ANN-GA approach for predictive modelling and optimization of NOx emissions in a cement precalcining kiln

Abstract

China’s cement production is increasing, but the sustainable development of the cement industry is hindered by pollutants, especially nitrogen oxide, with serious health impacts. This study reports the use of an artificial neural network and a genetic algorithm to control the operation parameters of a new dry-process cement kiln technology to predict and optimise the nitrogen oxide emissions. Comparing the predicted value and actual value, the error of the model is less than 2%. The GA is used to search the optimal operation parameters to achieve the lowest concentration of nitrogen oxide emission, which is 165.9 mg/m³ under optimal conditions. The results of sensitivity analysis show that the furnace temperature, raw material quantity and third air temperature have the greatest influence on the nitrogen oxide emission. This model prediction and optimisation can provide a reference for enterprises in controlling operation parameters to reduce nitrogen oxide emission.

Keywords:

• Precalcining kiln
• NOx emission
• ANN–GA