A Case for Slower Run Rates

I am going to be purposely vague about the product here and about the intricate details of the manufacturing process. They are not necessary to understand the solution or how this may be helpful on your own production lines. When I was consulting for a company on a line producing liquid filled containers that had a washer, filling, and lid machine in line we had an interesting issue. The machines were running at close to their tested maximum run rate, but the hourly average of units produced was around half of that. Before I go further, let me explain the setup of the production line.

The containers were loaded into a washing machine that fed into an accumulation that held about 2 minutes worth of containers before feeding into the filling machine. The filling machine was generally gravity fed, and had about 10 seconds of accumulation before feeding into the lid machine. The lid machine then fed into a stacking station. These machines did not communicate to each other, they were independently controlled and ran off of sensor inputs before and after each machine. These sensors were telling the individual machine there are enough containers to start, or that there is a backlog and they need to stop. When a machine would start back up it would ramp back up to speed. The last important detail is that there had to be a specific amount of fluid in each container.

Now that we know the set up this is what I would observe. The washing machine had no problem starting and stopping, neither did the lid machine. They both had consistent outputs that were within the quality standards during their ramp up times. The filling machine was a different story, because it was mainly gravity fed. When the machine was running at close to its max speed then stopped, on the ramp up it would overflow the containers. This was because it was going so fast that if it had filled to 16 ounces it was actually adjusted to try to draw 18 ounces of fluid, and it just didn't have enough time to fill all the way. However, on ramp up it did. The solution did take a bit of convincing the c suits, but I had enough data and was able to explain the fluid mechanics to them.

We set up a set of runs that would allow us to optimize the run rates of the machines. This was done by first running the washing and lid machine at their normal rates. Then we almost halved the speed of the filling machine. After a sufficient amount of time we increase it by 10 percent of the max run rate. What we did during that time was check the fluid amounts then stop the line and check them during the ramp up. This is how we came up with the maximum run for that fluid, in that filling machine for that container. The results were that we not only achieved more units produced per hour, but there was so little waste that we were surpassing the theoretical number of finished units per batch of fluid. This was because they had so much waste for so long that their waste percentage calculations did not have a good baseline.

There are instances even in production where slow is smooth, smooth is fast, and fast is efficient. This is why it is important to look at the data and look for the root cause. The lines might not be poorly maintained or broken, they might just not be optimized.