Implementing Artificial Intelligence on Manufacturing Floors
The following article was first posted in the Insights section of The Hartford’s website. It is reposted here with permission.
From 2001 through 2022, the average monthly growth in the workforce slowed to 0.6% year over year, which has caused many manufacturers to turn to artificial intelligence to enhance production and quality control on factory floors.
“Machine learning, which falls under the umbrella of AI, will be the gateway to adopting AI at large within manufacturing,” says Brian Kramer, underwriting officer and manufacturing industry practice lead for The Hartford.
“Maintaining efficiency is critical for business and the ability of this technology to support predictive maintenance in advance of a mechanical failure has an impact on productivity.
“Leveraging data to prevent issues before they surface will reduce waste and enhance efficiency.”
Uses of AI in Manufacturing
While manufacturing was a late adopter of AI into daily procedures, it’s quickly becoming successfully integrated, including:
Digital twins have grown in popularity due to their ability to create real life scenarios without wasting resources.
In manufacturing, they are primarily being used to create virtual copies of real-world components in the manufacturing process.
By implementing them into the design process, it allows the user to test new products and processes through systems engineering, modeling and multi-dimension simulations.
Virtual reality has many uses but is primarily credited for strengthening safety protocols.
VR can provide reality-based training that aids safety protocol education, such as navigating confined spaces, and helping workers understand the assembly process.
It’s also used for factory floor planning. In mass-production manufacturing, planning where to place tools, equipment and personnel is crucial for productivity and efficiency.
Machine learning is the process of using data to help a computer learn without direct instruction, which enables a computer system to continue learning and improving on its own, based on experience.
“Machine learning can improve process optimization that enhances productivity and reduce process-based waste,” says Kramer.
“Machine learning can also improve quality control and support efficiencies in navigating the supply chain.”
Predictive maintenance is a method of using AI to predict when a piece of equipment will likely fail, prompting maintenance to be done before it reaches that point.
The technology analyzes data from sensors and machinery on the factory floor to understand how and when failures and breakdowns are likely to occur.
“In the world of preventative maintenance, using artificial intelligence and sensor technology, you have the ability to be much more precise,” says Andrew Zarkowsky, global technology industry practice lead at The Hartford.
“This means you can use a machine right up until the point that has the highest probability of breaking, and then change out the part.
“You can operate the existing parts longer versus regularly scheduled maintenance and lower the likelihood that a machine will go down.”
Zarkowsky further explains that predictive maintenance can reduce the risk of business interruption over time.
“If you have a key piece of machinery that is critical to your process and it breaks, your operation is slowed or shut down,” he says.
“Predictive maintenance allows you to be more efficient and reduces the chance of machine downtime. This will reduce unexpected costs and over time lead to more consistent product creation.”
Cobots are designed to work alongside humans in a safe way, integrating our abilities with their own.
Many manufacturers are investing in this technology given it’s cheaper to operate because cobots don’t require dedicated space to function. This means they can safely work on a regular plant floor without the need for protective cages or segregation from humans.
“Manufacturing robotics has been around for a long time,” says Zarkowsky.
“What is changing is the environment in which a robot can work and the type of job it can do. Now, you can have goal-oriented tasks that don’t require a specific environment.”
For example, Zarkowsky explains, you can dump a box full of products on a conveyor belt and have a robot pick out all of the toothpaste and put it in a box.
“With the assistance of AI, cobots learn new functionality faster,” he says.
Cobots can carry out manufacturing operations like screwing, sanding and polishing. They can also complete quality control inspections using computer vision-enabled cameras.
Cobots are widely used by automotive manufacturers to perform tasks including gluing and welding, greasing camshafts and injecting oil into engines.
“You can start envisioning a world where you don’t have to create an environment for the robot,” says Zarkowsky.
“The robot can work in any environment, which creates more efficiency on the production floor.”
As more manufacturers rely on AI for production, there will be an increased need for oversight should obstacles arise.
“Right now, many manufacturing plants outsource their technology needs to some sort of third party, but I don’t think that is sustainable,” Zarkowsky says.
“Manufacturers are going to need an in-house technology team that understands their individual needs, which creates potential for additional costs.”
Additionally, as technology becomes the primary driver for processes, manufacturers need to be prepared for failure, accuracy concerns and cyber liability.
However, those costs are offset by the potential for AI to decrease manufacturing claims via digital twins, cobots, and predictive maintenance which can help reduce the potential for worker injury and business interruption.
Both Kramer and Zarkowsky agree that manufacturers should start small and research different use cases and applications for adopting this technology.
“The best advice is to find a project and get started. Conduct research on partner robotics and AI companies to see how they are implementing this,” says Zarkowsky.
“You can take small parts of your operation and test ways technology can make your machinery more efficient.
“This is not a future idea. This is a today.”
About the authors: Andrew Zarkowsky is a global technology Industry practice lead with The Hartford. Brian Kramer is an underwriting officer and manufacturing industry practice lead with The Hartford.
EXPLORE BY CATEGORY
Stay Connected with CBIA News Digests
The latest news and information delivered directly to your inbox.