A new research initiative backed by the Defense Advanced Research Projects Agency is set to reshape inspection and qualification processes in additive manufacturing, with a strong focus on reducing reliance on destructive testing.
The project, led by Associate Professor Dazhong Wu at the University of Central Florida, aims to develop an artificial intelligence-driven framework for predicting defects and mechanical performance in 3D-printed components. The effort has been awarded funding of nearly $500,000 under DARPA’s Young Faculty Award programme, with the potential for an additional $500,000 based on progress.
Additive manufacturing, widely used for producing complex and lightweight components across aerospace, automotive, and healthcare sectors, continues to face challenges in part qualification. Current inspection and testing processes are often time-consuming, costly, and heavily dependent on destructive methods, particularly for high-value materials such as titanium alloys.
The proposed AI-enabled approach seeks to address these limitations by enabling predictive inspection capabilities. Instead of relying on repeated physical testing cycles that result in part destruction, the model is designed to identify defects and assess performance characteristics through data-driven analysis. This shift aligns with broader industry efforts to integrate intelligent, non-destructive evaluation (NDE) techniques into advanced manufacturing workflows.
By reducing the need for trial-and-error testing, the framework is expected to lower inspection costs while improving consistency in part qualification. The approach also supports scalable inspection strategies, which remain a key barrier to wider adoption of additive manufacturing in safety-critical industries.
“I’m hopeful this AI-enabled additive manufacturing qualification framework will be used across many industries, including aerospace and, many more,” Wu says. “Bringing costs down is crucial to the additive manufacturing industry. To do that, we need to make sure every part consistently meets performance requirements.”
The research highlights a growing convergence between artificial intelligence and non-destructive testing, where predictive models are increasingly being explored to complement or replace traditional inspection methods. If successful, the initiative could contribute to faster, more reliable qualification processes, enabling broader industrial deployment of additive manufacturing technologies.
