Comparative Study of Linear Regression Techniques for Analyzing THPP of Solar Parabolic Trough Receiver with Twisted Tape Inserts using Machine Learning.

Abstract

This study presents a comparative analysis of four different linear regression techniques for predicting the THPP of a solar parabolic trough receiver with twisted tape inserts. The four techniques analyzed are: ordinary linear regression, interaction linear regression, robust linear regression, and stepwise linear regression. The study is aimed at determining the most suitable technique for predicting the heat transfer characteristics of the solar parabolic trough receiver with twisted tape inserts. This study's data set was developed through experimental study and includes various designs of twisted tape inserts, flow rates, and sun irradiation levels. The experimental design includes three twist ratios (3, 4, and 5), 5 perforation proportions (0.05, 0.1, 0.15, 0.20, and 0.25), and 3 wing depth proportions (0.1, 0.2, and 0.3), yielding 210 data sets. The effectiveness of the model is assessed by Root Mean Square Error (RMSE) and the coefficient of determination (R2). The analysis has revealed that the RMSE and R2 indices for stepwise linear regression are 0.076908 and 0.97, correspondingly. The results of the study show that the stepwise linear regression technique is the most suitable for predicting the THPP of the solar parabolic trough receiver with twisted tape inserts. The stepwise linear regression technique provided the best fit to the data and showed the highest accuracy compared to the other linear regression techniques. The findings of this study can be useful in designing and optimizing solar parabolic trough receivers with twisted tape inserts.