The present work aims to demonstrate an innovative way of application of modern data analysis tools for structural power plant materials, such as 9%Cr-steels. The purpose is the optimized prediction of the material behaviour considering the stress/time to rupture data and the determination of position of the specific melt in the scatter band for the given steel. This data is needed for an effective life time assessment of power plant components. The material behaviour is influenced by the multidimensional interdependencies between the individual elements of the chemical composition, the heat treatment parameters, product form, tensile properties and microstructure, which are difficult to describe using simple analytical methods. Modeling with neural network techniques therefore seems to be an interesting alternative. Moreover, the applied method takes away the requirement for long and expensive experiments. A large variety of data was used for training of a commercial neural network. By means of sensitivity studies the influence of specific data features was investigated with regard to an optimized correlation factor r. The interpretations of the neural network have been checked whether basic physical and metallurgical backgrounds are reflected. The results are compared with the real material behaviour gained by material tests. The applicability of the neural network tool for technical use in life time assessment was investigated.
[1]
John Hald,et al.
Behaviour of Z phase in 9–12%Cr steels
,
2006
.
[2]
Petra Perner,et al.
A Comparison between Neural Networks and Decision Trees
,
1999,
MLDM.
[3]
H. Bhadeshia,et al.
MICROSTRUCTURAL STABILITY OF STRONG 9 – 12 wt % Cr STEELS
,
2022
.
[4]
Petra Perner,et al.
A comparison between neural networks and decision trees based on data from industrial radiographic testing
,
2001,
Pattern Recognit. Lett..
[5]
B. Melzer,et al.
Verbesserte Lebensdauerabschätzung kriechbeanspruchter Rohrbögen mittels bauteilspezifischer Kennwerte
,
1993
.
[6]
Bruce E. Barrett.
Regression Analysis: Concepts and Applications
,
1994
.
[7]
Franklin A. Graybill,et al.
Regression Analysis-Concepts and Applications
,
1995
.
[8]
D. H. Hellmann,et al.
Feature Selection for a Real-World Learning Task
,
2001,
MLDM.
[9]
H. K. D. H. Bhadeshia,et al.
Design of Ferritic Creep-resistant Steels
,
2001
.