Predicting the vibration discomfort of standing passengers in transport

Abstract

It has previously been assumed that the vibration discomfort of standing people can be
estimated using the same procedures developed from for seated people. In this thesis, the
discomfort of standing people exposed to vibration was investigated to improve understanding
of the mechanisms responsible for discomfort and construct a model that may be used
to predict the discomfort of standing railway passengers.
The first of five experiments using the method of magnitude estimation and 6-s periods
of vibration investigated how the discomfort of standing subjects exposed to fore-and-aft,
lateral, and vertical sinusoidal vibration depends on the frequency of vibration. From the
judgements of 12 subjects at each of the 16 preferred one-third octave centre frequencies
from 0.5 to 16 Hz, frequency weightings were constructed for each direction. For vertical
vibration, the weighting was similar to that recommended in standards, but the weightings
for fore-and-aft and lateral vibration differed from that previously assumed. Horizontal
vibration caused loss of balance at frequencies less than about 3 Hz, and it caused discomfort
in the legs at higher frequencies. Vertical vibration caused discomfort in the upper body. To
adjust the frequency weightings according to differences in sensitivity between directions,
the second experiment with 12 subjects compared the discomfort caused by 4-Hz sinusoidal
vibration in the fore-and-aft, lateral, the vertical directions. It was found that sensitivity
was greater for fore-and-aft vibration than lateral vibration at frequencies less than 4 Hz
and weightings were determined to assist the evaluation vibration in all three directions.
The third experiment investigated the extent to which postural supports used by standing
train passengers (vertical bar, shoulder support, and back support) affect discomfort caused
by fore-and-aft and lateral vibration in the range 0.5 to 16 Hz. Supports that created a
new path for the transmission of vibration to the upper-body increased discomfort over the
range 4 to 16 Hz.
The fourth experiment investigated how the root-mean-square method, the basic evaluation
method in current standards but known to underestimate the discomfort caused by
motions containing occasional peaks, could be modified for the evaluation of non-sinusoidal
vibration. Using 1-Hz and 8-Hz random vibrations with a range of crest factors it was found
that the discomfort of standing subjects was better predicted with an exponent around 3,
rather than an exponent of 2 implicit in r.m.s. averaging. The final experiment determined
a method for predicting the discomfort of tri-axial vibration. The cube root of the sum
of the cubes of the discomfort caused by the single-axis components gave good estimates
of the total discomfort for both 1-Hz and 4-Hz tri-axial vibration. Since it was found in
the first experiment that the discomfort was generally proportional to the acceleration at
the power 0.7. these results suggest that the root-sum-of-squares of the accelerations gives
good estimates of the total discomfort for tri-axial vibration .
The results of all experiments were combined in an empirical model for predicting the
discomfort of standing people exposed to 6-s periods of vibration. It is concluded that there
are two distinctly different mechanisms responsible for vibration discomfort when standing:
postural instability and body vibration. Postural instability is dominant with horizontal
vibration at frequencies less than about 3 Hz, whereas body vibration is dominant with
vertical vibration and with horizontal vibration at frequencies greater than about 3 Hz.
The discomfort of standing people is similar to the discomfort of seated people for vertical
vibration, but fundamentally different with horizontal vibration due to postural instability
at low frequencies and vibration attenuation in the legs at higher frequencies

More information

Identifiers

ORCID for M.J. Griffin: orcid.org/0000-0003-0743-9502

Catalogue record

Export record