Control of a novel synthetical index for the local indoor air quality by the artificial neural network and genetic algorithm

Abstract

Abstract Based on the numerical simulation technology, a novel synthetical index of local indoor air quality (IAQ) was put forward. It can represent the ability of current air distribution in breathing zone to resist pollutant. The continuous and transient released HCHO and PM2.5 were selected as specific pollutants, and the impact of pollutant source location on the local IAQ was considered in the synthetical index. The synthetical indexes for the continuously released HCHO (Pc), transiently released HCHO (Pc-t), continuously released PM2.5 (Pp), transiently released PM2.5 (Pp-t) were obtained in two typical indoor models. The results showed that better local IAQ was obtained in the up-inlet and up-outlet ventilation model compared to the up-inlet and down-outlet ventilation model. The artificial neural network (ANN), genetic algorithm (GA) and integration of these two methods were separately used to control the local IAQ with two typical ventilation models. The ANN method showed higher efficiency when there was obvious correlation relationship between the control variable and objective. When the obvious correlation relationship was absent, the integration of GA and ANN showed more efficient than GA alone, and more accurate than ANN alone.