Soft sensing model of air leak status in grate during iron ore oxide pellet production

Aiming at improving the existing methods of air leak measurement, soft sensing models based on gas temperature difference and gas balance were established to detect the location of air leak anomaly in grate by characteristic analysis. Models were developed by simulating temperature field and gas balance to provide data for the soft sensing models. Software was developed and put into practice in a certain domestic pellet plant. The results show that the hit rate of air leak diagnosis is over 95%. And compressive strength of each product pellet is 86 N higher. First grade rate increases by 2.54%, operating rate of the device increases by 1.1% while coal consumption is 1.2 kg/t low.