Abstract
Performance optimization using process parameters of an indoor air filtration system is a requirement that has to be established through experimental and analytical means for increasing machine efficacy. A closed casing containing a motor-driven blower is placed in a glass-encapsulated control volume. Air flows axially through an inlet filter and is thrown radially by the blower. In the radial path, air is treated with free radicals from the UVC-irradiated nano-TiO(2) coated in the inner wall of casing. A known quantity of Staphylococcus aureus bacteria is populated (Courtesy: EFRAC Laboratories) in the glass-encapsulated control volume. The bacterial colony count is measured at different time intervals after the machine is switched on. Machine learning approaches are applied to develop a hypothesis space and the hypothesis based on best R(2) score is used as a fitness function in genetic algorithm to find the optimal values of input parameters. The present research aims to determine the optimum time for which the setup is operated, the optimum air flow velocity in the chamber, the optimum setup-chamber-turning-radius affecting the air flow chaos, and the optimum UVC tube wattage, which when maintained yields the maximum reduction in bacterial colony count. The optimal values of the process parameters were obtained from genetic algorithm using the hypothesis obtained from multivariate polynomial regression. A reduction of 91.41% in bacterial colony count was observed in the confirmation run upon running the air filter in the optimal condition.