Experimental investigation supported by artificial neural networks (ANNs) for predicting the heating performance of a cyclone separator coupled with induction heating coil

This paper proposes a novel dust separation and thermal inactivation system that couples an electromagnetic induction heating coil with a cyclone separator. The idea is aimed at the thermal inactivation of potential microorganisms (viruses and other pathogens) associated with dust-laden gas streams, particularly from road surfaces. The integrated system was tested in an experimental setup as a proof-of-concept investigation at different cyclone-induction heating temperatures of 100–400 °C (373–673 K) and air discharge flow velocities of 5–30 m/s. In addition, the influence of operating parameters such as cyclone heating temperature (as induction heating inputs), inlet air velocity, and heating time on the efficacy of the induction heating process were investigated. The time-averaged temperature measurements along the axial direction (z-down) showed predominant temperature fluctuations in the cylinder and cone sections for all cases, especially when the gas plummets down the cyclone. Temperature was highest in the conical section (z = 525 mm) at 320 °C (593 K) at 5 m/s and lowest in the outlet section (z = 375 mm) at 185 °C, which was expected. The results show that inlet air velocity plays a crucial role in affecting temperature dynamics. Conclusive results indicate that with increasing temperature, pressure drop decreases significantly. Along with the experimental results, multiple regression indices using an Artificial Neural Network (ANN) modeling were used to predict the cyclone's induction heating performance, showing good prediction quality in terms of the mean square errors (MSE and RMSE) and the correlation coefficient R2: 0.99. The study demonstrates the technical viability of integrating induction heating coil to a cyclone separator for wide temperature applications.