Business Reinforcement by integrating Artificial Intelligence in New advanced manufacturing Solutions

In a time marked by relentless technological progress and its resulting competitive advantages, the manufacturing sector leads the charge in innovation. However, the full potential of artificial intelligence (AI) remains underutilised. Building upon the successful outcomes of the PRISMA project (ERDF, October 2019 – April 2022), BRAINS aims to advance knowledge and insights in data learning, including Artificial Intelligence (AI). This project centres on facilitating the digital transition and automation within the manufacturing industry, progressing from manual to semi-automatic processes, and ultimately to fully automated systems, including machine learning automation.

While PRISMA primarily concentrated on vision technology development – encompassing data acquisition from production processes or machines through camera imagery and sensors, data processing, analysis, and subsequent adjustment of process parameters – BRAINS shifts its focus towards advanced process parameter adjustment, integrating vision and sensing technologies. It spans from intelligent operator instructions to automatic machine adjustments and self-learning process optimisations. This endeavour entails researching and developing new AI knowledge, culminating in the creation of a versatile data learning module.

The consolidation of this knowledge into a new data learning module is pivotal, catering to multiple industrial use cases within process automation. Through a consortium comprising companies and knowledge institutes, BRAINS addresses six distinct use cases, each focusing on specific challenges from various companies. This collaborative approach brings together expertise and experiences from different sources, helping to thoroughly explore possible solutions.

Project Consortium

Logo AWL
Logo Tembo


Use cases are valuable tools for evaluating and understanding systems, like software platforms. They help define, interpret, and organise system requirements, functions, and uses. For example, a manufacturing company implementing a new inventory management system might utilise use cases to outline inventory tracking procedures and anticipate potential bottlenecks in the production line.

In the context of BRAINS, the project examines six specific use cases focusing on automation, quality control, and process optimisation within manufacturing. By collaborating with industry leaders and leveraging their expertise, BRAINS aims to develop scalable AI solutions tailored to the unique challenges faced by manufacturing environments.

Tembo’s use case addresses the challenge of varying sealing strength leading to leaking pods, causing system shutdowns and cleaning. Their approach focuses on two main objectives: optimising parameter control to minimise leaks by determining and measuring influencing parameters, and reducing machine footprint by optimising relevant design variables and process parameters.


Zuidberg’s use case aims to optimise the PTO assembly process by collecting and using real-time process data for operator instructions. In particular, it looks at using AI to improve and simplify quality control for use with multiple product types.


AWL’s use case focuses on creating jig-free machine concepts for joining products, specifically through welding or assembling individual parts. Their proposed solution involves utilising a robotic system equipped with simple grippers, each capable of picking up its assigned part and bringing them together. Technical challenges include ensuring the absolute accuracy of the robots, generating optimal robot paths, developing flexible gripper concepts, and effectively connecting products using welding equipment or assembly tooling. Additionally, a quality inspection process will be implemented at the end of the assembly process, utilising vision technology and artificial intelligence for thorough quality control.


Bond3D’s use case focuses on improving the flow control of material during 3D printing, particularly for printing PEEK material whose physical properties are highly complex to model. The BRAINS project aims to replace the existing simplified model with a machine learning model which can also take this complexity into account. Because an accurate model is at the basis of good flow rate control performance, the new model will allow for more precise and consistent printing of PEEK material, ultimately enhancing product quality and manufacturing efficiency.


Bond3D’s other use case focuses on optimising autofill geometry during printing to improve efficiency, especially for long-duration prints where even small optimisations can save significant time. While standard geometries like squares and rectangles are well-optimised, the challenge lies in efficiently filling exotic shapes that are less common.


Vincent Blokhuis

Engineering Manager