用m个试样的测值,构成求k方程的差方和,将其对n个k值求极小(m>n),即令?sum from j=1 to m(M_j-sum from i=1 to n I_(ijk_s~i))~2/?k_s~i=0回归得出n个求k方程的各项系数.由△k=△_I~(-1)*△_M矩阵运算式解出n个k值,进而求得n个未知相量.利用回归降低总误差,以期克服多元方程求解过程中误差叠加的困难.方法应用于W_9Mo_3Cr_4V高速钢析出相的定量分析.
The sum of squares of deviations of the equation for calculating kwas constructed using the measured data of m samples. By deriving the minimumof the sum of squares of deviations of n equations, i.e. let ?sum from j=1 to m(M_j--sum from i=1 to n I_(ij)k_s~i)~2'/?k_s~i=0,the regressive coefficients of the n equations can be obtained (m>n). The k valuescan be deduced from the matrix equation Δ_k=Δ_I~(-1), Δ_M, and the n unknown phasesmay be calculated subsequently. It is expected that the trouble of error expansionduring solving the polynomial equation might be overcome by minimizing the sumerror through regression. The above method has been used to analyse precipitatesin high speed steel W_9Mo_3Cr_4V.
参考文献
| [1] | |
| [2] | |
| [3] | |
| [4] | |
| [5] |
- 下载量()
- 访问量()
- 您的评分：
-
10%
-
20%
-
30%
-
40%
-
50%