Kartläggning av påverkande faktorer vid härdning och anlöpning i värmebehandlingen

This thesis explores hardness variation in two tool steels, H13 and a chromium-alloyed stainless holder steel, based on industrial heat treatment data. The aim was to identify which process or material parameters most affect final hardness after hardening and tempering. Nearly 400 data entries were analyzed using two approaches: graphical correlation and a Random Forest regression model. The regression analysis showed weak linear relationships, with the highest R² value (0.547) for tempering time in H13 round bar. Alloying elements had limited influence across all datasets. The Random Forest model revealed that hardening and tempering times were consistently the most important variables. These results highlight that thermal parameters play a greater role than variations in chemical composition in predicting hardness. The study concludes that a multivariable approach is essential to understand hardness behavior in tool steels and can support improved heat treatment strategies in industrial practice.