Electroslag remelting is widely used to produce various special steels, mainly because of its ability
to provide high cleanliness and excellent homogeneity of ingot. Undoubtedly an optimal slag plays
a key role in removing inclusions and enhancing the mechanical properties of a specific special steel.
In order to improve the metallurgical performance of the slag in electroslag refining of die steel, a
variety of slags have already been developed using traditional trial-and-error methods. Nonetheless,
these slags may not be inherently optimal for the specific requirements of actual die steel refining
processes. In this work, we have designed a CaF2-CaO-Al2O3-SiO2-MgO quinary slag system using
genetic algorithms method aiming to remove the existing large size inclusions (D and Ds type) in
high quality die steel. The comprehensive effects of five components (CaF2, CaO, Al2O3, SiO2 and
MgO) were studied based on existing fundamental theory. The slag composition was screened
considering the parameters such as melting point, electrical conductivity, calcium ion concentration,
and appropriate viscosity etc. Furthermore, the candidate optimal slag was applied in the industrial
production of die steel by electroslag remelting method. The detection results indicate that majority
large inclusions (D and Ds type) have been successfully removed, showing significant advantage to
traditional slags. This result has demonstrated the genetic method to be efficient and reliable. In
addition, depending on the personalised demand of other steel and alloys, this method allows for the
adjustment of the screening frameworks and processes to develop the candidate optimal slags. This
method has not been reported previously in the field of electroslag composition design, and
compared with the traditional trial-and-error method, the new method will save a lot of time and
experimental costs.