next up previous contents
Next: Conclusions Up: Growth Model Previous: Programming

Results

The results consist of samples from the posterior distribution, together with kernel density estimates for the reliability R(N).

  figure868
Figure: A Typical Trace of Samples for tex2html_wrap_inline2431 .  

  figure873
Figure: A Typical Trace of Samples for m.  

  figure878
Figure: A Typical Trace of Samples for D.  

  figure883
Figure: Kernel Density Estimate for R(N) at 140kN. tex2html_wrap_inline3025  

  figure889
Figure: Kernel Density Estimate for R(N) at 160kN. tex2html_wrap_inline3029  

Figures gifgifgif show typical traces for the MCMC samples from the posterior distributions of tex2html_wrap_inline2431 , m, and D. These traces are from a single run of the chain for single cracks. It was necessary to draw samples from the posterior distribution for each of the crack specific parameters, that is for each crack in the specimen in order to calculate the kernel density estimates which are provided in Figures gif and gif. These kernel density estimates are approximations to the posterior distribution for R(N) for each of the sets of data, that is tex2html_wrap_inline3039 , as outlined in Section gif.



Cathal Walsh
Sat Jan 22 17:09:53 GMT 2000