CASE STUDIES
Data Registration and Machine Learning for Anomaly Detection
Executive Summary
Machine learning is a method of data analysis based on the idea that systems can learn from data, identify patterns and make decisions with minimal human intervention. Sigma Labs’ PrintRite3D® software uses machine learning to identify patterns in additive manufacturing processes to identify and predict anomalies in part builds. It can also be used to measure the accuracy of the diagnostic testing that analyzes part quality.
Additive manufacturers often rely on expensive CT scanning to analyze part quality. But because metal 3D printing typically uses powders that have high radiodensity, CT testing does not always yield accurate results, especially when a part has complex geometric structures within it. Sigma Labs has developed a method to ensure part quality with results comparable and complementary to CT testing, as demonstrated in a previous case study.
In this case study, Sigma Labs registered and mapped anomalies in a part already identified by CT scan in the PrintRite3D® software. This trained a machine learning algorithm to recognize patterns in PrintRite3D® metrics that corresponded to anomalies in the CT data. These patterns were then used to predict the anomalies found by CT scanning, and others not found by scanning.
The study found that with sufficient data, Sigma’s PrintRite3D® software, using machine learning, can accurately predict the size and position of anomalies, and can be used as a new, near-real time quality measurement to supplement other measurements, like CT scanning.