Prediction of the lime availability on an industrial kiln by neural networks

Neural networks are used to predict the lime availability on an industrial kiln for quality control. For this purpose, a predictive empirical model of the highly nonlinear relationship between important variables such as the kiln temperatures and the residual calcium carbonate, at the discharge end, is constructed using neural networks. It allows us to predict and monitor lime properties from kiln operation simulated data. With the help of neural networks the quality control of the industrial unit is achieved more quickly and is extremely cost effective. These capabilities further strengthen the kiln operator's decisions leading to energy savings and increased production of first-quality materials in the pulp manufacturing process.