Over the last few weeks we’ve been exploring the “factory of the future”, looking at trends such as automation and how the technologies featured on the TCT Show floor last month (where we delved further into these trends during a panel session on 3D printing and smart factories) are contributing towards Industry 4.0.
The connection between manufacturing hardware and software is a key part of that factory of the future vision. The ability to seamlessly communicate between various steps of the manufacturing process, which more often than not are operating in silos, and feed that data back to create smarter and more efficient workflows, is something that manufacturers are becoming increasingly aware of but remains in the early stages of deployment.
San Francisco-based, Oqton is aiming to overcome today’s manufacturing challenges with an AI (artificial intelligence) driven factory operating system. Founded by a group of manufacturing and AI experts, Oqton has developed a cloud-based platform, FactoryOS, linking data from design to delivery by combining real production data with simulations and applying AI techniques to optimise production and efficiency.
Ben Schrauwen, CTO and co-founder of Oqton, formerly a director for Autodesk’s Spark 3D printing platform, spoke to TCT about the digitisation of manufacturing and current and future realities of autonomous factories.
Oqton FactoryOS connects manufacturing software and hardware — what challenges are you aiming to tackle in the manufacturing environment with this platform?
The aim of Oqton is to bring the vision of autonomous manufacturing to reality, and that the path to finally doing that is through tightly connecting manufacturing software and hardware. Smart manufacturing is still in its infancy, but across the industry people see the value and are looking for ways to implement it. Most of the solutions that are on the market right now are either monolithic, requiring costly and complex implementation and replacement of existing systems, or only focus on solving point manufacturing challenges. Manufacturing processes remain in silos, so even though the volume of data available from production is increasing, making it useful across the interdependent workflows remains unsolved.
Part of the problem is the air gap between manufacturing software and hardware. Though technology is rapidly increasing in both areas, they aren’t connected. Manufacturing software doesn’t talk directly to the hardware, there are layers of translation and human intervention in between. And what’s happening on the production floor doesn’t loop back and inform future work in the software. Additive is an area where you can clearly see the impact of this problem manifest and impact adoption. Because of this gap, the benefits that we expect from this innovative manufacturing technology are cumbersome to achieve, or fail to deliver at the level of expectation.
You were previously at Autodesk and recently appointed Carl Bass to the Oqton board — what did you learn from your time at Autodesk about the digitisation of manufacturing?
Our time at Autodesk gave us perspective on what was happening at the forefront of manufacturing and allowed us to hear directly from manufacturers of all sizes about the challenges and disruption in the industry. Autodesk sits on the forefront of CAD and cloud technology and is actively exploring new and disruptive technologies in the space, so that gave us a deep appreciation of the need for companies and software providers to adapt to the new ways that technology can be used to make manufacturing more efficient. It also let us identify an opportunity to address this gap between manufacturing software and hardware that isn’t being directly addressed by others in the industry. We were able to leverage this knowledge to take a new perspective in a startup environment where we can operate with the nimbleness and dedicated focus that isn’t always possible in a large enterprise.
Who is going to benefit from this platform? Major manufacturers, service providers?
Manufacturers of all sizes stand to benefit from the Oqton FactoryOS platform, we see closed-loop data as the driving force behind the factory of the future. Of course, manufacturers who have already or who are looking to digitise their operations, early adopters of advanced manufacturing technologies such as AM, will be the first to benefit because they’re already on the path, we will help them accelerate the outcomes they are working toward.
Because FactoryOS allows companies to integrate information from across the manufacturing ecosystem, we also see that small and medium factories will be able to capitalise on these insights to become much more agile and efficient in a way that is currently only accessible to the big players who have the correspondingly big budgets to apply to it.
There is a fear of automation in the manufacturing world. What do you say to counteract that fear and what effect will this level of automation have on skilled engineers?
Historically, every impending shift in how people will work has instilled fear, and this is especially the case with automation. People are afraid automation will replace human positions. This is true, but if you look at the last 100 years of evolution in manufacturing automation, what you see is that automation is replacing tedious, dangerous, and low-skill jobs and making way for alarger number of skilled positions that increase the value of the workforce rather than reducing it. In this current phase of this evolution, fear is based on the idea that AI will out-think people. But we see automation being the most effective is when AI technology is combined with human decision making, not replacing it. Humans are able to make creative and complex trade-off decisions in ways that AI simply cannot, despite the science fiction. But automation can allow skilled engineers to make better and more informed decisions by aggregating data gathered from the entire production environment in a way that people would never be able to. I believe this technology will be incredibly powerful when people are using it to augment decision making rather than expecting AI function on its own.
In terms of AM, how does this “AI-driven factory” affect the various parts of the AM process from file, to fabrication, post-processing, etc?
With additive manufacturing technology right now, there is a bottleneck due to a lack of available additive-specific engineering talent. The gap between what software defines in the digital model and how the build performs on the hardware in the real world has to be filled by human expertise and trial-and-error processes. Metal additive in particular is incredibly complex, and these specialised engineers and operators know from thousands of hours of experience how to set up a build to meet quality and cost specifications, or what parameters might cause a particular build and machine combination to fail. But there aren’t enough of these experts to scale AM as a production method, and frankly repeating the same insights is not a good use of human ingenuity. AI can help alleviate this by continually monitoring and advising the build setup and fabrication process to alleviate the skills bottleneck.
There is a misconception that any component can be 3D printed, but if a manufacturer is looking for a specific level of tolerance, this isn’t always the case. AI technology can give an upfront assessment of manufacturability and can provide insights about how to make a product more manufacturable prior to beginning the fabrication stage.
Metal additive is also quite expensive, so it’s important to understand in advance what possible complications may exist. One of the capabilities of the Oqton platform is leveraging AI to provide an analysis of the printing process as a whole which can help decision makers understand expected yield and cost assessments.
How does the software learn? Are you doing any testing with industry users to gather data and validate?
Manufacturers are typically hardware focused, and fairly inexperienced when it comes to software development, which has resulted in the current gap between manufacturing hardware and software. With many manufacturing solutions, the software is created somewhere else and then implemented into the factory, but because it was developed in a silo it doesn’t interact closely with the machines and get the most out of the machine’s capabilities. Sure, today’s software technology is delivering process improvements, but they’re a fraction of what would be possible if the software and hardware were tightly paired. Oqton is closing this gap by working directly with global hardware partners to train and improve our models, taking a collaborative approach where the software is developed already integrated with the manufacturing hardware.
The Oqton FactoryOS platform then continues to learn and improve directly from real data in the production environment. The software is also continuously interacting with its users who, by deciding a course of action from the suggestions the platform provides, further improve the models over time.
What do you make of additional technologies like blockchain for the AM world – are they necessary/ is the AM industry at that level yet?
I’ve been bullish on the potential for blockchain technology in manufacturing for a while now. There’s incredible potential for this technology because the manufacturing industry is weighted down by cumbersome workflows developed strictly to ensure trust between players and processes, and what blockchain technology does is create a system in which trust is inherent and free. Specifically in additive, it is feasible to produce a part anywhere in the world, close to where the end customer is. Blockchain can be used to create a smart contract between the designer, the local producer of the part and the end customer. That smart contract can then ensure quality parts are delivered in time, as a decentralised escrow mechanism can be set up ensuring every player is incentivised to play his part in the transaction. Blockchain has the potential to be applied to manufacturing at large, not just for additive, but it’s in the early stage of the hype cycle and the technical foundation still has a way to go before it can be applied effectively.
When it comes to the factories of the future we’re seeing manufacturers looking at automation, connectivity, virtual reality and related technologies — what does it look like to you?
When the first wave of digitisation came through manufacturing, and the industry saw a large increase in visibility into production. Data from throughout the process has become much more accessible. But the next wave of innovation will make sense of all this data; extracting insights and making the data actionable, right where it is needed.
The factory of the future will apply data in new ways. Leveraging augmented reality and virtual reality, factories could start to see insights layered on top of reality right where and when it is most needed. Displaying data in real time within this physical context will make the factories of the future that much more efficient and agile, giving its workers the ability to make much more informed decisions.