2023-08-04 13:23:39
With the increasing industrialization of additive manufacturing, especially with regard to laser powderbed fusion and DED, the need for in-situ monitoring is growing. This is used to monitor the quality during the construction of the metal parts and thus to qualify and certify the process – and therefore the parts. However, technically there are still snags to in-situ monitoring, concludes ASTM International in a report on this subject.
ASTM International has published a detailed report on the technical maturity of in-situ monitoring in additive manufacturing. This Strategic Guide: Additive Manufacturing In-Situ Monitoring Technology Readiness forms the conclusion of a workshop that ASMT International has organized in recent years on this subject. Especially in a two-day workshop last year, more than 100 experts from more than 60 companies went into depth in this area.
In metal printing, the demand for in-process validation of workpieces via in-situ monitoring is growing
Monitoring is growing in importance
In the report, the authors write that the importance of in-situ monitoring in AM is growing. Literally, the document says: The role of ISM is becoming increasingly important as it has the potential to provide a holistic view of AM process health across a large number of interacting process variables. The ability to verify the state of control in an AM process and characterize any process anomalies is critical to the rapid, cost-effective qualification and certification of AM parts in a manufacturing environment.
Higher yield, no certification
The three monitoring options – monitoring the total system, monitoring the weld pool and monitoring the material deposition – make a valuable contribution to the AM process. They usually lead to a higher yield. However, the results are still insufficient to certify AM workpieces in the machine. Another conclusion is that closed loop systems are promising, but that the current certification frameworks do not allow closed loop control. The reason for this is that the data is not structured enough and some of the current ASTM standards do not allow parameters to be changed during the process. And that happens once more and once more with a closed loop system.
Machine learning holds promise, but lack of large datasets is still an obstacle
Challenge in data analysis
One of the other conclusions is that much still needs to be developed before a monitoring system that can detect errors during 3D printing can be used for product certification. The sensor technology is robust, but the data systems and data analysis are not. The biggest shortcoming is the lack of a clear framework for validation. Machine learning is seen as promising, both in the literature and by the participants in the workshops, but there are still a number of challenges that need to be solved. The biggest obstacle is that an enormous amount of data is needed to train the ML algorithms.
TRL levels
The report provides a clear picture of the TRL levels of the various components of a monitoring system, as well as a roadmap for further development. Although it is difficult to make a general statement, the workshop participants estimate the TRL level of in-situ monitoring to be between 4 and 5. Non-destructive inspection is still below that, between TRL levels 3 and 4. In-process control is a bit further, between TRL levels 5 and 6. A large number of manufacturers’ monitoring systems are described in the appendix to the report.
ASTM’s report can be read online for free and can also be downloaded. That can be done via this link.
1691212242
#Insitu #monitoring #suitable #product #certification
