ANFIS based effectiveness and number of transfer units predictions of MWCNT/water nanofluids flow in a double pipe U-bend heat exchanger

In this study, the heat transfer coefficient, Nusselt number, effectiveness and number of transfer units of water mixed multi-walled carbon nanotubes nanofluids passes through a tube-in-tube heat exchanger was experimentally investigated. Investigations were performed in the operating conditions of Reynolds number ranging from 3500 to 12000 and volume concentrations ranging from 0% to 0.3%, respectively. The obtained four parameters were predicted using adaptive neuro fuzzy inference system (ANFIS). The Reynolds number and particle volume loadings are the input data in artificial neural network analysis and heat transfer coefficient, Nusselt number, effectiveness and number of transfer units is output or target. The Nusselt number, heat transfer coefficient, effectiveness, and number of transfer units was enhanced to 31.3%, 44.17%, 2.51% and 2.76% at phi = 0.3% and at a Re of 10005, against base fluid. Implementation of ANFIS with various quantities of neurons in the mid layer provides 1-10- 6 with the correlation coefficient (R2) of 0.9978, and 0.9998 and root mean square error of 0.0018581, and 0.0014159 for heat transfer coefficient and Nusselt number, respectively. The above developed structure has been successful in predicting 96% of variation in all the parameters.

Automated summary

This study experimentally investigated the heat transfer coefficient, Nusselt number, effectiveness, and number of transfer units of water mixed multi-walled carbon nanotubes nanofluids passing through a tube-in-tube heat exchanger. The parameters were predicted using adaptive neuro fuzzy inference system (ANFIS). The developed structure successfully predicted 96% of variation in all parameters.