How AI supports smarter manufacturing
No matter what is being produced, making sure a factory is running smoothly and efficiently is of paramount importance. In an increasingly data-driven era, this means companies are now leverage AI to streamline maintenance, extract valuable insights from data and even package critical information into easy-to-use machine learning (ML) models.
In a recent podcast, guests Boris Scharinger, Senior Innovation Manager and Technology Strategist for Siemens Digital Industries, and Dr. Justin Hodges, Senior AI/ML Technical Specialist in Product Management for Siemens Digital Industries Software examined the various impacts of AI on the manufacturing and operations world. During this conversation, there was a particular focus on the topic of AI’s ability to serve as an interface between humans and complex data or systems, such as those found in a factory.
Check out the full episode here or keep reading for some of the highlights from that conversation.
AI streamlines factory operations
Keeping a factory running smoothly means managing a lot literal and metaphorical moving parts which, in turn, means a lot of data. This data takes many forms, ranging from live sensor data collected from smart machines to operations and maintenance manuals.
Both the collection and use of these various types of data present unique challenges in smart factories. AI offers a way to make data accessible on the shop floor for experts and non-experts alike while also streamlining the data collection process in many brownfield scenarios. Boris highlights AI vision as an area that has seen incredible advancement in recent years and not just in the ability for the models to accurately assess what they are seeing, but also in the way AI can help in the setup process – taking what used to require a team of machine vision experts and making it accessible to any engineer or technician.
As AI technology continues to mature, it may, itself, become the best way to deploy advanced technologies within specific domains such as machine vision on a production line. In this example, a separate AI model is used to help determine optimal part placement, lighting conditions, and camera settings when configuring an AI-based quality inspection system. A further supporting model, trained to identify features on parts, is then used to quickly teach the system what good and bad parts look like, allowing the entire deployment to be completed by a non-AI expert in under an hour.
Rethinking data packaging with AI
When it comes to the value of data, contextualization, packaging and usage are nearly as important as the raw information itself, while also being more difficult to implement efficiently. For example, while the current and historical information about how a machine is running is valuable, collecting that data into a model that simulates that machines operation or even a full digital twin has even greater value.
Packaging machine, part, and product data within AI models is one way to better incorporate both information and functionality into easy to share and use packages. Justin gives the example of using machine data to train an AI-powered reduced order model or ROM that could provide near instant access to detailed results on machine behavior all from a tablet on the shop floor. These kind of ROMs could do everything from simplifying maintenance to allowing operators to validate potential optimizations without the need for costly simulations and downtime.
Boris envisions taking this a step further, with suppliers packaging batch-specific data in the form of machine learning models along with their products, allowing manufacturers to calibrate machines for the specific characteristics of a certain lot of materials. Utilizing these models would help enable more efficient and less wasteful production, bringing manufacturing closer to the goal of right the first time production while supporting sustainability goals.
Data is both the lifeblood of AI, and the area which it can most help to improve existing processes and systems. By combining AI with the many types and vast quantities of data available within both a factory and the breadth of the manufacturing process, companies will be able to not only improve efficiency and support sustainable practices but open up new, undiscovered avenues of innovation as well.
To find out more, check out the full podcast here.
Siemens Digital Industries Software helps organizations of all sizes digitally transform using software, hardware and services from the Siemens Xcelerator business platform. Siemens’ software and the comprehensive digital twin enable companies to optimize their design, engineering and manufacturing processes to turn today’s ideas into the sustainable products of the future. From chips to entire systems, from product to process, across all industries. Siemens Digital Industries Software – Accelerating transformation.
