How AI Impacts an Industrial Machinery Manufacturer’s Design Process

AI has the potential to provide machine builders with superior data analytics abilities and improved human-machine interactions. (Image: Siemens)

Speed and flexibility are increasingly becoming the cornerstones of modern manufacturing, even as their continued adoption must align with existing values of cost and reliability all while keeping up with the demands for smarter, more complex products. This presents many challenges to machine builders since they must keep pace with the complexity of upcoming products while also being ready to meet the demands of the companies that will buy and operate these machines when it comes to efficiency, rapid production line ramp up, small batch sizes and high quality. Artificial intelligence will be a key tool going forward in achieving these results, offering the ability to more rapidly design, prototype, and implement changes and solutions through superior data analytics abilities and improved human-machine interactions.

Additionally, to help support the increasingly adaptable and data-driven landscape of manufacturing, machine builders must be ready to support the data collection required at every level of the manufacturing process. Adopting these new practices can be challenging, however, as manufacturing companies demand the same value and reliability of the past while needing to embrace the smart, flexible machines of the future. But by applying AI both to the machine design process and the finished machine itself, it’s possible to help smooth the adoption of new best practices and technologies without driving up costs or compromising on reliability.

Generative design is a powerful, AI-augmented tool to help designers better explore a design space and achieve optimal designs quickly and on budget. Rather than just optimizing for a minimum or maximum value, like in topology optimization, generative design tools help explore various combinations of parameters across an entire design space. This can help both highlight designs that would not have been found otherwise or achieve an optimal result faster, leading to a shorter design cycle.

AI and machine learning play a key part in the all-important step of picking results. A generative design process could yield hundreds or thousands of potential designs — far too many for even a team of designers to sift through. This is an ideal place to deploy machine learning which can easily filter results based on input criteria, past design data and even specific user preferences, making sure designers consider only the best designs, reducing or even eliminating the time spent on designs that don’t end up being used.

Beyond helping in design space exploration, AI can play an important role when it comes to using design tools. Learning how to use such complex tools will take new users months, and true proficiency likely won’t be attained for years. Even once mastery of these complex and feature-rich modern design tools is attained, accessing a single function could require traversing convoluted menus and submenus, hampering the design process with inefficiencies.

Teaching an AI model how to use the tools is a quick way to both flatten the learning curve and make the tool itself more user friendly. Using smart command prediction, an AI model that is aware of the design being worked on, common workflows and best practices can bring a selection of suggested functions right to the user’s (metaphorical) fingertips. This not only helps guide new users in the right direction, but also helps veterans get to the features they need faster.

There is no reason for AI to be limited to simple suggestions either; with the right training it can help automate boring and repetitive tasks. Smart part selection and classification can drastically reduce the amount of time spent on traditionally slow and highly manual tasks while intelligent automations turn entire workflows into convenient functions. Through smart automation and intelligent assistance these AI systems can help design not only new products but also the machines that build them faster and more efficiently than ever before — perfect for meeting the needs of modern manufacturing.

Generative AI can help further streamline the design process by assisting designers in leveraging the full capabilities of engineering tools by better linking vast databases of expert knowledge and past designs with complex design tool functionality. This not only broadens possibilities for machine designers, but also accelerates the design phase and ensures strict adherence to all standards for quality, reliability, and cost.

Generative AI can also help to optimize simulation and planning by improving resource allocation and ensuring the most efficient use of materials, time, and people. Through a holistic view of the entire design process and the ability to pull on existing design knowledge, a generative AI system could be used to validate changes or ideas in a design quickly and at minimal cost, allowing for a more rapid design process.

Data is the foundation of the modern world. This holds true whether it be in our personal lives or in the operation of a $10-billion gigafactory. However, data itself is not inherently valuable, nor in the confines of a factory can its collection be taken for granted. Instead, its value derives from the insights and applications of that data. IT/OT convergence represents one of the major consumers of this factory data, offering fast, adaptive solutions fueled by the data generated by smart machines. But for these solutions to reach their full potential, machines must be designed from day one to support this data-centric view of the factory.

However, designing machines to accommodate thorough data collection and analytics can increase the difficulty of the already complex task of designing them. Moreover, it can potentially cause companies to incur short-term cost increases and slowdowns. But the benefits to enabling IT/OT convergence at a machine level are impossible to ignore. By feeding AI models with the data coming off connected machines, predictive maintenance and holistic optimizations become possible while, as previously stated, AI itself can also help alleviate the challenges of designing machines for this new manufacturing paradigm.

While placing a physical sensor in an optimal location is the ideal when it comes to designing for a digitally connected manufacturing floor, that is not always possible. Physical limitations and constraints can hamper the placement of sensors into real machines but, by leveraging AI and the digital twin, it’s possible to work around these limitations. By combining a machine with its digital twin, real sensors can be augmented with AI-enhanced virtual sensors, allowing information to be gathered from locations that would normally be impossible to access physically.

In this era of rapid technological advancement, the way industrial machinery is designed and used must be reimagined. AI offers the potential to revolutionize the design process by streamlining and automating existing processes while enabling a new paradigm of design space exploration. Building machines with AI in mind also presents new challenges of its own, such as enabling comprehensive data collection and analytics starting on the shop floor. By designing machines from day one with AI and smart systems in mind, they can leverage AI in the design process and across the machine’s entire lifespan. This ultimately helps manufacturing companies reach their sustainability, flexibility, and future-forward goals so they can achieve the next generation of smart manufacturing.

This article was written by Rahul Garg, Vice President of Industrial Machinery, Siemens Digital Industries Software (Plano, TX). For more information, visit here .