QIMAL: a data-driven inspection system for the optimization of the visual inspection process in Metal Additive Manufacturing

Additive manufacturing (AM) technology is widely regarded as one of the most promising advanced manufacturing techniques. Compared to traditional methods like CNC tools, forging, and welding, AM offers several advantages, including the absence of tooling or molds, high material utilization, shorter product manufacturing cycles, and the ability to create intricate structures.

All AM systems follow the same principle of building structures layer by layer, and certain processes even allow printing with diverse materials such as polymers, metals, ceramics, and composites. AM is particularly well-suited for low-cost, rapid prototyping, and the production of large and complex metal structures.

 

The AM Challenge related to Quality

It is important to note that AM, on its own, may not consistently produce parts with the desired mechanical properties and surface quality required for many applications. AM components often exhibit various defects, such as powder agglomeration, balling, porosity, internal cracks, and thermal/internal stress, which can significantly impact the final part’s quality, mechanical properties, and safety.

To address these challenges, defect inspection methods play a crucial role in minimizing manufacturing defects, monitoring the process during manufacturing, improving surface quality, and ensuring the desired mechanical properties of AM components. By combining traditional defect detection techniques with high-resolution visual imaging, defect detection and fault prediction and diagnosis technology can be optimized, leading to enhanced overall quality and performance in additive manufacturing processes.

 

Why the Metal Additive Manufacturing processes?

Selective laser melting (SLM) is the premier metal 3D printing technology revolutionizing the additive manufacture of metal parts.

Selective laser melting (SLM) processes offer the ability to create functional metal parts with intricate geometries close to the final shape. However, these processes may suffer from imprecision and the formation of defects due to inherent randomness and irregularities associated with laser powder fusion. These defects can lead to 3D print failures, resulting in wasted time, materials, and resources. In large-scale industrial additive manufacturing (AM), the identification and mitigation of defects during the design phase become even more critical.

To tackle these challenges, the key lies in exploring the integration of process data from machines, CAD-simulation designs, and high-definition images within the manufacturing chamber during operations. Additionally, physical validation through laser scanning or mechanical tests can be employed. This rich data can serve as feedback for an in-situ control system and can be combined with existing defect recognition to create an AI-driven descriptive and prescriptive expert system. Such a system can then adjust processes or improve designs accordingly.

 

An effective process system should be able to:

1. Pinpoint defects in each processing layer accurately.
2. Detect any irregularities within the 3D CAD model.
3. Establish correlations between processing parameters and the occurrence of defects.

 

 

About QIMAL

Youbiquo, as part of the Reach Incubator Open Call, has proposed QIMAL, a cutting-edge data-driven in-process monitoring system that combines the analysis of final product images with 3D CAD design and machine operating parameters. Our innovative solution enables the identification of defects’ root causes and provides detailed data analysis within a trusted value chain platform.

QIMAL has been specifically designed to support machine operators, including Process Engineers and Application Engineers, who work with layer-based metal additive manufacturing. As part of their responsibilities, they analyze the 3D CAD design to be printed, adapt it for the chosen 3D printing process, and consider the customer’s material requirements. Additionally, they must consider the CAD topology to design the printing strategy, including structural support. The calibration of the machine and initiation of the printing job also fall under their purview.

The QIMAL solution aims to identify and address defects in real-time during the production process. Moreover, it thoroughly investigates and determines the root causes of these defects. Our monitoring system is designed to achieve the following objectives:

1. Optimize the printing process by comprehending the sources and occurrences of defects during manufacturing.
2. Enhance process efficiency through the development of descriptive-predictive-prescriptive data-driven models.
3. Elevate the level of sustainability in the manufacturing processes.

By attaining these objectives, we can significantly enhance the quality and reliability of metal parts produced through Selective Laser Melting (SLM) or Powder-Bed Fusion (PBF). Ultimately, this will maximize the efficiency and success of industrial additive manufacturing, making it more reliable and competitive in the market.