EN 14081-2 update (machine grading)

The much needed revision of the standard for machine grading settings, EN 14081-2, progresses.  Much needed, because the current version is literally unusable – except by a small number of specialists who know how to interpret the standard, in combination with the “TG1 decision list“.  But also needed because it contains a some new elements that should improve the way timber grading can be done:

i) A more sensible treatment for defining the geographical region over which settings can be applied (now called “settings areas”) – especially useful for very small and very large countries.

ii) The addition of an “adaptive settings” approach where the continuous monitoring of grading machine indicating property (IP) can be used to adjust the grading settings (currently, modern grading machines keep this information – which is very useful for spotting shifts in incoming timber quality – but no use is made of it for grading).

We recently blogged on the comments that came from the last (and second) Enquiry stage, which was passed with positive vote (just).  In October, CEN TC124 WG2 met to discuss those comments and amend the draft standard accordingly. It will shortly go to Formal Vote (a stage at which only editorial comments can be considered). There is some haste to do so – to maintain compatibility with the revised EN 14081-1 (hoping it will eventually be cited in the OJEU), and the amendment to EN 14081-3 (which made big changes to how output control is done).  Worth noting that EN 14081-3 will need to be amended again right away – since it currently points to the Enquiry draft of EN 14081-2.

In summary, the main changes made to prEN 14081-2:2017 (2nd Enquiry version) for FprEN 14081-2 (the version that will go to Formal Vote) are as follows:

Validity of existing settings

Machine settings that were derived in accordance with previous versions of the standard remain valid unless there is evidence that the required grade determining properties are not met.  It is only in that special case that new settings need to be derived according to the new version of the standard.

This differs from the previous text that said “Machine control setting that have previously been approved will need to be reviewed to ensure that the basis of the settings complies with the requirements of this European Standard where the settings area covers more than one country or standardized area” – something that was a problem for settings covering more than one country, approved prior to October 2012 (see here for more).

That a revision of EN 14081-2 alone could actually require recalculation of existing settings was, anyway, doubtful – especially with text in the (not-normative) introduction.

If settings are suspected to be producing timber that does not meet the required properties then there is clear and obvious need to check and, if necessary, recalculate the settings.  This follows through from the requirement in EN 14081-1 that settings need recalculating “whenever a change occurs in, the raw material, or in the method of production…, which would affect significantly one or more of the essential characteristics”.

The new version of EN 14081-2 includes a procedure for a local checking of settings, but the circumstances under which settings for a whole growth area would need recalculating are not entirely clear.

Fortunately, nobody suspects that any of the previous settings are unsafe, and since October 2012 the cases that may have been, were caught by the “country check” on the TG1 decision list (this date is marks the time at which much larger grading areas started be considered, following the Gradewood project).

The cost matrix

The cost matrix is a tricky procedure to calculate – but is considered by some to give some useful information about the performance of the grading machine.  It had almost been removed from the standard – changed into an additional machine verification procedure to determine the number of grades a machine is allowed to grade in combination.  There were only two problems with this new role for the cost matrix:
1) a way could not be found to make this a useful procedure
2) the standard applied the 0.2 limit to both the “incorrectly upgraded” timber and the “incorrectly downgraded” timber (instead of just the “incorrectly upgraded” pieces as now). This made the check impossible to pass.

So the cost matrix will revert back to the method in current use (a calculation made for every setting), but with a modification to permit greater flexibility in balancing yields between grades in production.

The cost matrix was commonly a problem for settings we calculated because we usually wanted to reduce reject rather than get maximum yield in the highest grade.  The greater flexibility is achieved by raising the limitation for “incorrectly upgraded” pieces from 0.20 to 0.40.  The term “incorrectly upgraded” is hugely misleading as they are only “incorrect” under the special circumstance of wanting maximum yield for the highest grade.  Relaxation of the cost matrix requirement does not cause any relaxation in the requirement for graded timber to meet the required characteristic values.

Procedures for the cost matrix are now grouped in Annex C and the requirements placed as a new step 3 in subclause 6.4.2.2 (verification of new settings).

Readability

Several editorial changes were made to improve the readability of the standard. They include creation of a new table dealing with calculation and verification of characteristic values (in subclause 6.4) to summarise the requirements.

Adaptive settings

Questions were raised about the adaptive settings procedure – its basis, and whether it has been shown to be safe or not.  Such evidence and justification cannot fit within the standard itself, so an explanatory background document will be distributed to justify the approach.

It is recognised that EN 14081-3 will need an improvement to consider factory production control for adaptive settings – but it was decided to keep this new adaptive settings approach in the revision since it is a clear step forward in making use of IP records.

Comparison with the existing procedure

There are understandable concerns that the new standard might have significant implications for grading yields for new settings – and that this might be unfair for new settings (new machines, countries newly adopting machine grading, and new timber species).  A full comparative study would be very costly – but spot checks have been done (including by us) that indicate that the changes introduced in the standard being put forward to Formal Vote do not have significant impact on machine settings compared to the procedures currently in use (the current version of EN 14081-2 and TG1 decision list).

The first draft for a new EN 14081-2 certainly was very different to the current procedure, but the version we have now is very similar.  The main difference will arise from the requirement to apply confidence adjustment when the number of specimens is very small (less than 40) – which will mostly affect settings that cover several countries (and do not have many specimens in each country).

The UK should note that the stiffness requirement for the country check is raised from >90% to >95% of the target stiffness, but this is not problematic in the vast majority of cases we have checked.