Gemba Keiei Chapter 20: What I Learned About Forging Changeover from Toyota do Brasil

Taiichi Ohno begins the chapter by saying “In order to achieve Just in Time you need to solve your changeover problems and reduce lot sizes. Forging processes are the most difficult.” This chapter should really be titled “Toyota Learned How to Do Forging Changeovers in Brazil”. It’s an interesting story.

Taiichi Ohno continues to describe what happens in a forging process, such as heating the metal, placing it in the die and hitting it. The combination of the right amount of heat, the right amount of metal, the right amount of time and dealing with the scale that results from oxidization all make die adjustment in a forging process challenging.
Because this adjustment is done by trial and error with two or three parts hit and the process adjusted, it takes a lot of time to changeover a forging press. Forging processes had the longest changeovers at Toyota and this caused Toyota to produce large lots of parts at the forging process.

“Back in those days at Toyota do Brasil we had installed one forging press so that we could bring all of our forging work in house” Taiichi Ohno tells us. This one press had to make more than 60 different forgings. They had only one press because they only build 2 or 3 automobiles per day. It was the smallest automobile company in the world, according to Ohno.

Because of the extremely low volumes, no other forging supplier would make the parts for Toyota. Forgings are typically ordered in 2,000 or 3,000 piece orders due to the long changeover times. This would have been many months of inventory at volumes of 2 or 3 cars per day. Toyota’s solution was to install a forging press and make all 60 parts internally.

According to Ohno, the rule was that you couldn’t make more than 10 of any single part at a time. If changeovers took one hour, you would run out of time very quickly in a day doing changeovers because the runs were so short. So the first target was a 15 minute changeover. If the changeover was 15 minutes and the run time was 15 minutes, you could make two different parts each hour.

By that math you could make 16 different forgings in an 8 hour day, and cycle through all 60+ forgings each week. For those of you struggling with kanban quantity reduction, heijunka or how to satisfy a wide variety of customer demands at low volumes, Taiichi Ohno just gave you the answer.

Ohno notes that while attempts to cut forging changeover times at Toyota in Japan would have been strongly resisted, Brazilians were more willing to listen to a Japanese instructor with a mustache who sounded like he knew what he was talking about.
With the efforts of the Brazilian workers at Toyota and the suggestions from Taiichi Ohno such as external changeover and using guides on the dies they were able to reduce the changeovers to less than 10 minutes. Their efforts were so successful that one forging press was able to make more than 60 different parts and never cause part shortages.

Several Japanese employees were sent to Brazil to learn about forging changeover from the Brazilians. As a result changeover reduction at Toyota in Japan became much more active.

Taiichi Ohno says that casting changeover reduction is not nearly so difficult, and goes on to describe the steps in a casting process. Because the volumes were so much higher at Toyota in Japan they did not have the same sensitivity to long changeover times and the need for lot size reduction. Casting runs of 2,000 or 3,000 were considered acceptable. In the case of casting also, Toyota do Brasil came up with innovative ways to reduce changeover time.

Taiichi Ohno gives a lot of credit to Toyota do Brasil as being a good model plant or test case for implementing high mix low volume Toyota Production System. At Toyota in Japan the volumes were so high that many lines were dedicated and practically no changeovers were done. When Ohno says “Toyota do Brasil may be doing TPS better than any other Toyota factory” he seems to be saying that true TPS is high mix, low volume and when you have high volumes and you do not need to implement SMED, you are not really doing TPS.

Ohno notes that when volumes are so much higher, companies can still turn a profit because of sheer volume, without needing to reduce cost and that the amazing thing was that Toyota do Brasil was profitable even as a high mix low volume automobile manufacturer.

“Toyota Production System is an approach that really ought to be used for mid-sized companies since it will be most effective” due to the fact that they will have lower volumes and more variety. TPS may not exist today had Toyota been a high volume manufacturer in the early days when Taiichi Ohno was helping to develop it.

Put on your fake mustache. Fly to an overeas factory. They may be more willing to listen and try kaizen. Then your home office can learn TPS from them.

---