anthropometric man-model used for preparing DHM
simulations did not suit the measurements of all workers
in real-life tasks, which probably accounted for
some deviations in AAWS posture scores. Better congruence
could have been reached if the man-models
were exactly designed to fit each worker’s extremities
and torso proportions. DHM simulations were prepared,
however, with the 50th percentile model because,
in preproduction planning, it is not yet known
who exactly will later carry out the task on the real production
line. Therefore, one purpose of this study was
to analyze how well the workload estimations based on
DHM simulations with an average size worker model
matched real-life assessments with workers with random
(i.e., uncontrolled) measurements. Second, although
real-life MTM-time and DHM simulation time
were not significantly different in mean, some single
time differences were found. Although a number of
DHM simulations were longer than real-life tasks, because
manikin movements such as walking or turning
around were more time-consuming than in reality,
other simulations were shorter because of technical
limitations (e.g., cable installations had to be simplified
as it was not possible to simulate flexible parts). These
deviations might have accounted for small differences
in ergonomics risk assessment results because the duration
of physical workloads is one major factor determining
AAWS scores. Third, some differences between
real-life tasks and DHM simulations actually have to
be expected because “it is unreasonable to assume that
a [digital] man-model will give the same angle values
as people performing the task [in real-life], as each
person performs the task differently [and not necessarily
as planned] according to the standard procedure”
(Carey & Gallwey, 2002, p. 841). For these three reasons
a perfect congruence between real-life tasks and DHM
simulations could hardly be reached. Considering this,
the moderate to good correlations that were found in
this study can be judged as good results.
Of course this study also has some limitations. First,
standard work sequences based on MTM-time were
used as the basis for preparing DHM simulations. For
detailed comparisons with real-life tasks it would have
been better to use videos and actual real-life time instead.
Yet, when applied in early planning phases there
is also no video of the (future) workplace, only a standard
description of work sequences. Second, real-life
tasks already existed, which is not the case when DHM
simulations are applied for proactive ergonomics.
This has to be considered when interpreting results,