| Summary: | 碩士 === 國立成功大學 === 環境醫學研究所 === 86 === Occupational exposure to heat stresses has been already known to
be dangerous to workers. To date, although more advanced
techniques have been widely adopted by industries, however,
workers might still suffer to heat exposures in some specific
environments. To protect workers from ill-health effects caused
by heat stresses, the government in Taiwan has promulgate the
PELs in 1974, which was mainly modified from the ACGIH TLVs.
However, the modification did not have any physiological basis
at that time. Considering the existence of differences in race
(Asian versus Caucasian) and living environment (humid and hot
versus dry and cool), the feasibility of the current PELs for
heat stress required further evaluations, especially on a
reasonable physiological basis. In this study we have tried to
build up heat exposure assessment predictive models which mainly
based on the Taiwanese measured physiological data. The
obtained predictive models were also used to compare with and
evaluate the predictive models published by ISO on the document
of ISO-7933.In this study, we found the predictive values of
Cres and Eres obtained from ISO-7933 suggested models were
significantly different from the theoretical values, which might
mainly due to the existence of racial difference between
Taiwanese and Caucasian. For this reason, two experimental
predictive models for Cres and Eres were developed in this
study, which were directly developed based on the measured
physiological data. Due to the predictive model for skin
temperature proposed by ISO was limited to specific
environmental conditions, the predicted values were also found
to be significantly different from the measured skin
temperatures obtained from workers under wider environmental
situations ( including temperature, wind velocity, and humidity,
etc. ). Through multiple regression analysis, a proposed
predictive model with 5 independent variables was proposed in
this study for skin temperature predictions in the future. For
metabolic predictions, the predictive model suggested by ISO
shows that the predicted values which based on heart rate
measurements did not consistent with the measured values. The
relationship between heart rates and metabolic rates was further
analyzed directed based on the measured values in this study,
the obtained relationship not only has a higher R2, but also
shows a more consistent trend. Finally, this study also shows
that the predicted allowable exposure time suggested by ISO-7933
might lead to under-estimated situations as compared to the
theoretical values, however, the predicted values obtained from
modified predictive models do show a batter trend as used to
predict allowable exposure time. It should noted here that the
trend mentioned above doesn''''t have a good correlation with the
experimental data, it might be due to the existence of
experimental outliers because of experimental errors or
intrinsic variations of tested subjects. Considering the
residuals of the regression results, it is recommended in the
future that the weighted-least-squares could be adopted for
regression analysis to eliminate the effects caused by outliers
in the future.
|
|---|
