How can engineers use AI?

Artificial intelligence is everywhere, so here are three ways for engineers to get practical value out of this trending technology.

Remember blockchain? 

There was a time when it seemed like it was going to transform engineering, from smart manufacturing to product lifecycle management (PLM) and everything in between. Things didn’t play out quite the way the evangelists hoped (after all, they rarely do), but for the average engineer, the stakes were fairly low to begin with.

Despite their innate curiosity, engineers are a generally conservative bunch when it comes to adopting new technologies into their workflows. “If it ain’t broke, don’t fix it,” as they say, though engineers may feel as though they hear a similar aphorism far more often: “If it ain’t broke, add another feature,” — which brings us to artificial intelligence (AI).

Although there are many hyperbolic claims about what AI can (and especially will) do, there are many real-world examples of how engineers can use it in their workflows today.

Here are three of the most impactful ways engineers are using AI today.

1. Generative Design

Certainly one of the most-discussed examples of using AI in engineering, generative design is an iterative process in which algorithms generate outputs based on a set of imposed constraints. Airbus used generative design to develop a bionic partition back in 2019. More recently, Airbus’ head of AI and advanced analytics said in an official blog post, “While it is unlikely that GenAI will be able to design future Airbus products from scratch, its ability to assist humans by enabling them to better manage complex and technical documents is proving promising.”

2. Computer Vision

Computer vision is even more venerable than generative design, but AI is proving to be similarly useful in taking the technology to the next level. In a typical computer vision setup in manufacturing, a system is trained on a series of images to identify defects during production. Generally, this boils down to a binary decision, with the system flagging parts as either good or not. The difficulty lies in judging edge cases, particularly when variations in lighting or other environmental factors that may affect the cameras and sensors. AI can help overcome these challenges by extracting more information from captured images and inferring more subtle patterns in the training data. Volkswagen recognized the value of this when it announced over 50 projects using AI and machine vision.

3. Fault Detection/Prediction

If there’s one thing AI is good at, it’s managing huge volumes of data, and one of the best examples of this is in quality assurance and control in manufacturing. Whether discrete or continuous, the amount of data generated in modern manufacturing is enormous and it’s continuing to grow with digital transformation. Manufacturers are betting big on AI, especially for predictive maintenance, where machine learning algorithms can identify patterns in production data that point to potential errors or machine downtime. PTC is one of several major software suppliers that are following suit with new AI features.