Assessment of the effect of the steels structure dispersion on its magnetic and mechanical properties

OBRABOTKAMETALLOV MATERIAL SCIENCE Vol. 23 No. 4 2021 Ta b l e 3 Regression statistics Multiple R 0.885 R -square 0.783 Normalized R -square 0.765 Standard error 0.014 Observations 14 Fig. 10. The value of the factor of different grain size Fz depending on the value of the coercive force for heat-treated specimens made of structural steels With a value of ± 2· ε , 95 % the data points are located within these de fi ned limits. The adequacy of the proposed linear regression model can be veri fi ed by examining the residuals of the model, which are determined for each X as: . i i i U Y Y = -  (4) The graph of the dependence of the residuals on the predicted values of Y is shown in Figure 11. There is a necessary condition that characterizes the adequacy of the analysed dependence for such graphs. It is the absence of characteristic “patterns” for the nonequilibrium distribution, depending on the Y values. For the dependence shown in the fi gure 11 in the location of the point cloud, there are no obvious patterns, which can tell us about the correctness of the found linear regression. Table 4 shows the values required for the regression analysis. The Y p coef fi cient corresponds to the Y value, provided that all parameters in the model are equal to 0. It means that the model does not introduce the other factors effects on the analysed parameters. X p1 shows the weightage of the parameter X over Y . The internal stresses within this model affect the uneven-grained factor with a weight of 0.0063. The sign before the number indicates the in fl uence exerted on the uneven-grained factor: the greater the internal stress, the greater the value of the uneven-grained factor. In addition, the values for the parameter Y at the intersection of the X axis with a con fi dence interval of 0.95 are presented.