Guest post – WoodSimulatR

This is a guest post from Andreas Weidenhiller from Holzforschung Austria.

Simulating values from destructive strength testing – the R package WoodSimulatR

Holzforschung Austria recently concluded the InnoGrading Project which focused on new machine strength grading methods in a multi-stage grading context (more about this will follow in a later article). InnoGrading followed an Open Innovation approach, where important research methodology should be shared with the public, in the form of extension packages to the open source statistics software R. The first fruit of this approach, the R package WoodSimulatR, has been released on the Comprehensive R Archive Network (cran.r-project.org) this November.

Research is foundationally dependent on data, data of destructive strength tests of sawn timber in the case of machine strength grading research. Such tests are expensive and, therefore, strength data is a limiting factor in our research. On the other hand, the industry extensively use modern strength grading machines which are capable of non-destructively measuring thousands of boards every hour.

For at least a decade, some researchers have been using such data from non-destructive testing (NDT) to simulate data from destructive strength tests in one way or another, but many have been intimidated by the complexities of simulating multiple correlated variables. The R package WoodSimulatR is now offering a unified, extensible and simple to use framework for this kind of simulations.

This framework is generic — you can use it to add any kind of simulated values of a correlated variable to an existing dataset as long as you have the information about the correlation with some variables already existing in the dataset, and as long as all involved variables are continuous.

Further, the package offers the option to create a dataset of correlated variables from destructive and non-destructive testing from scratch, simulating different origins of the material, with different pre-defined means and standard deviations, or even different correlations.

The WoodSimulatR package also includes a so-called vignette — in the R world, this is a kind of long form documentation in the style of an article, used to comprehensively explain important aspects of a package. This vignette explains how to simulate entire datasets and how to add simulated values from destructive strength testing to values from NDT measurements. The vignette can be read online at https://cran.r-project.org/package=WoodSimulatR.

If you already have R installed on your computer, adding the WoodSimulatR package is easy: just type install.packages(“WoodSimulatR”) into your R console while you are connected to the internet.


Comment from CWST: This is an exciting new addition to our wood research tool kit and we look forward to making good use of it. We will be posting some examples of WoodSimulatR use on this blog, including information you will need for simulating UK and Irish grown timber.

Be the first to comment

Leave a Reply

Your email address will not be published.


*