How AI is driving a green manufacturing revolution

Technology as if people and climate matter. First in a series.

Welcome to EarthWatch, an environmental news and opinion newsletter for people who think you should never turn your back on Mother Earth—written by me, Jerry Bowles, an ancient scribbler who has been around the Sun a few times and doesn’t need a weatherman to know which way the wind blows.

Ideas, tips, and feedback: jerry.bowles@gmail.com

Manufacturing has a monstrous carbon footprint. In 2018, direct industrial greenhouse gas emissions accounted for 22 percent of total U.S. gas emissions, making it the third-largest contributor, after transportation and electricity. Throw in indirect emissions associated with electricity and the number is 28.9 percent, which makes it the largest contributor of greenhouse gases of any sector. Steel and concrete manufacturers are the biggest offenders.

This startling and unsustainable figure comes despite a herculean quality and operations improvement effort that began in America in the early 1980s when Xerox discovered that Canon and Ricoh were selling printers cheaper than Xerox could make them and Toyota and Honda began pushing Chrysler and GM toward bankruptcy.

The Japanese “secret,” as it turns out, was a few operational efficiency tricks they learned in the post-war era from some obscure statisticians from the U.S. Census Bureau, as well as a lot of refinements of their own.

Things like “Just-in-Time” delivery of supplies so tons of inventory didn’t sit idle in warehouses; “kanban,” a workflow management tool to make sure the right parts get to the right places within a factory at exactly the moment they are needed; empowering every worker to be responsible for quality even to the point of stopping an assembly line if they see something seriously wrong. Quality circles, suggestion systems, analytic tools, a comprehensive system of constant analysis of every process by real scientists designed to reduce variability and make better products more cheaply.

It worked spectacularly and many of the big-name manufacturers that are still around today wouldn’t be here without that intense and sustained focus over the past three decades. However, there are signs that, like Kansas City, “they’ve gone about as fur as they can go.”

For all the billions spent on improving industrial quality and efficiency and the trillions made from that effort, there are still operational questions that simply can’t be answered by using the old techniques—most of which were developed before the imperatives of managing emissions, not just quality and productivity, were well understood.

Here’s an example of dealing with uncertainty from an interview I did with Berk Birand, CEO of Fero Labs, a New York-based startup that makes artificial intelligence and machine-learning software for improving processes and increasing the quality of manufacturing facilities. 

If you're a steel plant and you're taking scrap steel and you're melting scrap steel, you will never know ahead of time what it is that you're melting. You will never know if it has high chromium, low copper or high copper, you're just melting some dishwashers and car bodies and you never know what's going to be inside and there's no way you can forecast-until Fero came along--what's going to be inside, so there's always going to be uncertainty.

What we do essentially is build software that doesn’t try to reduce uncertainty but actually tries to measure the uncertainty factors and enable plans to adapt and make decisions based on this information. It’s focused more on let’s try to plan and minimize what’s going to be inside this specific batch.”

Sponsored Content

One of Fero’s earliest customers, Brazil-based Gerdau, which is among the largest producers of long steel made from scrap metal in the world, installed the Fero platform and connected all of the different sensors in its plants to first help identify data that was bad or missing so its quality departments could drill down into and clean up or create anew. As a result, Gerdau began getting more and more accurate predictions about the mechanical properties of the steel it made based on the data coming in.

The company then focused on creating live-prediction dashboards that show operators real-time results as they work the machines.  They communicate with the software in simple language the goals they want to achieve like “minimize raw material costs;” or “make sure that the steel beams are within spec and have a specific strength” or “minimize emissions and water use” or “reduce the possibility of production failure and having to rework the whole process.”

Essentially, they connect the software to the plant. They communicate all of these different complex goals and the software takes all of the raw data and the objectives they want to achieve, and then computes the best way of making steel, the best recipes to use so that they can minimize the raw material costs, maximize quality, and minimize emissions.

The software allows operators to melt a small sample and then measure what's inside and within seconds make a recommendation for how to process the particular load to meet the desired parameters. If necessary, workers can change the entire production plan so it maximizes the quality of the products and minimizes the emissions that are taking place. 

Fero calls its software "explainable ML" because it even tells the operators, who don't have to be engineers, why it is making its specific recommendations and the degree of certainty it has in them. The result is that fewer batches need to be scrapped or reworked because of production failure and that, in return, means thousands of tons of CO2 that don’t get released into the atmosphere. 

Fero is among a number of interesting new companies that are using AI to improve manufacturing processes. Bryan Walsh of Axios just had a worthy piece on one of them. Nanotronics, a Brooklyn-based company, has developed a platform that combines AI, automation, and computer imaging to identify anomalies in the manufacturing process.

Dig Deeper

Inventory of U.S. Greenhouse Gas Emissions and Sinks (EPA)

How the manufacturing industry can minimize its carbon footprint (The Fabricator)

Beyond Six Sigma - can Fero Labs' explainable ML drive new breakthroughs in manufacturing quality? (diginomica)

Obituary and Tribute to Dr. W. Edwards Deming (Jerry Bowles)

Building the AI-Enabled Factory (Axios)

Automated Industrial Quality Control (QC) Market Highlights the Impact of COVID-19 Highlights (2020-2024) | Importance for Accuracy in QC Processes to boost Market Growth (Technavio)

Never Turn Your Back on Mother Earth (Sparks)

You are reading a free version of EarthWatch. If you want to be sure to receive all updates and special alerts, as well as read, comment, and take part in the ongoing dialogue, you should subscribe. I’m an old guy living on a fixed income.