To all of my followers, I made a mistake on my posting number so I'm reposting this as Part 196.
Our piece on the Conflict Resolution Diagram is no quite ready to go to press, so in today’s posting I’m going to substitute what I think is an interesting and common place dilemma facing many manufacturing plants. I have been working with a client that produces medical devices and a couple of weeks ago I had an interesting discussion with the plant’s GM.
Our piece on the Conflict Resolution Diagram is no quite ready to go to press, so in today’s posting I’m going to substitute what I think is an interesting and common place dilemma facing many manufacturing plants. I have been working with a client that produces medical devices and a couple of weeks ago I had an interesting discussion with the plant’s GM.
I had done an assessment of
his plant earlier in the year and one of the most disturbing findings was the
level of work-in-process (WIP) inventory scattered throughout his plant. It was clear to me, after an in depth look at
his operations, that there was no synchronized flow in place. In fact the system in place was clearly based
upon the principles of “push” rather than “pull.” That is, keep the workers “busy” all of the
time. The plant was also characterized
by large batch sizes and no discernible production scheduling system to control
the level of WIP. Part of the discussion
I had with the GM was the impact these large batch sizes were having on the
overall production lead times which in some cases were as high as 6
months. This concept of batch size
versus lead time is the subject of this posting.
Before moving on, let’s
explore the differences between a synchronized and nonsynchronized manufacturing
environment. A nonsynchronized
environment is characterized by systems where products have long manufacturing
lead times and materials spend a large amount of time waiting in queues as WIP. Studies have demonstrated that in many
manufacturing plants, the majority of manufacturing lead time for materials is
actually spent waiting in these WIP queues.
In some plants, the actual processing time on a given order is as little
as 5 percent of total manufacturing lead time and this plant’s lead time
certainly met this scenario. Conversely,
synchronized manufacturing plants have relatively short manufacturing lead
times and materials spend very little time waiting in queues. Unlike the nonsynchronized plant, processing
of materials in a synchronized plant accounts for a relatively high percentage
of the manufacturing lead time.
There are several key
factors that impact synchronization in a manufacturing setting, but one of them
in particular, batch size, plays a very large role. As I explained to the GM, improperly chosen
batch sizes contribute significantly to nonsynchronized material flows. The direct result of this lack of
synchronization are increased levels of inventory, extended manufacturing lead
times, poor on-time delivery and needless operating expense. Many times these plants must resort to
overtime and expedited freight charges. Of
course, this GM did not agree with and asked me to prove it to him. One of the easiest ways to demonstrate the
impact of batch sizing decisions on synchronized flow is by looking at a simple
example which is what we did.
In this process example
there are two side-by-side machines used to clean the interior of the
parts. The two different media types are
used to remove things like burrs or other obstructions/impediments from the
parts. The figure below summarizes the
normal running conditions for the parts entering this 2-step process. The parts to be processed are received from a
feeder process in batches of 48 parts and 8 parts are loaded into a fixture which
is inserted into the first machine.
These 8 parts are run for 30 minutes and then removed, cleaned and
placed in a holding rack. The next 8
parts are run and the process repeats itself until all 48 parts are
completed. When all 48 parts are
completed, they are then moved to the second machine (Media Type 2) which
contains a second media type where, like the first machine, are run in batches
of 8 parts. From beginning to end the
process takes a total of 360 minutes of elapsed time from beginning to end to
complete all 48 parts (i.e. 180 minutes on Media Type 1 machine plus 180
minutes on Media Type 2 machine). This
example assumes no setup time.
This is a classic example of
a nonsynchronized flow, but what would this same scenario look like if the flow
was synchronized?
In the figure below we see
that instead of waiting for all 48 parts to be completed on the Media 1 machine,
after completing the first run of 8 parts, we immediately start processing the
parts on the Media 2 machine.
In this case our wait time
is the length of time to complete the first run of 8 parts or 30 minutes rather
than the 180 minutes of wait time in the nonsynchronized scenario. Now both machines are running simultaneously
until all 48 parts have been completed on both machines. Instead of taking 360 minutes to complete, as
in the unsynchronized flow, we see that the total elapsed time is now only 210
minutes. In effect what we have done is
change the transfer batch size from 48 parts to only 8 parts and the result was
an almost 60% reduction in lead time. It
didn’t cost us anything to achieve the seemingly impossible (impossible to the
GM that is).
It should be apparent then
that one of the key considerations we should always be evaluating is the
concept of transfer batch sizes within a facility as part of our synchronized flow
process.
Bob Sproull
5 comments:
Is there a mistake in numbering or the blog # 196 will come later?
Oh my gosh...yes, I made a mistake in the numbering scheme. I'm going to repost this one as 196. I apologize to all of my followers.
Bob Sproull
I believe Little's explains that if we decrease WIP then Cycle Time will decrease as well - exactly what your second example shows.
CT=WIP/TH
where:
CT - cycle time
WIP - work in process (batch size in your case)
TH - throughput
I meant Little's Law sorry for the omission.
Dimitar, yes you are absolutely correct in stating that Little's Law confirms the relationship between TP and CT. This law works equally well with individual processes or systems. Good point.
Bob
Post a Comment