Combined with simulation, in-process monitoring for additive manufacturing can improve quality and reduce waste, among other benefits.
One of the biggest challenges for the wider adoption of additive manufacturing (AM) as a production technology is the lack of repeatability and stability in AM processes. Real-time, in-process monitoring and closed-loop control systems can go a long way toward overcoming this barrier, though there is still much work to be done in their development.
What techniques are used for monitoring 3D printing processes?
While the particular 3D printing process in question will determine the most applicable monitoring techniques, this article will focus on discussing thermal processes, such as laser powder bed fusion (L-PBF), electron beam melting (EBM), directed energy deposition (DED) and fused deposition modeling (FDM). However, it should be clear that some of the techniques discussed below can also be applied to other additive processes, such as vat photopolymerization or binder jetting.
Examples of process sensing techniques that can be used for real-time monitoring of AM include visible-light and thermal sensing, integrated sphere radiometry, acoustic detection, Schlieren imaging and operando synchrotron X-ray imaging and diffraction.
Here’s a brief explanation of each technique as well as its strengths, limitations and applications:
Technique | Description | Strengths | Limitations | Applications |
Visible-Light and Thermal Sensing | Cameras monitor each print layer. Infrared cameras capture heat distribution and temperature variations during the process. | Non-invasive method of providing real-time data about layer adhesion and potential warping. | Visible-light cameras cannot identify internal defects and thermal results may be noisy due to room temperature fluctuations. | Used in both metal and plastic 3D printing, especially in powder bed fusion. |
Integrated Sphere Radiometry | Measures total radiative output during the printing process. | Useful for capturing total energy output, which is useful for understanding overall process stability. | Less effective for detecting local defects and microstructural variations. | Used in high-energy processes, typically involving metal 3D printing. |
Acoustic Detection | Acoustic sensors capture soundwaves generated by the printer to identify changes in material behavior, including cracking, delamination or material flow issues. | Capable of detecting internal defects not visible on the surface of a part and is particularly sensitive to mechanical changes, such as stress fractures or voids. | Signals can be noisy and distinguishing between normal operation and actual defects can be challenging. | Primarily used in metal and composite 3D printing. May benefit from machine learning and other advanced analytics. |
Schlieren Imaging | Used to visualize changes in pressure and temperature, based on light refraction through areas of varying density. | Captures high-resolution visual data of thermal gradients and airflow, which can help identify problems with gas shielding or heat dispersion. | Requires highly controlled conditions, making it less viable for industrial settings. Primarily used for research rather than in-process monitoring. | Used in metal 3D printing research to study melt pool dynamics and gas flow behavior. |
Operando Synchrotron X-ray Imaging and Diffraction | Captures real-time, high-resolution images of internal structures during printing. X-ray diffraction tracks changes in the crystalline structure of materials. | Useful for understanding thermal and mechanical behavior of metals at a granular level. | Expensive and complex to implement. Primarily used for research and critical applications. | Most often used in aerospace and biomedical engineering research. |
How does real-time, in-process monitoring inform 3D printing simulation?
While there are many good reasons to deploy in-process monitoring in additive manufacturing, in the context of AM simulation, feeding the real-time data from in-process monitoring of 3D printing into predictive simulation models can help engineers anticipate potential defects, such as warping or residual stresses.
Moreover, combining feedback monitoring with simulation models enables engineers to predict where these issues are most likely to arise in the part. In an ideal scenario involving closed-loop controls, this information can be used to take corrective actions during the printing process, such as dynamic adjustments to laser power or print head speed.
As machine learning (ML) and artificial intelligence (AI) continue to evolve, there will be more possibilities for applying these powerful statistical tools to both process monitoring and simulation. Ultimately, AI and ML could be used to create a positive feedback loop between 3D printing process data and simulations of those processes in real time, reducing defects and material waste and increasing production speeds and part quality.