has demonstrated high predictive accuracy for evaluating the thermal and thermo-hydraulic performance of an evacuated tube collector solar air heater fitted with a pierced twisted tape insert, according to an experimental study published in Discover Applied Sciences. The researchers developed and compared multiple machine learning algorithms to simultaneously predict three crucial operational outputs from empirical measurements, establishing that support vector machine regression delivered the strongest predictive capabilities.
Evacuated tube collector solar air heaters are well suited for low-temperature solar heating demands, but their overall thermal and thermo-hydraulic efficiency depends on intricate, nonlinear interactions among operational, environmental, and geometrical parameters. While prior research in solar thermal engineering has employed machine learning models to forecast isolated indicators such as outlet air temperature or thermal efficiency, little attention had previously been directed toward simultaneously predicting outlet air temperature, thermal efficiency, and thermo-hydraulic efficiency for an evacuated tube collector solar air heater specifically equipped with a pierced twisted tape insert from measured operational and geometrical variables.
To bridge this gap, the study experimentally explored how incorporating the pierced twisted tape insert influences the thermal performance of the collector system. Gathering reliable training and validation data required the experimental measurement of eight separate operating, ambient, and geometrical parameters across testing conditions. These empirical input parameters comprised solar irradiance, mass flow rate, ambient temperature, wind speed, pitch length, hole diameter, pressure drop, and temperature rise.
Using the recorded experimental data, the research team developed and comparatively analyzed three distinct machine learning regression models. The tested algorithms included Support Vector Machine Regression, Stochastic Gradient Descent Regression, and Extra Trees Regression. Each regression framework was configured to handle the multi-parameter input dataset and generate simultaneous predictions for the target metrics: outlet air temperature, thermal efficiency, and thermo-hydraulic efficiency. By testing these different mathematical regression strategies against identical experimental datasets, the authors systematically assessed which algorithm best captured the complex parameter interactions governing the solar air heater.
The physical testing revealed clear behavioral patterns in how fluid flow dynamics influenced thermal output. Experimental results demonstrated that as the air mass flow rate through the evacuated tube collector increased, the outlet air temperature, the overall temperature rise, and the thermal efficiency all experienced a measurable decrease. At the same time, the rate of heat recovery increased alongside higher mass flow rates.
When comparing algorithm accuracy on the unseen testing dataset, Support Vector Machine Regression exhibited the best overall prediction ability among all models evaluated. For predicting outlet air temperature, the support vector machine achieved a coefficient of determination of R2 = 0.98256. For predicting thermal efficiency, the same regression model attained an R2 value of 0.98791. Finally, when evaluating thermo-hydraulic efficiency, the support vector machine achieved an R2 of 0.98502.
The researchers noted that these machine learning models can be utilized to quickly estimate evacuated tube collector solar air heater performance across the experimentally tested operating region. By providing rapid computational estimations, the data-driven framework supports performance evaluation and engineering optimization while significantly lowering the necessity for repeated experimental trials.
Subject of Research: Machine learning prediction of evacuated tube solar air heater performance
Article Title: Machine learning based prediction of thermal performance for an evacuated tube solar air heater with pierced twisted tape
Article References: A., V., C., R., T., K., V., R., S., P., & Terefe, T. O. (2026). Machine learning based prediction of thermal performance for an evacuated tube solar air heater with pierced twisted tape. Discover Applied Sciences. https://doi.org/10.1007/s42452-026-09817-1
Image Credits: AI Generated
DOI: 10.1007/s42452-026-09817-1
Keywords: evacuated tube solar air heater, pierced twisted tape, support vector machine regression, stochastic gradient descent regression, extra trees regression, thermal efficiency, thermo-hydraulic efficiency, outlet air temperature, temperature rise, mass flow rate, heat recovery rate, pressure drop, solar irradiance, ambient temperature, wind speed, pitch length, hole diameter
News Source: Denise Maddox. (November 11, 2026). Support Vector Machine Regression Accurately Predicts Evacuated Tube Solar Heater Performance. Scienmag.



