It’s Alive! Engineers Are Starting to Design Self-Aware Machines

With Digital Twin technology, the stage has been set for intelligent machines to collaborate with each other and optimize themselves.

The way engineers design machines is changing, says Paul Goossens, vice-president of engineering solutions at Maplesoft, an industrial software company.

The old way was build-it-and-tweak-it. Engineers would do some structural and steady-state analysis, figure out the size they need, and then make something that fits… hopefully.

“It’s typically during the prototype stages or when doing your first build for a customer that a lot of problems arise,” says Goossens. “Historically, for machine designers, addressing those issues was just seen as the cost of doing business. But that cannot happen anymore.”

With the onset of the digital industrial revolution, designs are becoming much too complex, and the competition for efficiency is simply too fierce to risk costly setbacks. There’s a paradigm shift underway, says Goossens, and it’s being driven by designers who take a “systems approach” to their designs.

To do this, engineers start by creating a virtual prototype based on the available information about the mechanical systems, such as the actuators, motors, hydraulics, and so on. By taking this step, engineers can then design a physics-based model – or “Digital Twin” – that functions as a simulation of the actual machine, whether it’s been built yet or not.

Autonomous vehicles lead the way

The fruit of this approach can already be seen in drones and self-driving cars. “What made drones a tenable technology is that they use physics models,” says Goossens. Ditto with self-driving cars. “They have a physics representation of the vehicle itself, which they can use to predict behavior if the vehicle needs to go into an avoidance maneuver.”

Digital Twins are already prevalent in the aerospace and energy sectors, but the technology is still far from reaching its full potential. “It’s all very new,” says Goossens. “The door’s just been opened.”

Despite the relative novelty of the technology, GE is already very active in this space. GE engineers design Digital Twins for all kinds of systems and equipment, from engines to capital projects to power turbines to sewage treatment plants. Why all the investment?

A baseline for awareness

As a starting point, Digital Twins offer a reliable baseline that allows operators to spot malfunctions before they happen. “When you have a virtual simulation running alongside the actual machine, any discrepancy between the two lets you know that something’s wrong,” Goossens explains.

When the machine drifts from the simulation, operators can then take a follow-up step and model different solutions. For example, they might use the Digital Twin to figure out a compensation strategy that will cut downtime. 

Digital twins not only prevent costly breakdowns, but they also lay the foundation for machines that monitor themselves and manage their own maintenance. In other words, machines that are self-aware.

An intelligent machine becomes self-aware when it recognizes that it’s drifting from the virtual model, and then either automatically makes the necessary adjustments, or else schedules its own maintenance and even orders its own replacement parts.

That’s not all. “You’re going to see collaborative machines that communicate with each other,”
Goossens predicts.

As machines learn to identify and anticipate when they will need replacement parts, they will send that information to other intelligent machines at the manufacturing facility. Those in turn will coordinate big data analytics to forecast what products should through the assembly line. “Machines will be able to re-configure supply to meet that impending demand,” says Goossens.

Want more stories like this delivered to your inbox?

subscribe

Related Stories