Summary:
- The need for in-process monitoring is crucial for industrial manufacturing.
- Control of the additive manufacturing process is needed to ensure quality assurance.
- A manufacturing plan with lean strategy is required for operational sustainability.
- 3D Systems (DDD, Financial) risks devolution if they cannot implement in-process monitoring technology and devise a strategy for industrial manufacturing.
- Without technological and operational ingenuity, investors are at great risk for financial loss from declines in demand for 3D Systems products and services.
There are two metrology companies that were considered in writing this article, Plasmo Industrietechnik GmbH and Sigma Labs (SGLB, Financial). Sigma Labs was chosen for illustration of in-process monitoring technology, quality assurance methodology and strategic plans for lean manufacturing operations. The decision for choosing Sigma Labs was based on the company having an approach that aligns to scientific research, analysis techniques that can be corroborated elsewhere, commercialized products in market, and comprehensive strategy for lean metal additive manufacturing operations.
3D Systems is at risk of competitive elimination without technology to ensure quality assurance for metal additive manufacturing. Metal AM is being used in more applications and a change is taking place with a shift from prototyping to large scale production. Nonetheless, there are barriers metal AM has to overcome for manufacturers to fully adopt the technology. There are variations in part quality which restricts metal AM parts from use in high-stress applications. The aerospace and automotive industries would like to add 3D printing technology to their manufacturing scope but part inconsistencies jeopardizes safety. The National Institute of Standards and Technology (NIST), the body in governance for safety, recognizes a critical need for in-process monitoring and control of the AM fusion process. Without a means to control the additive process and ensure quality assurance, 3DS risks technological obfuscation in an evolving landscape towards industrial manufacturing. If 3DS cannot evolve to meet the needs of the industry, investors are at great risk for losing their investments as the demand for 3DS products and services diminishes.
To ensure that each layer is properly constructed, in-process sensors are required to examine the transformation of metal powder to melt pool. The sensor must have the capability to collect the information with consistency. The information that is needed concerns the internal composition of melt pools and whether nor not the composition contain porosity (voids), unfused metal powder, and cracks. NIST researchers noted that cameras and photo-optical systems are insufficientas sensors as visual and thermographic sight can be easily obstructed by any density encountered. Cameras simply cannot penetrate the surface of the melt pool. NIST researchers found that thermal emissions may give unreliable temperature readings as the melt pool element contains homogeneous liquid, unconsolidated powder, and solidified surfaces.
MTU Aero Engines and Sigma Labs have successfully utilized acoustic energy to gather in-process information. MTU Aero Engines, Germany's premier aircraft manufacturer, presented their findings in the research article, "Online Monitoring of Additive Manufacturing Processes Using Ultrasound" which deduced that acoustic energy is a reliable source of information. Their findings were confirmed with the correlation of acoustic energy to destructive test results in tensile strengths and fatigue. They concluded that acoustic energy is reliable in gaining qualitative information for microstructure and mechanical properties. Sigma Labs has gone further in technological development and has formulated a statistical analysis system to prevent errors from occurring in the additive process.
There is a critical need to understand and control the fusion process. Process characterization is the mechanism of eradicating fusion errors to produce 3D printed parts with uniformly dense consistency. This procedure is commonly known as the"fingerprinting" of an approved part for production. Process characterization is divided into two processes that are interconnected to define the fingerprint, process parameter normalization and destructive testing. To reveal nominal parameters, each layer of the additive process is recorded to give a map of the whole part. The part is then subjected to tests in fatigue life, tensile strengths, and operation conditions (secondary characteristics) for physical limits and spliced for visual inspection. A failure in either tolerance test or visual examination is correlated to off-nominal signals from the map. The correlation procedure is performed by reducing the immense amount of raw in-process data and identifying key process physics that causes faults. The 3D printing machine parameters are then adjusted to correct the printing process. Multiple iterations in part build, testing, and machine process parameter are performed until errors are eradicated. The fingerprinting process of a fully dense part becomes more defined with each passing examination. Depicted in the diagram below is the fingerprinting process where multiple iterations in builds, destructive testing, and machine parameter adjustments are performed. Process characterization is mapping of the part as a whole and identifying nominal signals. Process qualification is achieved when one successful build contains only nominal signals.Process validation (repeatability / reliability)is accomplished when up to 30 consecutive runs have completed with only nominal signals detected. A quality certification report showing the results of in-process monitoring will also be generated as a warrant that a part is built free from defect.
The process of identifying nominal signals requires metallurgical expertise. Knowledge of material science and understanding of signatures emitted by the fusion reaction is needed for the classification of in-process signals. This is important as the fingerprinting process will not be performed properly and off-nominal signals can be incorrectly identified as passing signatures. As mentioned earlier, Sigma Labs has progressed further in technological development and has formulated an algorithmic software program from analyzing acoustic energy. Their statistical analyzing system, INSPECT, is able to prevent errors in the fusion process from occurring. Sigma Labs' engineers have perfected their skills in signal classification through their years of experience in metallurgical and welding research at Los Alamos National Laboratory. Needless to say, misinterpretation of the fingerprint signals will produce parts that will not pass destructive testing and become a waste of time, effort, material, and valuable resources. Therefore, knowledge of material science is crucial for proper process characterization and lean operation in manufacturing.
A manufacturing operation needs to lean in expenditure for the process to remain viable. Andrew Good, quality manager at Jesse Garant & Associates' Metrology Center, notes that overall manufacturing cost is doubled by default when parts undergo post-process quality assurance. To evaluate voids using conventional methods, parts must be cut open for manual evaluation, which destroys the part. Non-destructive inspective (NDI), accordingly, has significant cost benefit. NDI can be performed by in-process monitoring and employment of an intelligent automated system to bypass post-process inspection. By knowing the characteristics that constitutes quality, Sigma Labs' INSPECT can be trained to recognize conditions that cause errors in the fusion process and prevent them from occurring. As displayed in the diagram above, notice that destructive testing is no longer routinely performed during the production phase of manufacturing and post-process QA is only needed on "for cause" parts (those that are suspect). Destructive testing is performed on suspect parts so that the conditions that created faults (root cause of faults)can be identified. By knowing the root cause of faults, the conditions can be entered into INSPECT so that the system can better recognize and prevent errors from occurring. In this manner, errors are eradicated and production speed is accelerated to the fullest manufacturing potential.
A strategic plan is needed for metal AM to be a solution in industrial manufacturing. Production rates will dramatically increase with the utilization of part databases and enhancement scheme. A series of 3D printers can be tasked to refine part blueprints. With a database server, each machine in a build group will work in synergy in fine-tuning the parts blueprint by having each machine report suspect ("for cause") conditions to the database. The intelligent automated system will then be able to recognize the root causes of faults faster with more machines working on the same blueprint. Manufacturers can be reassured that parts are built to design intent as the database is continuously improved and fine-tuned. With a database server and machines to improve error detection, the entire factory floor will reach full capacity and the production stream freed of obstruction.
In-process quality assurance (IPQAâ„¢) will be the gateway to allow manufacturers to use AM for more applications. Manufacturers will find metal AM as a solution when the costs of post-process inspection and materials are calculated. Withtime and cost reduced over the entire parts life cycle, manufacturers will be able to include AM as a routine fabrication procedure for metal parts.
It is hard to have confidence in 3D Systems (3DS) with their inability to describe their quality control method. The only instance 3DS has discussed quality assurance was during the Stephens 2015 Spring Investment Conference. 3DS responded to a question asking how they will overcome the hurdles for adoption in higher levels of manufacturing and the only description the audience was offered is that they are "working on statistical process control algorithms that allow metrology level certifications for the printer".
Otherwise, there are no further explanations. 3DS has not offered any details for the method by which in-process data will be collected nor have they explained the tests being performed on parts to ensure their resiliency. Neither does there appear to be any sort of strategy for quality assurance in going forward at their resource website. There are questions of whether 3DS even knows about the research performed by the NIST which revealed the inferiority of cameras and photo-optical systems for in-process monitoring. 3DS states that they are a solutions provider, yet there is no proof in methodology or ingenuity in strategizing for industrial manufacturing.
How is anyone to believe when there is such a large and substantial amount in the lack of evidence? 3DS' metal printers are NOTdescribed in having in-process monitoring capability, and therefore, quality assurance cannot be ensured. How then will 3DS sustain itself with the focus in metal AM shifting from prototyping to large-scale production? How will part variation and inconsistencies be remedied? There is a huge void which 3DS needs to answer to ensure its survival within a competitive landscape. Without addressing these delinquencies, 3DS is at great risk of fading into obscurity like many other technologies that have gone outdated.
3DS has a responsibility to investors to describe its technology for in-process monitoring and plan details in becoming a solutions provider for industrial manufacturing. Shareholders and investors deserve an explanation as a return for vesting. Without clarification, stockholders are kept from knowing whether or not 3DS has the capability to adapt and meet the needs of the industry. If 3DS does not have the capacity to be a solution for industrial manufacturing, then investors are at great risk for loss as the demands for 3DS products and services will diminish. Investors have a right to know if a company cannot perform to task so that they can prepare and protect themselves from financial deterioration.
