Ingenieurwissenschaften und zugeordnete Tätigkeiten
Filtern
Dokumenttyp
- Beitrag zu einem Tagungsband (2)
- Vortrag (2)
Sprache
- Englisch (4) (entfernen)
Schlagworte
- Thermografie (4) (entfernen)
Organisationseinheit der BAM
Eingeladener Vortrag
- nein (2)
The formation of defects such as keyhole pores is a major challenge for the production of metal parts by Laser Powder Bed Fusion (LPBF). The LPBF process is characterized by a large number of influencing factors which can be hard to quantify. Machine Learning (ML) is a prominent tool to predict the outcome of complex processes on the basis of different sensor data. In this study, a ML model for defect prediction is created using thermographic image features as input data. As a reference, the porosity information calculated from an x-ray Micro Computed Tomography (µCT) scan of the produced specimen is used. Physical knowledge about the keyhole pore formation is incorporated into the model to increase the prediction accuracy. From the prediction result, the quality of the input data is evaluated and future demands on in-situ monitoring of LPBF processes are formulated.
In this paper shortwave infrared (SWIR) thermographic measurements of the manufacturing of thin single-line walls via laser metal deposition (LMD) are presented. As the thermographic camera is mounted fixed to the welding arm, an acceleration sensor was used to assist in reconstructing the spatial position from the predefined welding path. Hereby we could obtain data sets containing the size of the molten pool and the oxide covered areas as functions of the position in the workpiece. Furthermore, the influence of the acquisition wavelength onto the thermograms was investigated in a spectral range from 1250 nm to 1550 nm. All wavelengths turned out to be usable for the in-situ process monitoring of the LMD process. The longer wavelengths are shown to be beneficial for the lower temperature range, while shorter wavelengths show more details within the molten pool.
By allowing economic on demand manufacturing of highly customized and complex workpieces, metal based additive manufacturing (AM) has the prospect to revolutionize many industrial areas. Since AM is prone to the formation of defects during the building process, a fundamental requirement for AM to become applicable in most fields is the ability to guarantee the adherence to strict quality and safety standards. A possible solution for this problem lies in the deployment of various in-situ monitoring techniques. For most of these techniques, the application to AM is still very poorly understood. Therefore, the BAM in its mission to provide safety in technology has initiated the project “Process Monitoring of AM” (ProMoAM). In this project, a wide range of in-situ process monitoring techniques, including active and passive thermography, optical tomography, optical emission and absorption spectroscopy, eddy current testing, laminography, X-ray backscattering and photoacoustic methods, are applied to laser metal deposition (LMD), laser powder bed fusion and wire arc AM. Since it is still unclear which measured quantities are relevant for the detection of defects, these measurements are performed very thoroughly. In successive steps, the data acquired by all these methods is fused and compared to the results of reference methods such as computer tomography and ultrasonic immersion testing. The goal is to find reliable methods to detect the formation of defects during the building process. The detailed acquired data sets may also be used for comparison with simulations.
Here, we show first results of high speed (> 300 Hz) thermographic measurements of the LMD process in the SWIR range using 316L as building material. For these experiments, the camera was mounted fixed to the welding arm of the LMD machine to keep the molten pool in focus, regardless of the shape of the specimen. As the thermograms do not contain any information about the current spatial position during the building process, we use an acceleration sensor to track the movement and synchronize the measured data with the predefined welding path. This allows us to reconstruct the geometry of the workpieces and assign the thermographic data to spatial positions. Furthermore, we investigate the influence of the acquisition wavelength on the thermographic data by comparing measurements acquired with different narrow bandpass filters (50 nm FWHM) in a spectral range from 1150 nm to 1550 nm.
This research was funded by BAM within the Focus Area Materials.
The aim of this work is to illustrate the contribution of signal processing techniques in the field of Non-Destructive Evaluation. A component’s life evaluation is inevitably related to the presence of flaws in it. The detection and characterization of cracks prior to damage is a technologically and economically significant task and is of very importance when it comes to safety-relevant measures. The Laser Thermography is the most effective and advanced thermography method for Non-Destructive Evaluation. High capability for the detection of surface cracks and for the characterization of the geometry of artificial surface flaws in metallic samples of laser thermography is particularly encouraging. This is one of the non- contacting, fast and real time detection method. The presence of a vertical surface breaking crack will disturb the thermal footprint. The data processing method plays vital role in fast detection of the surface and sub-surface cracks.
Currently in laser thermographic inspection lacks a compromising data processing algorithm which is necessary for the fast crack detection and also the analysis of data is done as part of post processing. In this work we introduced a raw data based image processing algorithm which results precise, better and fast crack detection. The algorithm we developed gives better results in both experimental and modeling data. By applying this algorithm we carried out a detailed investigation Variation of thermal contrast with crack parameters like depth and width. The algorithm we developed is applied for various surface temperature data from the 2D scanning model and also validated credibility of algorithm with experimental data.