TonyBagnall
commented
3 years ago great, thanks for the donation, I'll post here when it is on the website
Open vivekmahato opened 3 years ago
TonyBagnall
commented
3 years ago great, thanks for the donation, I'll post here when it is on the website
Please give a brief description of the data
The data comes from the AconityMINI 3D printer during the manufacturing of stainless steel blocks with a designed cavity. The temperature data is collected, and it comes from pyrometers that monitor melt pool temperature. The pyrometers supplied by KLEIBER Infrared GmbH detect the heat emission light in the range of 1500 to 1700 nm via two detectors. The pyrometers track the scan of the laser to provide a time-series sampled at 100 Hz.
The dataset is sampled from the mid-section of these blocks and is organized into two datasets (Plarge and Psmall). The Plarge dataset covers cubes with large pores (0.4 mm, 0.5 mm, and 0.6 mm). For these cubes, the laser is set at 180 W, the scanning speed was 800 mm/sec and the layer thickness was 20 μm. The Psmall dataset covers cubes with small pores (0.05 mm and 0.1 mm) with the same process parameters except the thickness is set at 40 μm.
Where can we get the data?
https://github.com/vivekmahato/mlots/tree/main/input/AM_Datasets
The datasets are named: AM_Plarge.csv, and AM_Psmall.csv.
Is there a publication related to the data
https://researchrepository.ucd.ie/handle/10197/11740
http://ceur-ws.org/Vol-2771/AICS2020_paper_31.pdf
Is your data multivariate or univariate? Univariate.
Is there a default train/test split for the data A stratified random split of 67-33% works.
Are there any missing values/padding? The time-series have been post-padded with 0's to bring each sample to 300 data-points.
Is there a picture for the data we can put on the website?
Fig. 2(a) in the paper: https://researchrepository.ucd.ie/handle/10197/11740
How would you like the donation attributed on the website?
The dataset comes with a ODBL license.
Can you provide a paragraph description, including the meaning of the class values, for the website?
The IR data samples represent raster scans during 3D printing of these stainless steel blocks. If a raster scan passes through the cavity, it is labeled as 1 (or porous). The non-anomalous raster scans are therefore labeled as 0.