- March 20, 2019
- Posted by: David Marshall
- Category: Digital Transformation, Leadership, Manufacturing
A few weeks ago, I had the privilege of serving on a panel at the Plastics Executive Conference in Naples, Florida. The panel was entitled Digital Transformation and Plastics Manufacturing, and it was moderated by my good friend, Nanette Gregory, senior partner at NSG Consulting.
I was joined by Rajiv Menon, founder of Informulate; Gary Stein, CTO of Alpha Proto, and Willem Sundblad, founder and CEO of Oden Technologies.
Nanette started the discussion by asking who in the audience considered themselves part of a digital-forward thinking company.
One hand went up, and we knew we were in the right place.
Here’s a brief summary of what we each shared with our audience:
Rajiv Menon graduated from his MBA program and went into mobile banking and banking applications at the corporate level. He had an itch to understand things at a higher level, which ultimately led him to start his own company. He works with clients on their higher level strategic problems. Over the last 13 years, he’s gone from a software company to an agile, continuous improvement company to an innovation company.
Rajiv has worked with different companies on a variety of projects in the area of digital transformation and continuous improvement. For example, he’s helped Stanford University build a program to help them facilitate wellness among their employees. Stanford liked the program so much, they asked Rajiv to help them find new and interesting ways to engage their employees and keeping them moving forward in their healthy behaviors. They channeled the employees’ creativity and looked for ways they could increase their engagement.
One of the realizations they had was that people tend to be a little lazy, but they found a 20% improvement in engagement just by sending reminders to people after the participants had completed a previous step or milestone.
Rajiv is able to find areas of improvement for his customers through measuring their performance, trying new tweaks and adjustments, and then measuring the new results.
I was up next, and I told the packed hall one of my favorite stories, A Tale of Two Plants, which is all about the transformation of our Duoline factory, which made fiberglass epoxy-reinforced liners for oil field tubulars.
I showed a photo of a semi-mechanical operation that used a pre-curing agent that used hot oil, pumped from seven boilers underground to the curing stations. It was an environmental nightmare.
Nothing was automated, it was all manual. And the quality checks were all done at the end of a shift. If you found a quality problem at the end of the shift, you had a good possibility of having to scrap the entire shift’s production. We had ongoing remediation costs more or less $ 1million a year, and scrap costs of $2 – $2.5 million a year, in a $17 million business.
So we scrapped the entire operation and built an entirely new factory from the ground up. We used artificial intelligence and IoT to run the operation, and created a series of autonomous work cells that were each operated by one person, bringing us down to 19 employees.
Most importantly, the system was able to capture reams of data on an automated basis and provide traceability of the entire process from the vendors’ raw materials to our end product. That way, we could find the cause of any problems that might come up.
We were able to completely eliminate our environmental remediation costs because we switched to steam (the only byproduct was water), and our waste was cut from $2.5 million to $25,000 per year. The end result was that we paid off the new factory in eight years thanks to our digital transformation.
Willem Sundblad stepped up next and described what the new Smart Factory of the Future was going to look like, based on his own work in the industry.
“The value of your business is the sum of all the problems you solve,” said Willem. “A digital system helps you solve more problems faster than you otherwise would have.”
Instead of taking days and weeks, you can now do it in minutes, and you can solve a problem you might not have otherwise been able to do. There are algorithms that will help companies predict failures and identify potential problems before they ever happen. So Willem’s expertise lies in connecting machines and capturing the data coming off the production lines and integrating the data into the ERP.
Willem says there are three levels of achieving the smart factory.
- Level 1: Connect the system with the right architecture
- Level 2: Add more intelligence, like predictive algorithms, environmental analysis, predictive analytics.
- Level 3: Instead of the insights going to a person, they go back into the AI system. Willem says we’re not quite there as an industry, but his company is already doing it for a single customer.
It’s important to remember, said Willem, that this is a journey, and you can’t skip ahead without going through the three steps. You have to gather plenty of data in order to understand when and how you’re going to take that next step.
Finally, Gary Stein described how he earned his Ph.D. from the University of Central Florida in machine learning and started Alpha Proto with a few of his fellow PhDs. They do a lot of robotics, such as building robotics cars and surveillance systems for the military and border patrol. They also build technology for oil and gas and injection molding industries.
Gary talked specifically about digital twinning and what it means for manufacturers undergoing their own digital transformation.
A digital twin lets a person see all the data and information that the manufacturing sensors are picking up, and then compare them to another manufacturing setup somewhere else — similar to comparing all the work cells in my old factory — and figuring out why there’s a difference. This can even be done with manufacturing setups around the world, not just in the same building.
You can also use digital twinning to do virtual testing to see what will happen if you make some changes to your system and then apply those changes to the real world.
The upside of digital twinning is that optimization and pattern recognition are easier. You can more easily find patterns and problems before they become too big and expensive to fix. You can also do modeling and prediction to see what will happen if you change things, such as changing a step in the process and seeing what will change in the output.
The efficiency and distribution of digital twinning means that you can transfer the settings and information to different locations. Rather than having one expert in the company who flies all over the country doing setups and adjustments, you can create a digital twin of your system, all the inputs and outputs, and then reproduce it in your other factories.
After our introductions, we answered some questions and from Nanette, and I’ll discuss those in my next blog post.
I’ve been a manufacturing executive, as well as a sales and marketing professional, for a few decades. Now I help companies turn around their own business through processes like digital transformation. If you would like more information, please visit my website and connect with me on Twitter, Facebook, or LinkedIn.