TC20 Inclusions

Co-Chairs	Prof Y Murakami	Kyushu University, Japan Contact E-mails: ymura@mech.kyushu-u.ac.jp
Co-Chairs	Prof CW Anderson	University of Sheffield, UK Contact E-mails: c.w.anderson@sheffield.ac.uk
Secretary	Prof S Beretta	Politecnico di Milano, Italy Contact E-mails: stefano.beretta@polimi.it

TC13 UPDATE 2006

MINUTES OF TC20 MEETING

Meeting began at 10.00 and after the particpants introduced themselves there was a technical presentation by E. Henault entitled “Statistical analysis of non metallic inclusions for the estimation of rolling contact fatigue range and quality control of bearing steel”.

The discussion touched the following points: i) interest in this kind of simulation for addressing the open problem of the number of areas; ii) sensitivity of results to what is taken as the objective of the study; iii) realistic choice of the initial distribution for the simulation; iv) combination of measurements.

Then there were two contributions describing the results so-far obtained about the first round-robin: Beretta showed the results obtained by the different participants which show similar results (except for two participanst whose measurement method has affected the results). Anderson discussed the consistency of results by the different labs results firstly by pair-wise comparisons among the samples, then he showed that adopting a Fisher correction for multiple testing there is no significant difference between laboratories. Another more discriminating approach was based on likelihood testing: results form different labs share the same common but reveal differences in the location parameters; no practical difference for fatigue limit; evidence of curvature from some laboratories.

The following discussion can be so summarised: importance of trying of recognizing different population of particles; present results (considering that the amount of material tested is small and small differences are expected) suggest that for comparisons of steels measured in different labs the between-lab variability should be taken into account; another problem is the correct definition of measurements when the particles have a mixed composition. Girodin recalled that the in-batch variability can be relevant and this is not covered by the round-robin.

The discussion of the second round-robin on the bearing steel was introduced by a technical presentation by Girodin entitled “Statistical analysis of non metallic inclusions for the estimation of rolling contact fatigue range and quality control of bearing steel”.

E. Henault summarized the set-up of the second round-robin onto a 100Cr6 bearing steel and he showed the results by Ascometal. Then Butler showed his measurements onto the bearing steel emphasizing the difficulty of recognizing the different types of particles and different types of oxides. Then the discussion focused on the treatment of samples and the different type of particles recognised during image analysis.

Some action have been agreed for improving RR1: i) replication of measurements by some of the RR1 participants in order to estimate the within-lab variability; ii) circulation of a set of images for comparing the different elaboration by RR1 participants; iii) results of RR1 will be backed with a set of simulations by E. Henault; iv) theoretical comparison; v) fatigue tests for estimating the size of extreme defects.

The actions considered for RR2 are: i) record as much details as possible about particle measurement; ii) comparison of the particle identification.

The main activity of TC20 in near future is to publish by the end of 2006 a paper which summarises the results of the first round-robin. Cirilli then proposed to apply for 2006 to a European Project with the aim of finding a financial support to future TC20 activities.

Next meeting will take place in Milan on 13 September 2006 because of the presence of Murakami in Europe for ‘Crack Path’ Conference in Parma.

Present results of RR1 will be posted on TC20 website.