Applied AI in Manufacturing

“Traditional industrial automation requires hundreds of hours to reprogram, making it very impractical to change how the task is performed.” - Jim Lawton, Chief Product and Marketing Officer of Rethink Robotics

Industrial AI will play a central role in the factory of tomorrow

The AI revolution in manufacturing is part of a bigger change often referred to as Industry 4.0 or Smart Manufacturing. And that is exactly what AI will do to the factories of tomorrow; make them smarter.

Manufacturing has come a long way since the early 1900's when row upon row of workers pieced together a product as it ambled along the first assembly lines one repetitive task at a time symbolizing what we today label as Industry 2.0. 

The 1960'ies marked the beginning of Industry 3.0 with the the first industrial robot making its debut at General Motors in 1961. Coming in at a price of $25.000, the 1.8 tonne Unimate robotic arm lifted and stacked hot metal parts in the GM plant in Ewing Township. The early robots were essentially dumb machines that could only perform a single task. Now, the industry is in the first stages of its next automation breakthrough - Industry 4.0 - using AI to support production decisions in real time. 

What defines industrial AI?

A lot of applications of AI is possible within the factory, also termed “Industrial Artificial Intelligence”. Industrial AI is often related to technologies addressing product and service innovation, productivity improvement, and insight discovery.

A Danish factory is churning out, say, water pumps, when a sensor spots a defect. That data is fed to a computer in the cloud, which immediately pulls the defective part from the line and orders a replacement. That's 's real-time problem solving that can save manufacturers billions of dollars in recalls, repairs and lost business.

Specific technologies pushing to the development of AI within Manufacturing includes more affordable sensors and the automated process of data acquisition; More powerful computation capability of computers to perform more complex tasks at a faster speed with lower cost; Faster connectivity infrastructure and more accessible cloud services for data management and computing power outsourcing.

“Industrial companies need to become digital to survive, we must turn information into insights and insights into outcomes.”

- GE chairman and CEO Jeff Immelt

In the last four years more than 6 billion dollars has been invested alone in start-up companies developing smart manufacturing technologies.

Some of the investments are in enterprise solutions while others are focused on areas where the possibilities between the artificial intelligence and the plant floor makes for a perfect fit.

AI plays an essential role in manufacturing by being the catalyst for whole new types of technical solutions like Predictive Analytics software. Likewise embedding AI on the factory floor will enable new levels of productivity. Take a look at the opportunities in for example Collaborative Robotics and Additive Manufacturing, 3D Printing, Industrial Augmented Reality and Virtual Reality below.

Four of the most promising focus areas for Industrial AI

Collaborative robots – Adaptive manufacturing

Rethink Robotics advocates the “co-bot” model where humans and robots work side by side for maximum effectiveness. While industrial robots have long performed heavy lifting and tedious work on assembly lines, they’re typically designed for a single task and require hours to reprogram. Rethink’s smart collaborative robots, are able to learn a multitude of tasks from demonstrations, just like their human counterparts can.

“Training a robot is nearly as simple as training a human,” claims Rethink. “Companies that don’t have programming expertise on staff and can’t afford to spend hundreds of thousands of dollars on a traditional industrial robot can instead leverage more affordable, flexible automation and adapt to market changes.”

More than $3 billion has been invested since 2012 in robotics startup companies just looking at venture capital. Besides Rethink Robotics, some interesting businesses within collaborative robitics include Universal Robots, Vicarious, and Geek+.

Predictive Maintenance
Industrial equipment is typically serviced on a fixed schedule, irrespective of actual operating condition, resulting in wasted labor and risk of unexpected and undiagnosed equipment failures.

Once instrumented with sensors and networked with each other, devices can be monitored, analyzed and modeled for improved performance and service. This creates a “digital twin” to any physical object, reflecting the state and attributes of the object – regardless of time, state and position.

“Twinning” a piece of equipment allows human operators to constantly monitor performance data and generate predictive analytics. Complex twins like those of gas turbines interpret data from hundreds of sensors, understand failure conditions, track anomalies, and can be used to regulate production based on real-time demand.

But “digital twins” can be crated for both simple objects and for whole factories. Insights can be applied to a lot of business issues helping increase quality and efficiency.

Companies working within predictive maintenance- and analytics include; DataRPMC3IOTUptake and SparkCognition.

Additive Manufacturing / 3D printing
In its simplest form a 3D printer contains no AI software as such. But todays solutions use AI many ways – directly in the 3D printer or indirectly in the algorithms used in the CAD/CAM tools or in the software supporting the printing process. Examples of this are when the CAD/CAM tools directly or indirectly suggests optimized designs using AI algorithms, simplifying a structure, making it lighter or cheaper. Another example is additive printers connected to the cloud. Intelligent algorithms inspect the way a structure is printed and, if needed, optimizes each print.

Another, and fairly new discipline is generative design. Generative design mimics nature’s evolutionary approach to design, and uses AI as part of the process, generating potentially thousands of design solutions.

Through an iterative process, narrowing the solution space, the designer can choose the designs with the best fit depending on the material, structural need, production, etc.

Companies working within additive manufacturing and 3D printing include Carbon and Desktop Metal.

Industrial Augmented & Virtual Reality
Augmented reality, more than virtual reality, is getting traction for industrial use. One example is complex assemblies, where information in a head-mounted display assists the worker. Another is off-site assistance, where the display provides a live feed and presence into far way locations, making it possible to provide remote assistance to co-workers or customers on complicated tasks. 

AI, or just sophisticated algorithms, plays a role in many of the solutions when it comes to creating the “artificial world” and bringing objects together in new ways.

Companies working within industrial augmented- and virtual reality include ScopeAR, Meta and Upskill.

Which industrial AI companies should you keep an eye on?

Below are five sample companies delivering AI supported solutions for manufacturing, with AI playing a larger or smaller role in the software component of the solution.

Whether the companies will succeed needs to be seen. But – their solutions are available today supporting very different scenarios for AI use: industrial augmented reality glasses, additive manufacturing or 3D printing, collaborative robots, predictive analytics and “fabrication-as-a-service”.

With over 16,000 employees and an annual production of more than 16 million units, Grundfos is a global leader in advanced water pump solutions and a trendsetter in water technology. Grundfos_Logo.svg.png

Grundfos' products span over a wide range of industries and serve numerous different purposes, including industrial pumps, wastewater pumps, groundwater pumps, water pressure pumps, dosing and disinfecting pumps and more.

Below is a selection of key figures from the Grundfos organisation: 

Manufacturing facts.png 

Grundfos’ manufacturing processes uses a lot of technology and software, and has been through a continuous technological evolution ever since the company was founded in 1945. Grundfos currently runs a digital value chain initiative trying to solve challenges and exploit new opportunities:

  • Use data to create higher predictability, demand, maintenance, resources
  • Reducing the number of defects
  • Use robots and autonomous vehicles to do the ”heavy lifting”, thus improving working conditions
  • Use “The Digital Twin” and traceability to get increased transparency across production and supply chain
  • Use data to get more insight into quality and production
  • Advanced 3D printing and manufacturing
  • Augmented and virtual reality to improve training, collaboration, service and complex assembly
  • Reduce development time and simplify development cycles

Where can AI be used at the factory floor?

The digital journey of Grundfos is focused on the above-mentioned opportunities and challenges. Some of the solution scenarios and technologies discussed contains AI elements and three scenarios are described in more details below.  The cases are not fully implemented at the plant floor, but they are either in a pilot phase or discussed as a project candidate for Grundfos’ digital journey.