Regression Analysis of Thermal Conductivity Based on Measurements of Compacted Graphite Irons

摘要

A model describing the thermal conductivity of compacted graphite iron (CGI) was created based on the microstructure analysis and thermal conductivity measurements of 76 compacted graphite samples. The thermal conductivity was measured using a laser flash apparatus for seven temperatures ranging between 35 °C and 600 °C. The model was created by solving a linear regression model taking into account the influence of carbon and silicon additions, nodularity, and fractions of ferrite and carbide constituents. Observations and the results from the model indicated a positive influence of the fraction of ferrite in the metal matrix on the thermal conductivity. Increasing the amount of carbon addition while keeping the CE value constant, i.e., at the same time reducing the silicon addition, had a positive effect on the thermal conductivity value. Nodularity is known to reduce the thermal conductivity and this was also confirmed. The fraction of carbides was low in the samples, making their influence slight. A comparison of the thermal conductivity values calculated from the model with measured values showed a good agreement, even on materials not used to solve the linear regression model.