Artificial Intelligence (A.I.) and Mechanical Engineering

Artificial Intelligence (AI) is revolutionizing the industry with smaller, faster and safer technologies.  

In the associated field of machine learning, computers trained to “think” like humans offer insights beyond statistics and data points. One example is Autodesk's Dreamcatcher computer-aided design (CAD) system. Users input design problems, and Dreamcatcher generates designs that bypass those problems, while adhering to user-supplied parameters such as cost, structural requirements, material composition and performance criteria.

In an industry where precision is everything, mechanical engineers are wise to welcome the latest advances in AI.

Machine Learning and 3D Printing Tech 

3D-printing technologies, like those recently employed by Airbus, allow for quicker, in-house production of machining components. 3D-printed parts are up to 55% lighter and require up to 90% less material. In aerospace, every kilogram shaved corresponds to a 25-ton reduction in CO2 emissions.

3D-printing used in conjunction with machine learning promises to create materials of unmatched strength, lightness and complexity. MIT's new Inkbit combines a 3D-printer with machine vision and machine learning. It allows the AI to predict how an object will respond to stress. Then it makes real-time adjustments during construction. Inkbit's robotic vision is powered by optical coherence tomography (OCT), which ophthalmologists use to peer into patients' eyes. Inkbit will use its OCT to see through surface layers of material and provide resolution at a fraction of the width of a human hair.

Engineers Use AI for Maintenance, Repairs

AI is also being used to predict maintenance and repair. The Karlsruhe Institute of Technology (KIT) developed an algorithm to monitor ball screw drives, which guide the machining of cylindrical components. The algorithm analyzed thousands of images displaying various degrees of structural damage, learning to distinguish between actual defects and harmless discoloration from build-up of dirt. 

In the aviation industry, unplanned maintenance accounts for 30% of all delay time.  To prevent unplanned maintenance, Airbus' data platform, Skywise, can collect 24,000 data points on an A320 aircraft, where previous methods could only supply 400. Predictive maintenance has already provided tangible results for British airliner easyJet, which entered a partnership program with Skywise in 2018. Delays caused by unforeseen technical issues have been reduced to three per 1,000 flights, compared to 10 per 1,000 flights in 2010.   

AI-powered predictive analytics combine information from wide-ranging sources, such as environmental data, project reports, financial statements or even readings from sensors embedded in industrial equipment. And according to management consulting firm McKinsey & Company, analytics can yield a 25% reduction in downtime and labor costs, along with a 10% reduction of maintenance costs.

Engineers Find Improved Safety with Predictive Technology

These predictive analytics can also foresee occupational hazards to improve worker safety and reduce worksite injuries. Over the past few years, Smartvid.io and the Suffolk Construction Company have been testing a predictive AI system named Vinnie. Smartvid.io supplied Vinnie with 10 years of Suffolk safety data, including incident reports and jobsite images. Then Vinnie analyzed a mountain of variables — including weather, safety equipment usage and specific project details — to predict workplace accidents.

Smartvid.io compared Vinnie's predictions against three years of incident data and found that Vinnie had predicted 20% of incidents with 80% accuracy when allowed to make four alerts per year. If Vinnie were allowed 12 alerts a year, it could predict 40% of incidents with 66% accuracy. Since each incident costs an estimated $36,000, a company that avoids even 25% of these incidents could save millions of dollars per year.

Over the past few years, Suffolk employed Vinnie alongside the Safety Observations module, an app that logs workers' safety observations. As of December 2019, Suffolk experienced a 28% reduction in recordable incidents and a 35% reduction in lost time.  

According to Smartvid.io, companies using their predictive-based safety systems have benefited from prediction accuracies of up to 86%, which have resulted in up to 60% fewer recordable incidents.

Augmented Reality Reduces Training Efforts, Enhances Build Time

Another benefit of AI is the emergence of augmented reality (AR), which bridges the virtual and physical worlds. One simple but effective example is HOMAG's (a German wood processing company) intelliGuide. The intelliGuide operator assistance system uses lasers to project pictograms or directions directly onto the workpieces. Operators no longer have to look away at their monitors while cutting or shaping panels, which has led to improved efficiency and reduced training efforts.

Smart glass technology, like Microsoft's HoloLens, offers an affordable and accessible platform for AR. By overlaying virtual information across one's field of few, architects can see and even walk through a 3D representation of their designs. Similarly, mechanical engineers can view machines or components from every possible angle, no longer relying on a static, 2D blueprint. The interactive displays can be adjusted, tested through virtual simulations and instantly shared with collaborators. Any 3D-printed parts or projects can quickly be compared and checked for faults against project models viewed in AR.

AR provides personal benefits as well. Workers trained with AR headsets no longer have to suffer reading through manuals or attending training sessions. Instead, they can learn tasks by actually performing them through virtual means. And with AR, new types of collaborative efforts are possible. Users can connect with faraway experts who can remotely view and guide the operators at the jobsite.

Just the Tip of the AI Iceberg

Neural networks and machine learning systems are shaping a futuristic world in ways that seemed unimaginable even several decades ago. AI has given us self-driving cars, but imagine convoys of automated supply trucks crossing the highways or virtual assistants ready to help with daily tasks. New technologies are ushering in medical treatments that work on minuscule scales to target the very parts of our genome that cause disease. 

All that’s in store for us, thanks to the power of AI.