Ozone pollution prediction around industrial areas using fuzzy neural network approach

This paper presents the prediction of ozone pollution as a function of meteorological parameters including wind speed and direction, relative humidity, temperature, solar intensity, concentration of primary pollutants consisting of methane, carbon monoxide, carbon dioxide, nitrogen oxide, nitrogen dioxide, sulfur dioxide, non-methane hydrocarbons, and dust around the Shuaiba industrial area in Kuwait by a fuzzy neural network (FNN) modeling approach. A subtractive clustering analysis was performed for the input data to produce a concise representation of the system's behavior leading to the minimum number of rules. In addition, Sugeno–Takagi–Gang fuzzy inference and hybrid algorithm were used to prepare the FNN system. It is perceived that the FNN model is more accurate and reliable than artificial neural network model to forecast the pre-mentioned concentration. Finally, sensitivity analysis was applied. It was found that temperature, solar radiation, and relative humidity are the dominant parameters affecting the ozone level.