Page 224 - How To Implement Lean Manufacturing

202 Cha pte r T h i r tee n

The first evaluation to make on a balancing chart (see Appendix C in Chap. 7), is to
evaluate the waiting time. Recall that the waiting time is the distance from design time
to the station cycle time, plus the time for transportation. In this case, the waiting time
for station 1 is two seconds, plus two seconds for transportation for a total of four sec-
onds, station 2 waiting time is three seconds plus two for transportation for a total of
five … and so on, rendering the total wait time as 108 seconds per unit produced. This
is the design waiting time, not the actual waiting time. Quite frankly, that is both huge
and amazing. The actual work time is 228 seconds per unit for a total labor time of 336
seconds per unit, this means the designers planned for labor losses, due to waiting
alone, of 32 percent [100 × (108/336)]. However, since the process is not able to perform
as designed, our labor used is 462 seconds per unit (22 seconds × 21 stations). The actual
waiting losses are 234 seconds per unit [(22 × 21) – 228]. And now it is obvious that the
time losses exceed the actual work time. To put this in simple terms, “we need a lot
more time and a lot more labor than we planned for.” So, in summary:
• We have 228 seconds of planned work in each unit.
• We are paying for 462 seconds of labor per unit.
• Our actual wastes due to waiting are 234 seconds per unit.
• The planned waste was 108 seconds per unit.

These represent huge losses that we may be able to exploit in the redesign.
Second, we evaluate the balance, qualitatively. In this case, the balance is very bad.
There are five stations with a 13-second cycle time, and seven with cycle times of ten or
less. That is a 30 percent variation, which then translates into a lot of wasted labor.
Clearly, rebalancing will benefit this process.
Third, we find the bottleneck. In this case, we only evaluated to the nearest second,
so it appears there are five bottlenecks. There is, by definition, one bottleneck in any
system.
Although we do not know which station is the bottleneck, the evaluation is extremely
revealing. First, it shows the average times are all below takt, yet we know the process
cannot sustain this average rate. The graph also shows that there are five stations right
near the 14-second maximum time. Consequently, if a slight variation in cycle time
occurs, the operator would likely hit the delay button and this would delay the entire
process for five seconds. Since on every unit of production, this opportunity is pre-
sented seven times, I was actually amazed we were able to produce as much as we did.
Recall that our real cycle time was 22 seconds. If we add the designed cycle time of
16 seconds (14 seconds of work plus 2 seconds of transportation) to the time of one delay,
5 seconds, and compare this total of 21 seconds to the actual cycle time of 22 seconds, it
means, that on average the delay button is struck once per production unit. It is clear
that a redesign is needed.
How Many Cells? How Many Stations per Cell?
With the time study and balancing study completed and evaluated, we needed to come
up with a reasonable cell design. As you recall, little capital was available to spend on
this project, so we were not able to change the end of the line, which had some expen-
sive machinery. Stations 17 through 21 consisted of two tests and three packaging sta-
tions, which culminated in pallets of 12 units being shrink-wrapped and ready to load
on a semi. These five stations all produced at cycle times below takt and had a total of