The last decade has seen manufacturing change at a staggering rate. Production series are shrinking, product lifespans are shortening, and custom designs are becoming the norm. This trend is challenging conventional manufacturing planning software to unprecedented levels.
To manage large-scale production workflows efficiently, factories normally use a Manufacturing Execution System (MES), which plans, executes and monitors the fabrication of goods. An MES is well-equipped to handle large series production, but smaller batches require a never-before-seen degree of flexibility. In addition, an MES is configured to execute a single, fixed plan and struggles to accommodate rescheduling or reworks. It’s often oblivious to changes on the shop floor because the planning, execution and monitoring are isolated, and organized as a waterfall. Deviations from the plan are resolved manually and an attempt to shoehorn the manual changes into an existing system would likely result in an inefficient workflow.
For a genuinely flexible production setup that can accommodate new technologies, you need a system that brings together information from various machines and applications in a unified, easy-to-configure solution that will both detect issues and suggest solutions. Our approach to such a system is called a Manufacturing Operating System (Manufacturing OS). It is a natural evolution of an MES with a few key differences.
A Manufacturing OS brings together information from various machines and applications in a unified, easy-to-configure solution that will detect issues and suggest solutions.
To make workflows fast and flexible, it’s essential to radically reduce their reliance on human input. Whether it’s an AM engineer, a dental lab technician or a robot programmer, staff are still an indispensable element in a lot of manufacturing operations. Yet these skilled and hard-to-find professionals often have to perform repetitive, non-value-adding tasks that should be automated in a manufacturing workflow.
With a Manufacturing OS you can automate processes at different levels of manufacturing. For example, programming welding robots was until recently a tremendously time-consuming task. A highly skilled engineer would spend hours using complex software packages to guide the robot safely through the desired welding sequence. Oqton’s Manufacturing OS now automates that process, generating robot programs directly from the 3D CAD data.
Similarly, dental labs can now autonomously print thousands of parts a day, as we automatically recognize part categories and surface features to orient the part, generate support structures, nest the parts and generate instructions for the 3D printing and CNC milling.
Having the right information at the right time is essential if you want to rapidly adapt production workflows. Usual industrial internet of things (IIoT) platforms go a long way to achieving this, with sensors, instruments, and other devices connected to dashboards showing a machine's performance. In theory, having this information at our fingertips should help us increase productivity and efficiency. In reality, a few crucial elements are missing.
While we do track information about machinery involved in manufacturing, in an MES we don’t have a simple way of extracting insights and acting on them. Reports are frequently siloed, languishing as a file in a database no one ever looks at.
The idea behind a Manufacturing OS is to gather all the data from different sources in one place that other systems would access and use to generate suggestions for improvements. To work together seamlessly, systems need to share data, speak the same language and understand every role in the bigger picture. That is exactly what the Manufacturing OS aims to do.
In inspection, additional contextual data could help us move beyond a basic understanding of how a part is deformed towards identifying the root cause of deformation and how to avoid it. Think of quality inspection software like Control X that generates reports based on 3D scans of a part which flag up any problems. By integrating Control X into the Manufacturing OS, we could link its data with the production details about the part, and get a deeper understanding of the deformations. For 3D printed parts, for example, we could see if a deformation is directly linked to their orientation or position on the build plate; or see if a certain part of the machine is degrading over time and producing more defective parts.
We also see great potential to improve quality in robotic welding by adding more contextual data to the IIoT. We’ve installed sensors on welding robots to monitor the voltage and current, and microphones to listen to the weld. Because multiple systems share information in the Manufacturing OS, we can also complement the sensor data with the CAD model of the weld, the welding parameters, the thickness of the plates, and the position of the welding torch. All this gives us a picture of exactly what the robot is doing and allows us to make an informed decision on how to adapt.
A Manufacturing OS could give us a full picture of what a welding robot is doing by connecting information that is usually siloed.
In a modern, agile production environment, it’s impossible to predict everything. New products are constantly being added and workflows are rapidly evolving. A traditional MES has been designed to handle predefined workflows and is difficult to (re)configure. As a consequence, companies normally need to hire expensive consultants to set up their MES and, if you’re a small business, you may not have the necessary budget to even get an MES solution.
We’re developing the Manufacturing OS to be easily configurable and re-configurable. Our goal is to provide manufacturers with the building blocks for the platform, which their staff would use to set up the workflow independently.
In a market that is shifting towards short product cycles, configurable workflows and interfaces help manufacturers deliver orders swiftly. Since every new product requires a new production route, and establishing the best one is often an iterative process, an easily reconfigurable setup is helpful in reducing the time it takes to make the necessary adaptations. When it comes to interfaces, configurability gives users the ability to personalise the settings for quick and easy access to information that is most relevant to a specific workflow.
Configurability is especially important for production routes in advanced manufacturing. An important way of increasing productivity in 3D printing is nesting, the process of arranging and orienting many parts on the machine’s build plate in a way that achieves optimal use of space. Easy-to-configure nesting means having a quick setup for whole part families, with specific geometric, material and finish variants, which ultimately enables the mass production of custom parts.
Oqton is empowering Belgian start-up Amnovis to achieve high-speed and flexible production of high-end AM products.
With configurable routes and interfaces, enhanced contextual information and autonomy, a Manufacturing OS prepares manufacturers for short product cycles and mass customisation. While it retains a few MES features, with AI and IIoT, a Manufacturing OS is more capable of adopting new technologies into manufacturing workflows and making production leaner. In addition, as we continue to collaborate with customers on specific solutions, the Manufacturing OS will become better at supporting a wider range of segments and industries.