Civil Engineering Reference
In-Depth Information
In the SHARP Study, it is not possible to perform accurate vibration measurements while also collecting a
large amount of other physical exposure parameters for each of the subjects. The alternative is to obtain
the declared vibration values of the tools that the workers use, and then perform a time study on the
video recordings to measure the actual time that the vibration tools are activated. This estimation
may not reflect the real vibration exposure of the workers, as other factors such as tool balancing,
work surface conditions, and individual work techniques can influence the true vibration level, but it
can give an estimation of the vibration exposure.
44.3.4 Posture Analysis, Event Based vs. Time Based
Posture analysis is based on observations of recorded tasks as well. There are two types of posture ana-
lyses, “event based” and “time based.” For “event-based” posture analysis the overall postures (the most
common posture and the worst posture) for the different body parts when performing a specific task are
determined. In “time-based” posture analysis, postures are measured at a particular time for a specific
task. There are two types of analyses for the time-based posture analysis: (1) continuous observation
and (2) time-sampled observation.
In the continuous observation, the analyst observes the postures of the different body parts continu-
ously, and marks down the changes whenever the body part moves from one predefined angular category
to another. This analysis allows the determination of the distribution of the angles of the different body
parts and the movement frequency between the different predefined angular categories. However, this
type of analysis is very time consuming, particularly when there are several predefined angular categories
for each body part. Also, it is very difficult to observe several body parts simultaneously. Therefore, the
analyst must play the video several times in order to complete the analyses for the different body parts. In
the time-sampled observation of the time-based posture analysis, the analyst observes the postures at a
number of preselected times during the task performance, and a distribution of the postures is calculated.
Although this method significantly reduces the data processing time, it is not possible to obtain inform-
ation about repetitive movements of the different body parts in combination with the postures.
In the SHARP Study, both event-based and time-based posture analyses are used. One of the reasons
for using both methods is to compare results obtained by the two methods. However, due to the large
amount of data processing and analysis in this project, only the time-sampled observation method is
used for the time-based posture analysis, rather than the continuous observation approach.
In the event-based posture analysis, predefined postures are used (Table 44.2). Posture distribution
results can then be calculated and used for epidemiological modeling. An illustrative distribution
result is shown in Figure 44.5, where job A seems to have more wrist extensions and flexion compared
to job B and workers at job B maintain more neutral wrist postures compared to workers at job
A. Depending on the need of the epidemiological analysis, some of the predefined angular categories
may be consolidated.
Using the event-based posture results, one can also calculate certain indices for the different body
parts, such as the RULA scores. To obtain the final RULA score, apart from the posture results, additional
information such as forces and muscle use should also be obtained. Details on the computation of RULA
scores can be found in relevant articles (McAtamney and Corlett, 1993).
In the time-based posture analysis, postures of the same body regions used in the event-based posture
analysis are measured. However, instead of giving overall estimated posture values for an entire task
(event), postures are estimated at certain points of time during a task performance. In the SHARP
Study, postures are estimated for numerous randomly selected frames during a 15-min task recording
(75 frames for a single-task job, 80 frames for a two-task job, 90 frames for a three-task job, and 100
frames for a four-task job). To lower individual analyst variation, the frames are assigned to two analysts
for processing.
One of the potential problems with predefined angular categories is that the analyst may be biased by
the nature of the job. For instance, when a posture is on the threshold of two predefined angular
categories (e.g., posture is approximately 30
8
while the categories are 0 to 30
8
and 30 to 60
8
), the
Search WWH ::
Custom Search