Presentation of the Melt-specific Creep Rupture Strength of Selected High-temperature-resistant Steels Based on Data Processing Via Neuronal Networks
Project Number 316
At relatively high thermal stress temperatures, steels exhibit time-dependent strain behaviour (creep). Accordingly, high-temperature components have to be designed on the basis of time-dependent parameters.
The long-time stress-rupture strength exhibits a scatter range of ±20 % in the direction of the stress. The reason for this spread of values lies in the microstructure of the molten material in each case.
This spread may result in an increase in the fracture time by as much as several 100 % (in the case of +20 %) or in a reduction of up to 80 % (in the case of -20 %) relative to the mean value. Both the component design and the assessment of the component’s residual lifetime are based on the lower limit of the long-time stress-rupture strength scatter range.
Within the scope of the preceding project 273 (AVIF 198), the potential presented by data mining tools (neuronal networks) in respect of the delimitation and quantitative assessment of microstructure-related variability in long-time stress-rupture strength parameters as applied to X20CrMoV12-1 martensitic steel was successfully and convincingly demonstrated.
In the course of this follow-on project, the knowledge acquired is to be applied to other materials and deepened on the basis of these data.
The following tasks are planned:
- Application of the findings to high-temperature-resistant steels that are already being used in new power station construction projects (P91 and P92, depending on the data available).
- Determination of the degree of uncertainty entailed by the method itself.
- The characteristic values are to be supplemented by additional quantities that, in the case of the 9-12% chromium steels, have an impact on long-time stress-rupture strength and are therefore suitable for strengthening the sensitivity of the neuronal network and improving its accuracy.
- Cross-link to Research Project 274 (Damage Development III, AVIF 229). The microstructure tests and component computations being carried out in the course of this project are to be backed up by the evaluations produced by the neuronal network and the different sets of results are to be compared.
The project is to be carried out between July 2008 and September 2010 under the coordination of the FDBR Research Foundation and will be headed by Dr. Karl Maile (MPA Stuttgart) and Dr. Wolfhart Müller (IMF Freiberg). It will be financed as Project Number AVIF 257 mainly from funds provided by the Research Association of the Arbeitsgemeinschaft der Eisen und Metall verarbeitenden Industrie e.V. (Association of the Iron- and Metalworking Industry / AVIF). The project will be supervised by Working Group W 12 (“Residual Lifetime”) of the Arbeitsgemeinschaft Warmfeste Stähle (High-Temperature-Resistant Steels Association). The investigations will be monitored on the part of the VGB by the Technical Committee
Materials and Quality Supervision