Japan’s Komatsu takes its productivity quest global with IoT and AI
Komatsu is one of the world’s top makers of excavators, bulldozers, and other heavy equipment. Its bright yellow machinery and vehicles are hard at work across the globe every day – at construction sites, mining projects, infrastructure works, forestry areas, as well as in factories, warehouses, recycling plants, and more.
The company has decades of expertise in making machines that can dig, lift, and move all sorts of physical stuff.
But when it comes to gathering and handling data to boost its own manufacturing capabilities and productivity, Komatsu needed help. So, it turned to the Azure cloud and the specialists at Microsoft.
Komatsu is an innovative manufacturing enterprise that competes in an increasingly unpredictable international marketplace. Ever-shifting economic and other forces – like booms and busts in resource markets – are constantly pushing demand for its equipment up and down from country to country.
Maintaining production momentum in the face of this sort of uncertainty can be a big challenge for factory managers.
“Keeping pace with these fluctuations is our primary issue,” explains Nobuyoshi Yamanaka, who is General Manager for Komatsu’s Manufacturing Engineering Development Center Production Division. “The best way to do that is by raising our productivity. And, to do that … we need data.”
With the right data and the right insights, decisionmakers can visualize situations. From there they can opt to speed up or slow down production runs, manage supply chains, and accommodate factory downtime for retooling and maintenance.
They can also optimize the use of personnel – a key factor in Japan’s sophisticated manufacturing sector, which is grappling with a shortage of skilled workers as the nation’s demographics age.
Acknowledging that it had a need for data, Komatsu went about seeking advice on what technology and data solutions would be best for its ambitious productivity quest. They searched widely and settled on Microsoft.
Adopting a cloud solution
“Microsoft asked us what we wanted to do and how we wanted to expand the solution in the future, then it gave us exactly the right support,” says Yamanaka whose team is now studying how artificial intelligence (AI) and Intelligent Edge solutions might further boost efficiencies.
To best explain Komatsu’s cloud adoption story, let’s go back to 2009. That is when the company first set out to collect production data by using on-premises servers. Five years later, it went further and launched “KOM-MICS” – an Internet of Things (IoT) system that collects data from sensors that are installed on a myriad of machine tools and welding robots.
“Komatsu uses a high-mix/low-volume manufacturing system. Plant equipment is not always operating at full capacity as machines may be down for many hours due to setup changes, and so on,” Yamanaka says. “Visualizing this situation and reducing machine downtime increases manufacturing output without extra equipment or personnel. Our ultimate goal is to double productivity while reducing equipment and personnel.”
KOM-MICS was a success. And, soon so much information was coming in that Komatsu realized that its on-premises approach to data needed a rethink. It also wanted to collect and visualize data from a network of outside partners and other factories, both in Japan and abroad, that contribute around 80 percent of its overall production work.
So in 2016, it began looking around for a cloud solution.
“We needed to roll out KOM-MICS to our partners and overseas manufacturing bases to increase the overall productivity of Komatsu,” explains Keisuke Tsuboi from Komatsu’s Numerical Controller Team, Advanced Technology Promotion Office.
“Because KOM-MICS collects 20 to 30 GB of data from each machine tool per year, adding the required resources to the on-premise system, and increasing the number of connected machine tools, would have been difficult. So, we decided the cloud could overcome these problems.”
Weighing up the options
Komatsu compared several cloud services and moved its data onto Azure in early 2017. According to Tsuboi, a primary reason behind the choice was trust: Azure has extensive security measures backed by Microsoft’s expertise. Azure also made Komatsu’s data capabilities immediately compliant with GDPR – the European Union’s new globally important data protection measure.
“It is difficult to keep track of security technology advancements. So we believe it is better to entrust the professionals at Microsoft,” he says. The flexibility and scalability of Azure were also deciding factors that is allowing KOM-MICS coverage to ramped up almost seamlessly.
“We are connecting 100 to 200 extra machines to KOM-MICS per year,” Tsuboi says. “We have around 700 connected machine tools and 350 connected welding robots. Komatsu has around 1,200 machine tools and 700 welding robots that can be connected to KOM-MICS.This scale of data is no problem for our system on Azure.”
Komatsu connected its Thai and Indonesian bases to KOM-MICS in 2017. Since then, the number of Komatsu’s partners connected to KOM-MICS has been increasing rapidly.
“The transition to Azure instantly expanded the potential scope of the KOM-MICS rollout. The meticulous support of Microsoft enabled us to complete the migration in a short time,” explains Yamanaka.
With earthquakes and typhoons a constant threat in Japan, Azure has also bolstered Komatsu’s
disaster response capabilities compared with the previous on-premises system.
In the end, more data from more machines in more places means the company can improve quality measures, plan and adjust with agility, and better anticipate equipment failure.
“Before we started collecting data, we didn’t know to what extent our machines were working within a 24-hour period,” says Tsuboi. “With KOM-MICS, data is visualized so we can work on improving production efficiency by increasing areas with low production conditions to be equal to those that are high.
“By analyzing the machine data from a certain production line we have been able to increase the machine operation rate by about 25 percent.”
A future with AI and the Intelligent Edge
Looking ahead, Yamanaka believes artificial intelligence (AI) on the Intelligent Edge can potentially deliver more productivity dividends, such as freeing up the time of skilled workers and opening the door to predictive maintenance.
“I believe that data can be used in a variety of ways,” he says. “We would like to automatically realize optimal machining conditions and have AI do some tasks that are currently handled by skilled workers.
“Also, there is quality. We would like features that can automatically detect signs of failures before they happen. We need to make use of AI. But because processing data in the cloud takes time, we are thinking about adopting Azure IoT Edge so we can run Microsoft Azure services on IoT devices.