Dynamic modelling of BOF process : comparison of model performance with the plant data

In a top blowing steel making process (BOF), the refining reactions of C, Si, Mn and P are related to several physicochemical phenomena occurring in different zones (e.g. slag-metal emulsion, jet impact zone, slag-bulk metal zone) inside the converter. In earlier publications, the authors have demonstrated a dynamic model, based on the fundamental approach of multiple zone reaction kinetics to simulate the refining of elements (C, Si, Mn and P) in a top blowing steelmaking process (BOF).[1,2] After successful model validation with the literature data from a 200 ton converter[3], simulations have been carried out to assess the model performance with the plant data. Off-line heat data obtained from a 330 ton converter at Tata Steel, Netherlands was used for the model validation. The BOF shop in Tata Steel operates with a wide range of operating and process conditions such as (i) different scrap mix, (ii) dynamic flux addition strategy, (iii) dynamic change in lance position and (iv) top and bottom flow rate. The model predictions of hot metal impurities were validated with the two sub-lance measurements and the simulated slag compositions were compared with the end blow measurements. The possible effect of the uncertainties associated with the measured (or reported) input variables in industrial conditions on the accuracy of the model calculations has been investigated. Further, the role of dynamic change in lance height and flow rate on the slag formation and the hot metal refining has been studied. The present study identifies the critical input variables required for the accuracy in the prediction of the dynamic model in plant conditions and provides a fundamental understanding to control the dynamic process variables in a BOF operation.