Efficient and optimized product design – AI transforms industrial R&D

Generative AI will aid the work of various professionals at an increasing rate as it is becoming more accessible and easier to use, leading to enhanced productivity and streamlined operations across industries. But this is merely the beginning for industrial R&D.

“With enough data, the future of product design could involve automating processes where a designer describes their vision verbally to the AI, and the model on the screen continuously updates accordingly,” Tero Hämeenaho, Head of Additive Manufacturing and AI Program Director at Etteplan, envisions.

While future visions are still far ahead, there are good grounds to predict that generative AI and large language models will profoundly shape the playing field in R&D. For example, industrial players have already seen increased efficiency and cost savings in pre-emptive maintenance and quality assurance.

“Generative AI can analyze data and predict outcomes more accurately. It can potentially improve information-based decision-making and decrease costs and lead times. This will undeniably speed up innovation in R&D”, Hämeenaho explains.

Real-time data processing will become easier thanks to expanding edge computing and AI chips. Success will require high-quality data and proper data management.

In various design tasks, such as equipment or process design, it is important that plans are based on up-to-date standards and use existing design guidelines. However, there are tens of thousands of standards alone, and finding the right information takes designers' valuable time.

"SFS needed a solution that would help technical designers easily find the right information. Together, we created an AI-powered conversational bot that swiftly identifies the appropriate standards or design guidelines from a vast data pool and multiple sources", Hämeenaho says.

Both design guidelines and standards include many images, illustrations, and charts. The AI had to be able to find relevant information in different formats to offer meaningful assistance to its users, so every piece of data fed into the service was vectorized.

“Increasing efficiency with AI requires a sufficient amount of refined data. For instance, if a company has enough product design examples, AI could automatically select and modify models based on the company's guidelines, speeding up the design process,” Hämeenaho explains.

Another long-standing topic has been the utilization of operational data from sensor-rich devices. By feeding this data into simulated environments and digital twins, AI could assist in creating optimized products.

“Companies may already possess plenty of data, but it is scattered across different systems and platforms. Companies need to ensure they gather relevant, quality data to gain any advantage from AI now or in the future”, Hämeenaho points out.

Many industrial applications, like robotic automation, require real-time data processing to function efficiently. Until recently, this has hindered many attempts to utilize AI, as the latency caused by cloud-based processing is often unacceptable in industrial environments.

The shift towards edge computing, where AI is embedded directly into devices via specialized chips, is a game changer for industrial companies. It reduces data transmission needs and enables real-time, on-site processing.

According to Hämeenaho, advantages such as AI-based performance optimization will soon be accessible to even more companies. “With edge computing and AI chips, it is possible to create smarter, safer, and more autonomous systems without the delays of cloud-based computing.”

To get started, R&D should explore the possibilities AI can create for their teams and the entire business. While current AI tools may have limitations for future use cases, it is crucial to accumulate experience and offer useful learnings for future endeavors. That way, R&D teams will stand ready to reap the benefits from early on.

“Companies with gained experience will stand among forerunners when future AI tools, like physics-based AI models, become available. They already understand how AI can serve their business and create advantages,” Hämeenaho points out.

Successful AI projects do not necessarily mean heavy investments in internal resources or competencies. A basic understanding of AI and its strengths and weaknesses is valuable, but companies should have a clear grasp of the limitations within their own processes and methods at a minimum.

“Etteplan has helped its customers by breaking down their processes to identify opportunities for AI to deliver improvements and added value. Our industry expertise allows us to create meaningful solutions that give our clients a competitive edge,” Hämeenaho concludes.

Are you looking to leverage R&R’s full potential to stay ahead of the curve? Check out our guidebook here!