A Survey of Work Measurement
Techniques
Dr. W. A. Woeber, PrEng, DSc(Tech), FIMechE, FIProdE
Seventy-five years have passed
since Frederick Winslow Taylor presented his Paper on "A Piece-Rate
System" to the American Society of Mechanical Engineers. It was he who
adopted the term time-study and who set the pattern for the study of
time elements and cycle times. Today, in operational research, we feel that
indirect measures of fact are preferable to direct operational measures with
the result that statistical and mathematical techniques have, in many
instances, replaced the basic time study technique.
Although stopwatch time study and
elemental time values are useful for measuring repetitive short cycle labour operations,
most indirect activities are characterised by relatively long, irregular cycles
requiring a work measurement technique based on statistical principles that
allow generalisations about the entire time span to be made on the basis of a
sample of that time interval.
Rating research:
The object of modern time
study is to assess the work content of a specific task in terms of the time it
should take a fully trained and experienced worker to carry out that task at
"standard performance". This is achieved by comparing the performance
of the worker with the observer's own concept of standard performance which is
rated at 100 on the B.S. Scale. In addition to assessing the worker's rate of working,
the time taken to complete the task is measured.
By multiplying the rating by
the time taken, the Basic Minute Value (BMV) of the task is obtained. Certain relaxation
allowances, which take into account the ergonomic and environmental conditions
under which the task is performed, are added to the BMV to derive the Standard Minute
Value (SMV).
The random variations in performance times of a particular worker may be
assumed to be normally distributed. The distribution will have a mean y. and
a standard deviation a. The standard error will be equal to a divided by
\/n, where n, the number of observations, represents the sample
size. The actual size of the sample will be determined by the degree of
accuracy which the analysis demands. To assess a "qualified" worker's
rate of working raises the question of how the attributes of such a worker are
to be defined.
In
the Westinghouse System the
worker is evaluated in terms of four factors. These are skill, effort,
consistency, and working conditions. Briefly, skill is defined as the
proficiency at following a given method; effort, as the will to work;
consistency, as the degree of variation in performance times; and conditions,
as the characteristics of the physical environment which affect the worker,
such as noise, light, heat and humidity.
Firstly, if normal working times (NWT) for the
elements are available, a company can develop standard data for future use to
describe a new job. Secondly, elemental time data permit making better
estimates of the probable effect of method changes. A variation and extension of effort rating was
developed by Mundel5. He claims to have reduced the degree of subjective judgment
to what he has called "objective rating". A modified version of the
objective rating system has been proposed by Nadler6. Known as "pace
rating", it follows the same procedure except that the rating of the
worker's operating pace is obtained by selecting a speed, which is comparable,
from a "step film". The film loop consists of a series of gradually
increasing working paces of the performances of a simple job and each different
pace is preceded by a symbol identifying the performance level. The pace rating
system, like all the others, has its limitations; difficulties are encountered
in selecting typical jobs and , judgment plays an inordinate role in their
selection. An interesting research project was undertaken by Desmond who
developed a method of analysis to measure three major defects in the techniques
of work measurement, namely errors in the concept of normal performance, "flatness"
of rating, or the inability to appreciate proportionate changes in speed, and
residual inconsistency of rating, or the inability to recognize the same speed
when seen on more than one occasion. A simplified graphical version of
Desmond's "reciprate method",
based on regression analysis, is shown in Fig. 1.
The ratings are plotted
against the times on specially prepared graph paper, in which the rating scale
is proportional to the reciprocal of the rating. A rule is then placed through
the origin and rotated until it produces the best fit. The time corresponding
to the intersection of this line with the normal performance line is then the
estimated normal time. The next step consists of drawing the study line by eye and
measuring, on a linear scale, the height of its intersection with the original
line through the origin. This distance is represented by the symbol E in
Fig. 1. The height of the intersection of the study line with the rating axis
is also measured and it is represented by the symbol D; this can be a
negative quantity when the particular study is steep.
The ratio D/E is then
the flatness of the study, and records of variation from study to study give a
good indication of the ability to maintain a certain quality of rating with
respect to this defect. Such records are conveniently kept as control charts,
although calculation of control limits is subject to considerable difficulty.
The inconsistency of
the study is determined graphically by means of the range of scatter about the
estimated operation line. The plotted point is chosen which lies at a maximum
vertical distance above the line, and this distance is represented by the
symbol A. Similarly, the distance B represents the semi-range
below the estimated operation line. Then the sum of these distances A + B = to
is the range of reciprate scatter and can be used to estimate the standard deviation
about the line. Desmond concluded his paper by stating that the method had been
applied to the analysis of some 32 000 observations which had been collected during
a nation-wide survey of time study rating accuracy.
Work sampling techniques:
Work
sampling is based upon statistical principles that allow generalisations about
the entire time span to be made on the basis of a sample of that time interval.
The technique, also described as the ratio-delay method, originated with L H.
C. Tippett10of the Shirley Institute of the British Cotton Industry Research
Association. The fundamental principle of work sampling may be stated thus: the
number of observations is proportional to the amount of time spent in the
working or idle state. The accuracy of the estimate depends on the number of
random observations and on pre-set precision limits and confidence levels. The
treatment of the data requires the setting up of a statistical model following
a binomial distribution; statistical quality control methods can be used in
analysis. Hence, from the simple formula for mean proportion:
and tables have been
developed which give control limits. (Barnes). See Fig.
Quantitative approaches to
the solution can be made by assuming that the nature of the service and arrival
times follows a Poisson distribution. Essential to waiting-time analysis is the
ratio of service time to between-service time. This ratio is defined as K.
Work sampling provides a convenient means for the collection of the data.
If (in fractions)
m = machine
time,
b = service time
and
d = machine idle
time,
then
m + b + d = 1
and
K = b/m
Once K is known,
tables, such as those compiled by Docent Conny Palm, may be consulted to find
the changes in the various times of idling, service, and operation with changes
in number of machines per serviceman.
Another approach is Monte
Carlo simulation. This method calls for estimating on the basis of past
experience the probabilities of occurrence to be associated with the various
possible service and arrival times.
Physiological work measurement:
In
physics and the engineering sciences "work" is defined as the scalar
product of force and displacement. In
Physiology, however, the
concept "work" has a more complex meaning. A physiologist sometimes
speaks of "work", where it would be better to use the term
"effort". "Energy" is the capacity for doing work. Physical
working capacity, or the physiological limit for sustained work, is measured as
the worker's maximum oxygen intake. Economy of effort in the performance of a
task is expressed as the rate of oxygen in litres per minute, the
physiologist's measure of energy expended to the work performed. Energy
expenditure data are also stated in terms of calories per minute and heart rate
data in beats per minute.
Relaxation allowance:
This is defined2
as "an addition to the basic time intended to provide the worker with the
opportunity to recover from the physiological and psychological effects of carrying
out specified work under specified conditions". To determine allowances
for recovery from fatigue, the time study analyst uses tables which are based
upon estimates of the physiological strain the worker is likely to experience
as a result of the physical effort of the task and any environmental stresses
which might add to the physical effort.
Physical effort is estimated
from the metabolic activity of the worker, the air exhaled by him being collected
in a Douglas bag. This is then quantitatively analysed and the rate of oxygen
consumed per unit of time is computed.
Under ideal conditions, if
£ = total energy expenditure
per minute,
r = rate of working,
n = energy expenditure per
minute due to normal metabolism, and
m = energy expenditure per
minute not due to normal metabolism,
then
£ = mr + n,
or
m= (E - n)/r
Summing up, it can be stated
that where physical effort is involved in an operation, information can be
obtained by means of energy expenditure measurements, the physiological strain
involved and the time necessary for circular respiratory recovery.
Conclusion
As a result of increased
mechanisation in industry, the muscle load on the worker will be reduced. The
man machine system will give him less and less to do with his legs, arms and
trunk, and more to do with his eyes and his fingers in the way of watching and
manipulating controls, pointers and switches. This is the change of the work
load from the more physiological elements to the psychological aspects of man.
The physical load is diminished and the perceptual load increased.