ADAPTIVE NEURAL-NET (ANN) MODELS FOR DESULFURIZATION OF HOT METAL ANDSTEEL

Neural nets can be adapted to complex patterns of interrelated input a nd output variables in a process even if the data sets contain a lot o f noise. In this work two specific examples for the application of ada ptive neural nets (ANN) in steel industry are described. First, the su lphur content of hot-metal, obtained at the end of calcium carbide pow der injection into 400 t torpedo ladles is predicted as a function of hot-metal weight, treatment time, initial sulphur content, gas flow ra te and powder injection rate. The values predicted by the trained ANN model for a completely new set of input test data compare well with th e actual values obtained on the shop floor. In the second example, the sulphur content of steel, obtained at the end of blow is predicted as a function of liquid-metal weight, total amount of oxygen injected, a mount of iron ore added, and the temperature, contents of carbon, mang anese, phosphorus and sulphur determined by in-blow sampling in a 300 t converter. The ANN predicted values of sulphur content of steel at t ap (without reblow) also agree well with the values obtained on the sh op floor.