Regression analysis for tensile strength of powder metallurgy low alloy steels
FU Jiaqian1.，YU Yongliang2.，ZHANG Dejin2.，LI Songlin1.
（1.State Key Laboratory of Powder Metallurgy, Central South University, Changsha 410083, China； 2.Shandong
Luyin New Material Technology Co., Ltd., Laiwu 271105, China）
Abstract：Powder metallurgy low alloy steels are an important component of iron-based powder metallurgy mate‐
rials, which are widely used in structural components. With the development of the powder metallurgy industry
towards high-density, high-strength and complex-shaped powder metallurgy products, higher demands have been
placed on the mechanical properties. The influence of chemical composition and related manufacturing process
parameters on the mechanical properties of powder metallurgy low alloy steels is still unclear. The purpose of this
study is to investigate the influence of various factors on the mechanical properties of powder metallurgy low alloy
steel, to obtain a relationship model between chemical composition and manufacturing process parameters and the
tensile strength, and to achieve effective prediction of material properties. Fe-Ni-Cu-C powder metallurgy low al‐
loy steels with LAP100.29 water-atomised iron powder as the base powder were prepared after pressing and sinter‐
ing, tensile strength was tested and a tensile strength database was established. A multiple linear regression model
was established for the tensile strength of the powder metallurgy low alloy steels with pressing pressure, carbon
content, nickel content, copper content and iron powder purity as independent variables. The regression model was
tested for significance to determine the validity of the model and the performance of the regression model was as‐
sessed by the coefficient of determination and the mean relative error. Specimens of Fe-1Ni-1Cu-0.3C, Fe-1Ni-
1Cu-0.5C and Fe-1Ni-1Cu-0.8C powder metallurgy low-alloy steels were prepared under the same experimental conditions for tensile strength testing and metallographic observation to validate the multiple regression model.
The multiple linear regression model for tensile strength is TS=-334 652.22+0.34 P+165.96 C+12.76 Ni+32.42 Cu+
3 358.06 Fe. Significance tests show that the linearity of the model is significant and that the linear relationships
between each of the variables in the model and tensile strength are significant. The R2 of the model was 0.873 and
the R 2
adj was 0.863, indicating a good fit of the model. The mean relative error of the predicted values of tensile
strength values is 6.39%. The mean relative error between the predicted and experimental values of the validation
system is 4.43%, which is less different from the mean relative error value of the regression model. The multiple
linear regression model has a good fit and high prediction accuracy, which can be used to effectively predict
the tensile strength of powder metallurgy low alloy steels. More influencing factors present in the powder metallur‐
gy process can be taken into account to improve the mechanical properties database and establish a multivariate
mechanical properties model.Applying it to actual production can reduce the number of experiments and tests, pro‐
viding a practical and effective solution to the problems of long product development cycles, low efficiency and
high cost.
2023, Vol. 33 
