To get design values (the strength of something, for example) designing to Australian/ New Zealand loading code you do $n$ tests and estimate the coefficient of variation of your parent population, $V_{sc}$,). With that, you get $k_t$ from lookup table below and divide you lowest test value(arguably ought to be the mean?) by $k_t$. enter image description here I have tried to work out what these $k_t$ values are - where they come from. Getting nowhere.

I think it 'works' on the principle that you want $\gamma$ confidence that the estimated strength, $R_{p,cover}$, covers the parent population $p$-fractile, $R_p$, so $P(R_{p,cover} < R_p) = \gamma$ NOTE: I am precising from an old uni text book!

So, assuming that the parent population mean equals test population mean, $\bar{R}$

Parent population Standard deviation, $s$

COV, $V_{sc}=s/\bar{R}$

$R_{p,\gamma}=m+\kappa(p,\gamma).s$, so $ x_{p}=\bar{R}(1+\kappa(p,\gamma).V_{sc})$, so:

$k_t = \bar{R}/R_{p,\gamma} = \frac{1}{1+\kappa(p,\gamma).V_{sc}}$

The problem I am having is what an earth $\kappa(p,\gamma)$ is, to match the values in the lookup table. I do not know what $p$ and $\gamma$ are, though I have an inkling that they are 0.05 and 95% respectively.

I am working in R, gist here


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.