VALID3D | Valid generative design for 3D printing

Region:

Flanders

Financed by

Towards automated quality control in metal additive manufacturing

Metal additive manufacturing (AM), particularly laser powder bed fusion (LPBF), offers many advantages such as flexibility, lightweighting and cost reduction. However, its adoption in demanding sectors such as aerospace and medical remains limited. The main reason is the lack of reliable and accessible tools to guarantee the quality of critical parts. Today, quality control can represent up to 50% of manufacturing costs. VALID3D addresses this challenge by developing an automated, data-driven approach covering the entire production chain.

Target group

The project is primarily aimed at:

Aerospace companies producing critical parts
Medical device manufacturers requiring certified parts
Additive manufacturing service providers seeking to improve traceability
Suppliers of AM software and equipment

Context

The global metal additive manufacturing market is rapidly expanding, with more than 21,000 machines installed and over 230 companies active in equipment, software, materials and services. This growth is accompanied by a diversification of players and increasing requirements in terms of quality and traceability, particularly in high-value sectors such as aerospace and medical.
Several challenges remain. Quality control costs are still high. Data management is complex due to the volume and heterogeneity of data. Finally, the lack of advanced tools and suitable software solutions limits the ability to guarantee reliable and compliant parts, especially for critical applications or personalised medical devices.

Objectives & results

VALID3D aims to improve quality control and process understanding in metal additive manufacturing through automation and artificial intelligence.

Project objectives

Automate the ingestion and integration of data from multiple AM process sources
Reduce manual work related to quality monitoring by more than 90%
Accelerate root cause analysis (RCA) using AI, from several weeks to a few minutes
Replace up to 70% of non-destructive (NDT) and destructive testing (DT) with digital analyses, leading to manufacturing cost reductions of up to 50%

Key results:

Automated tools for data ingestion and processing
Artificial intelligence modules for quality monitoring and defect analysis
Digital methods to reduce DT and NDT inspections
Reports and guidelines on efficiency gains and industrial impact

For a typical additive manufacturing unit in the aerospace sector (10 machines, 2 productions per week per machine), these advances translate into an efficiency gain of more than 90% (around €160k/year) and a reduction of NDT and DT-related costs of more than €1.1M/year. They also allow engineers to focus on higher value-added tasks and support the transition towards a more digital and intelligent industry.

VALID3D 2 pictures — Image 1: LPBF machine in operation at Sirris / Image 2: AMIQUAM eddy current system integrated in the Sirris LPBF machine

Approach

The project is structured into five work packages:

WP1: Management and dissemination
Project coordination, external communication, knowledge transfer, valorisation of results and definition of best practices for data exchange and quality
WP2: Process integration
Harmonisation of data formats, mapping of process steps and connection of systems across the entire additive manufacturing chain, data collection and structuring
WP3: Technical and system developments
Development of automated process data evaluation, integration of new sensors, inter-layer control and open monitoring tools
WP4: Quality and artificial intelligence
Use of collected data to model product quality using AI, training of predictive models, use of data from various applications
WP5: Use cases
Specification, implementation and validation of application cases to demonstrate the overall effectiveness of the project