Open Data Challenge

In-situ quality process monitoring in Additive Manufacturing

An Open Science project collaboration between Trumpf GmbH and Politecnico di Milano

Open Data Description

Additive Manufacturing (AM) processes and technologies have experienced continuous growth in their adoption across a wide variety of industrial sectors. They have impacted, among other domains, the biomedical, aerospace, racing and automotive, oil and gas, tooling and molding, and creative industries. From a statistical process monitoring (SPM) perspective, the AM paradigm entails a layerwise production that enables the in-line and in-situ collection of a vast amount of signal data. These data can be used to determine process stability and to accelerate the detection of anomalies and defects during the manufacturing process.

The Open Science project focuses on in-situ anomaly detection in Laser Powder Bed Fusion (L-PBF) [1]. Among diverse sensing configurations that are available for this process, a very effective and largely studied configuration consists of using the optical path of the laser to measure the radiation emitted by the melt pool and its surroundings (Fig. 1). The melt pool is the region where the laser beam exposure melts the material, and it is known to be a primary feature of interest in any process that involves a beam-material interaction aimed at achieving a local melting of the material.

Fig. 1 – Left: The co-axial monitoring setup that utilizes two photodiodes aligned to the optical path of the laser. Only the InGaAs photodiode is considered for this dataset. Right: Schematic side and top views of the melt pool along the laser scan direction.

The Open Data is available to anyone is interested in working on real data for the study and development of in-situ monitoring solutions. Here below, you can find detailed information about the dataset and istructions about how to access the data.

This dataset has been designed to develop and test in-situ process monitoring solutions that can detect such anomalies as soon as possible while achieving the best compromise in terms of false positives and false negatives.

Please cite the dataset as follows:

“The dataset is part of an Open Data Science project between Trumpf GmbH and Politecnico di Milano and it is available at”.

[1] An illustrative video of the L-PBF process:



Marc Gronle, Frederik Schaal

TRUMPF Laser- und Systemtechnik GmbH, Johann-Maus-Straße 2, 71254 Ditzingen, Germany,

Politecnico di Milano

Bianca Maria Colosimo, Marco Grasso, Emidio Granito

Politecnico di Milano, Dipartimento di Meccanica, Via La Masa, 1, 20156 Milano, Italy,

2021 QSR Open Data Challenge Competition

The Open Data made available by Trumpf and Politecnico di Milano is used for the 2021 QSR Open Data Challenge Competition. More information can be found here.