5 insider tips to access the IIoT

How to swim through the big data lake without drowning

By Mike Bacidore, editor in chief

If you’re like most people, you’re still trying to figure out how and when to leverage this Industrial Internet of Things (IIoT) that is everywhere and nowhere at the same time.

An Hitachi distribution center increased its productivity by 8% through dynamic scheduling of orders based on data aggregation, analytics and machine learning.

What? How can something be nowhere when it's everywhere? Is this some sort of ancient Egyptian riddle designed to stump the unworthy? Don't worry. We've got your back.

The IIoT is nowhere because it’s not a tangible system and, in many companies’ views, it is still a futuristic concept that is decades away, if it ever happens at all.

The truth is IIoT is everywhere. In its most advantageous form it literally includes every piece of machinery on the planet, sharing data, predicting outcomes and optimizing production based on global, real-time information. It’s also everywhere because it dominates most industrial conversations, from the supply chain through production and right on to fulfillment and distribution. It has become the staple of every industrial trade show and conference, as well. Any event without at least one exhibitor or educational session not focused on the IIoT is, well, Nowheresville.

There is so much to learn and so many real-world IIoT implementers to learn from that the conferences have become one of the best places to gather what you need to break from the concept phase and see the possibilities. At the North American Manufacturing Excellence Summit (NAMES17) in Wheeling, Illinois, a parade of speakers from varied companies, such as Ford Motor, Kellogg, Siemens, GE and Newport News Shipbuilding, discussed how to feed the production beast, and IIoT seemed to permeate each spoonful.

Download our special report on trends in industrial networks

Greg Kinsey, vice president of Industrial IoT Solutions at Hitachi Insight Group, served up a heaping helping of advice in his presentation on manufacturing digitization, which included five key considerations for a successful implementation.

  1. Aggregation of big data enables the value. You have to build the analytics and break down those silos.
  2. Collaborative creation is better than off-the-shelf technology solutions. You cannot buy the factory of the future in a box.
  3. Begin with the business case. When you can take the business case to your manager and show a return that is three times the investment, that ROI makes for a good business case.
  4. Integrate with a continuous-improvement system/culture. Embrace digital as a way to reinvent kaizen.
  5. Execute a five-year transformation roadmap. Having a five-year plan allows you to do some quick wins and move forward on that transformational journey.

The majority of Hitachi’s $90 billion in annual revenue comes from Japan, where it’s involved in a wide range of industrial businesses, but it’s also the 12th largest software company in the world. “Some people think of us as the Japanese GE,” said Kinsey. “We’ve all heard about the fourth industrial revolution, which those of us in the room are supposed to be driving, but there are really three key opportunities for manufacturers; they are smart manufacturing, smart products and smart services.”

Smart manufacturing includes productivity gains, quality improvements, flexibility, mass customization and virtualization. “Most businesses are focused on improved manufacturing,” explained Kinsey. Smart products are offerings such as tracking, monitoring, remote control, diagnostics and maintenance. “Even Philip Morris has announced it’s switching its business from tobacco to smart products,” he noted. And smart services mean the monetization of data.

At a company like Hitachi, it would be easy to simply fall in love with technology, but Kinsey explained that its approach is to step back and look at how the technology can create value as it builds the factory of the future.

“Excessive time and resources are spent on fighting fires, he said. “Production shifts rarely go as planned. End-to-end visibility of operations is poor. Productivity and quality are impacted by variation. Anomalies cost you productivity; they cost you quality; they cost you money.”

A large number of existing manufacturing IT systems have been out there a long time. In any give factory, there might be many different types of systems or software that don’t talk to one another. Kinsey discussed the “data lake” concept, in which you can put together data from different formats, including video, spreadsheets, sensor measurements and many others. By analyzing what’s in the data lake, a company is able to process and use that data for something useful. “You can start to get value and overcome some of the problems in your IT systems,” explained Kinsey. “On one hand, you have to run your business; and, on the other hand, you want innovation. The key is it has to be aggregated and scalable.”

What can you do with the data to optimize manufacturing? Aggregation, streaming and integration of data enables the creation of correlation, logic and algorithms. Data analytics allow you to find patterns, combinations and exceptions. Millions of iterations of machine learning help to drive continuous improvement and leaner operations.

“We came up with a digital manufacturing roadmap because most companies we work with weren’t sure where to start,” said Kinsey. “Ours includes six levels of maturity.”

  • Level 1: visualization—people downstream can see what’s going on upstream
  • Level 2: integration
  • Level 3: analysis—making sense of the data
  • Level 4: predictive—predicting problems
  • Level 5: prescriptive—suggestions on what to do
  • Level 6: symbiotic—setting up a global network of factories.

“Maybe your roadmap will have fewer or more steps, but it should focus on capabilities, which will drive maturity on the map,” explained Kinsey, who then explained three use cases, all of which can work from the same database of big data. They are getting rid of bottlenecks through dynamic scheduling, which is digitization of theory of constraints; predictive quality which is digitization of Six Sigma; and maintenance optimization, which is digitization of maintenance excellence.

An Hitachi distribution center increased its productivity by 8% through dynamic scheduling of orders based on data aggregation, analytics and machine learning. And the Hitachi Omika factory improved lead time by 50% in a high-mix, make-to-order production operation through dynamic scheduling. By aggregating customer order data and operational data sets, bottlenecks are predicted and reduced or avoided, said Kinsey.

 

And Daicel, a global producer of pyrotechnic injectors for automotive airbags, made a strategic decision to digitize its quality management system. “The goals were to increase certainty of product quality, reduce cost of internal rework and root cause eradication,” said Kinsey. “Through digitization of man, machine, method and material, aggregation of 3D-image analysis with data from IT systems and IoT devices enables defect prediction and improved quality management processes.” And then digitization, analytics and predictive quality are integrated with the continuous-improvement system.

“We started to measure processes and materials,” said Kinsey. “To monitor the workers, they installed video cameras and were able to do a physical analysis, thinking if they knew how people were working they could help to correct new workers who were doing things wrong and developing carpal tunnel syndrome.” While this type of employee monitoring might not be feasible in the United States, the implementation in this case has lowered worker injuries and increased production.

Show Comments
Hide Comments

Join the discussion

We welcome your thoughtful comments.
All comments will display your user name.

Want to participate in the discussion?

Register for free

Log in for complete access.

Comments

No one has commented on this page yet.

RSS feed for comments on this page | RSS feed for all comments