2

Process Monitoring and Control

The first workshop session provided an overview of measurements and modeling for process monitoring and control in additive manufacturing (AM). Speakers described systems measured in situ and in real time as well as challenges of each resolution and process signature. Bianca Maria Colosimo (Politecnico di Milano), Jarred Heigel (National Institute of Standards and Technology), Marvin Siewert (University of Bremen), Kilian Wasmer (Empa), Ben Dutton (Manufacturing Technology Centre), and Amit Surana (United Technologies Research Center) each discussed research, challenges, and future directions relating to the following questions:

• How can systems be measured in real time?
• What AM measurements enable uncertainty quantification?
• How can the precision of a measurement be certified?
• How can measured data be employed to understand the full state of a system?
• What mathematical and statistical methods could be applied to AM? How can resources from other disciplines be integrated?
• What can be measured in situ and in line? What are the main challenges of coaxial and off-axis sensing in terms of accuracy, frequency, and spatial and temporal resolution?
• What is the correlation between process signature and product defects? How does the probability of detecting flaws connect with the qualification of an additively manufactured item?

• How can models and solutions be used to transfer knowledge from machine to machine and from laboratory to laboratory? How does this change depending on the material and geometry selected to make a part?
• What are the impacts of false positives and false negatives? What are the economic advantages of in-situ monitoring?
• What are the challenges of moving from monitoring to feedback control?

MEASUREMENTS AND MODELING FOR PROCESS MONITORING AND CONTROL

Bianca Maria Colosimo, Politecnico di Milano

Colosimo described Politecnico di Milano’s AddMe.Lab, a laboratory combining industrial machines and novel prototypes for AM processes such as selective laser melting, electron beam melting, directed energy deposition with powder and wire feedstocks, and binder jetting. She explained that in-situ monitoring can help reduce major industrial barriers for metal AM technologies, such as process instability, lack of repeatability, and defect rates (Mani et al., 2017; Tapia and Elwany, 2014; Everton et al., 2016; Spears and Gold, 2016; Grasso and Colosimo, 2017).

Defects in AM products originate in a variety of ways, including the equipment, process, design choices, and feedstock material. Colosimo shared several references for defect sources, as shown in Table 2.1. The process signature, which represents the manufacturing process through which data are collected from control systems and sensors, can give insights into approaches to control the quality of the final product. Ideally, in-situ monitoring could identify defects in real time and correct the process accordingly.

Colosimo provided examples of different levels of in-situ monitoring.

• Level 0: Using the existing signals (without additional sensors) to appropriately analyze all of the available information via statistical machine learning in order to predict defect onset from monitoring and fusing signal data (Grasso and Colosimo, 2017).
• Level 1: Monitoring the powder bed to assess uniformity of the powder coverage, the geometry, and possibly the temperature distribution of the melted layer. These assessments can be done using high-resolution images in the visible and infrared bands. At this level, it is possible to detect delamination defects as well as geometrical deviation between the actual and the nominal shape

TABLE 2.1 Defect Sources and Categories by Publication

SOURCE: M. Grasso and B.M. Colosimo, 2017, Process defects and in-situ monitoring methods in metal powder-bed fusion: A review, Measurement Science and Technology 28(4):1–25, 10.1088/1361-6501/aa5c4f. © IOP Publishing. Reproduced with permission. All rights reserved.

• printed at each layer (Tapia and Elwany, 2014; Mani et al., 2017; Grasso and Colosimo, 2017; Everton et al., 2016).

• Level 2: Monitoring the printed layer, using high-speed videos in the visible or infrared ranges. Hot and cold spots (i.e., areas that remain hot or cold for a long period of time and can cause geometrical or volumetric defects due to over-melting or under-melting) can be detected in the thermal signature. Infrared video cameras can aid in computing the spatial gradient and temporal gradient, which can be used to predict the final microstructure (Land et al., 2015; Krauss et al., 2014; Caltanissetta et al., 2018; Arnold et al., 2018; Trushnikov et al., 2016; Grasso and Colosimo, 2016; Colosimo and Grasso, 2018; Brumana et al., 2018).
• Level 3: Monitoring the AM track to assess the spatter signature, the plume, the shape of the track, and the cooling rate left by the beam behind it. The spatter signature can relate to the expected porosity, and an excessive plume can lead to job failure (Repossini et al., 2017; Ly et al., 2017).
• Level 4: Monitoring the melt-pool size, shape, and temperature. Since the laser directly impacts the melt pool, feedback control could be implemented to keep the melt-pool signature stable by varying the laser power and/or speed (Doubenskaia et al., 2012; Berumen et al., 2010; Kruth et al., 2007).

Levels 1 through 3 are considered “off-axis monitoring” because they need sensors that are not placed coaxially with the laser beam.

Colosimo emphasized that in-situ sensing can improve understanding of the AM process, allow for calibration of the AM process simulations, increase part quality (e.g., by detecting, preventing, or even compensating for defects), and support process qualification. Some pending issues, however, include correlating the process signature with product quality and modeling defects appropriately. She also outlined key sensing questions: How should the appropriate sensors and their spatial and temporal resolutions be chosen? How could in-situ sensing accuracy be certified? What methods and tools should be used for multisensor data fusion?

A goal is to move from “sensorized” machines that collect data to “intelligent” AM systems that use data to make decisions. This transition requires a combination of statistical methods to visualize effectiveness and efficiency. Colosimo stressed that multidisciplinary research is needed to enable new ideas in in-situ sensing, monitoring, and control.

MEASUREMENT SCIENCE FOR PROCESS MONITORING AND CONTROL

Jarred Heigel, National Institute of Standards and Technology

Heigel described the eight project areas of the Measurement Science for Additive Manufacturing program within the Engineering Laboratory at the National Institute of Standards and Technology (NIST):

1. Precursor material qualification,
2. AM machine and process qualification,
3. AM part qualification,
4. Metrology for multiphysics AM model validation,
5. Metrology for real-time monitoring of AM,
6. Machine and process control methods for AM,
7. Data-driven decision support for AM, and
8. Data integration and management for AM.

The primary objective of this program is to “develop and deploy measurement science that will enable rapid design-to-product transformation through advances in material characterization; in-process process sensing, monitoring, and model-based optimal control; performance qualification of materials, machines, processes, and parts; and end-to-end digital implementation and integration of AM processes and systems” (NIST, 2019). Heigel’s presentation focused on measurements and sensors used for real-time monitoring, challenges of real-time monitoring and control, and the path forward.

Real-time monitoring, Heigel stated, includes any sensor measurements that are continuously recorded during the AM process and used to ensure that the machine and process are performing as expected.¹ Common optical sensors include high-speed cameras, pyrometers, in-line cameras, and in-line photodetectors. These optical sensors can provide great insight into each layer but are limited to observing only the surfaces. Ultrasonic sensors can be used to detect subsurface defects by sending ultrasonic waves through the part, and acoustic sensors can detect melt-pool quality and part failure by monitoring the acoustic emissions from the melt pool and cracks.

Heigel explained that real-time monitoring enables both statistical process control and feedback control. Statistical process control involves comparing the data from a new build with historical data of other builds

___________________

¹ In the context of this presentation, layer-wise imaging or intermittent measurements are not considered real-time monitoring.

to determine whether the process is performing within an acceptable range. It also involves collecting data from process signatures and comparing them with control limits, which are calculated for the expected measurements of the process output. In contrast, feedback control relies on real-time monitoring and high-rate continuous measurement analysis that can then be used to modify the process.

The industry is striving for rapid processes and rapid certification, with the help of real-time monitoring and associated control. However, Heigel explained that the largest current barriers for industry are high capital costs, a lack of robust correlations, and difficulty interpreting what is being measured. Different monitoring approaches are being deployed to balance cost and speed constraints. Some AM machines are enabling layer-wise imaging and melt-pool monitoring. Coaxial photodetectors enable low-cost monitoring at sufficiently high speeds (compared to high-speed imaging) but lack fidelity to interpret processing quality. For directed energy deposition systems, coaxial melt-pool imaging is currently being used for real-time monitoring and control and feedback control because the process dynamics are comparatively slower.

Heigel elaborated on some challenges for real-time monitoring. The first challenge mentioned was measurement fidelity, which involves the trade-off between high spatial resolution and high temporal resolution. Thermal cameras can provide high spatial resolution but are temporally limited to 10³ Hz. Photodetectors can provide higher temporal resolution but cannot directly determine dynamic size variations in the melt pool. Another challenge for real-time monitoring is correlating the sensor data with the physics underlying the AM process. Better understanding the physics helps to inform decisions about what types of sensors to use, how to interpret the measurements, how to calibrate those measurements, what control algorithms to use, and how to prioritize research and development.

However, real-time monitoring and feedback control cannot correct flawed designs or processes. Heigel explained that the process must be improved to minimize variability, and the build strategies must be designed to optimize the process. The importance of modeling and validation efforts toward achieving this goal cannot be overstated. In addition, an improved understanding of the relationship between defects and real-time monitoring signals must be developed. This requires improvements to the post-process detection of defects and consideration of how the real-time data are processed and stored. Finally, metrology improvements, such as better calibration of the sensors, will play an important role in allowing data acquired across machines to be compared.

During the question and answer portion of this presentation, a participant asked Heigel what to do if a defect is detected during the monitoring

process. Should the part be discarded or is there a way to fix it? Heigel responded that, first, one has to identify the type of defect accurately and determine whether it is fixable. For example, it may be possible to re-scan a pore relatively close to the surface and to release the associated void, but a part with a crack may have to be discarded. Real-time monitoring can help determine whether a defective part should be discarded, corrected, or ignored.

Workshop co-chair Wing Kam Liu (Northwestern University) asked about the challenges of powder-bed technologies versus multiple-head machines. Heigel said that NIST focuses on powder-bed systems due to limited resources. While both are being used in industry, the focus tends to be on powder-bed technologies. Heigel emphasized that both methods have important considerations and would benefit from additional research. He noted that there are some challenges in powder-bed versus directed energy deposition. For example, ultrasonic measurements tend to be transferrable, but differences in process speed can create different sized melt pools and cause a different formation. Also, differences in the plumes and powder delivery result in different types of problems in powder-bed and multiple-head technologies. Lessons learned from measurement science about different optimal sensors, ultrasonics, and acoustics could be applied to both technologies.

SIMULATIONS: A CHANCE FOR KNOWLEDGE-BASED IMPROVEMENT OF ADDITIVE MANUFACTURING

Marvin Siewert, University of Bremen

Siewert began by explaining four competences in AM: (1) part design (e.g., topology optimization, residual stress and distortion, compensation of distortion), (2) pre-processing (e.g., part orientation, support structures, nesting of parts), (3) process (e.g., scan strategies, thermal management, microstructure properties), and (4) post treatment (e.g., hot isostatic pressing, milling, heat treatment). He provided several examples of how these competences work together in practice.

The first example was a simulation of residual stress and distortion. The classical thermomechanical approach calculates the temperature field using the initial condition and suitable boundary conditions. Next, thermal strains and force equilibrium are calculated at several time steps. Siewert noted that while this approach can be informative, it can also be difficult to calibrate and validate as well as time-consuming and cost-intensive to run. In contrast, the mechanical process equivalent method requires inserting the inherent strains of the union of multiple layers as loads into a mechanical calculation. This approach can be calibrated more

easily, even for large parts, and is implemented within software at the Integrated Status and Effectiveness Monitoring Program (ISEMP). The Additive Works GmbH is a spin-off company of ISEMP and provides (among other things) simulations based on the mechanical process equivalent method in its software Amphyon.² Some goals and applications for this simulation are fast computation of residual stress and distortion, fast estimation, identification of critical areas, adaptation of the design, and simulation-based adaptation of support structures. Ade Makinde (General Electric Global Research Center) wondered about the accuracy of the mechanical process equivalent method. Siewert noted that shrinkage from every new layer calibrates well and does not require data to be uploaded, which enables faster computations.

Siewert’s second example described mesoscopic and macroscopic simulations of a temperature field. On the mesoscopic scale, a Goldak heat source was used to analyze the melt-pool size and shape. This type of analysis can be used to calibrate heat sources by comparing micro-sections with simulated melt-pool shapes, to explore the influence of local geometry on the melt-pool size (e.g., overhanging regions with different angles), and to estimate cooling rates. On the macroscopic scale, energy input is realized by element activation at a certain temperature. This type of analysis can be used, for example, to understand the influence of different hatch orders (i.e., the order in which material is filled within the boundaries of AM parts) and to identify critical areas (e.g., hot spots). The models on both scales use fast finite-difference method/finite-element method calculations and are currently undergoing experimental validation.

Siewert explained that the vision of predicting and controlling all parameters in the entire AM process requires broad and deep thinking. Since the quality and reliability of the produced part is influenced by the whole process chain, every step needs to be understood as well as possible. Simulation methods and algorithms are needed to understand the process and measured data, to predict critical situations, to adapt to and overcome these situations, and to optimize the process. He emphasized that data to validate and calibrate methods as well as improved mechanisms to get adapted parameters into the process are critical to realize this vision.

In response to a question about employing measured data to understand the full state of a system, Siewert said that one could use measured data to calibrate and validate a simulation model. A reliable simulation can give a deeper understanding of the process within the system.

___________________

² To learn more about Amphyon, see https://web.altair.com/2017-introduction-to-additive-works, accessed October 26, 2018.

Another participant asked how the accuracy and precision of a measurement can be certified. Siewert explained that although applications vary, different measurement techniques and the calibrated simulations can be combined to improve the reliability of measurements.

DISCUSSION

Heigel, Wasmer, Siewert, Dutton, and Surana participated in a panel discussion moderated by Colosimo. Colosimo asked each panelist to comment on sensor issues. Wasmer said that the cost of sensors will always be a consideration as sensors may be too expensive for some low-cost applications. He also noted the value of a central resource for results from various types of sensors. Researchers have to understand as much as they can about both the process and the limits of the sensors in order to minimize error; often this can be done by measuring the piece directly, independent from the process parameters. A participant from industry asked which sensor data are most helpful. Surana responded that sensing could be used in many different ways, including off-line model validation and in-line detection of failure. However, the sensing process also depends on the scale being modeled and the techniques being used.

A participant asked the panelists to comment on the repeatability of sensor data, how consistent the sensors are across machines and manufacturers, and standards for these sensors. Heigel mentioned that NIST has been working toward understanding both sensor variability and machine/process variability and that a lack of standards or best practices for calibration is a barrier. NIST has been conducting an interlaboratory study for the past few years to investigate powder-bed fusion variability, and some irregularities have been observed. NIST is also researching the physics behind sensor measurement and developing calibration procedures. Dutton added that it is important to develop these tools to enable the sensors to scale up into other ranges. He added that a structural model of the part capabilities as well as the type(s) of defects and sizes that the part can handle would be helpful in establishing quality requirements for a part.

In response to a question about the effect of sensor distribution, Dutton mentioned that most current sensing methods are only looking at the top surface and can miss deeper defects. Other methods not yet applied, such as laser ultrasound, could cover both surfaces and material within about 2 mm of the surface. Improved sensing during the layer-by-layer AM build process may enable more thorough defect detection. Colosimo added that it is difficult to learn across machines because sensor integration tends to be manufacturer-specific. She also mentioned that multiple sensors could be used to increase the robustness of results and detect

when problems with a specific sensor occur. Another audience member asked whether NIST is considering any compression or filtering techniques to reduce storage demands for large data sets. Heigel responded that although NIST is not specifically looking into this, data storage is an important consideration. Colosimo encouraged widespread data sharing to enable faster progress. Heigel commented that although data sharing can help identify mistakes, open communication is essential; data can be easily misinterpreted, and challenges exist with repurposing data to new applications. Surana suggested that some data may be prioritized for storage along with summaries of where the supplemental information can be found.

In response to a question about the importance of process parameters, Wasmer emphasized the value of determining the exact moment something happens so that that moment can be explored using other techniques, such as machine learning. Surana agreed that this is an important opportunity. Another participant asked the panelists to comment on the challenges associated with part inspection. Colosimo responded that in-situ monitoring allows some visibility into the process during the build but is not as helpful when defects depend on the post-processing steps (e.g., thermal treatment and finishing).

Liu asked the panelists for their thoughts on short-term, intermediate, and long-term goals in AM. The panelists suggested the following areas for improvement:

• Short-term goals
  • Improving imaging capabilities (Colosimo);
  • Clarifying what to monitor and when (Dutton);
  • Setting expectations for assessing what can and cannot be done (Heigel); and
  • Establishing calibration procedures (Heigel).
• Intermediate-term goals
  • Facilitating real-time feedback control (Colosimo);
  • Improving the use of models and statistical analyses to determine the optimal level of feedback, taking into consideration the design and purpose of the AM part (Dutton);
  • Improving modeling capabilities to predict and design the process (Heigel); and
  • Advancing fast computations (Surana).
• Long-term goals
  • Improving the fundamental understanding of the processes, especially for varying shapes and materials (Colosimo);
  • Designing processes to be consistent across machines and sensors (Heigel); and

REFERENCES

Doubenskaia, M., M. Pavlov, S. Grigoriev, E. Tikhonova, and I. Smurov. 2012. Comprehensive optical monitoring of selective laser melting. Journal of Laser Micro/Nanoengineering 7(3):236–243.

Edwards, P., A. O’Conner, and M. Ramulu. 2013. Electron beam additive manufacturing of titanium components: Properties and performance. Journal of Manufacturing Science and Engineering 135(6):1–7.

Everton, S., M. Hirsch, P. Stavroulakis, R. Leach, and A. Clare. 2016. Review of in-situ process monitoring and in-situ metrology for metal additive manufacturing. Materials and Design 95:431–445.

Ferrar, B., L. Mullen, E. Jones, R. Stamp, and C.J. Sutcliffe. 2012. Gas flow effects on selective laser melting (SLM) manufacturing performance. Journal of Materials Processing Technology 212(2):355–364.

Foster, B.K., E.W. Reutzel, A.R. Nassar, C.J. Dickman, and B.T. Hall. 2015. A brief survey of sensing for metal-based powder bed fusion additive manufacturing. In Proceedings of SPIE—The International Society for Optical Engineering 9489: 94890B. Dimensional Optical Metrology and Inspection for Practical Applications IV, Baltimore, Md., April 20–21.

Fox, J.C., S.P. Moylan, and B.M. Lane. 2016. Effect of process parameters on the surface roughness of overhanging structures in laser powder-bed fusion additive manufacturing. Procedia CIRP 45:131–134.

Gong, H., K. Rafi, N.V. Karthik, T. Starr, and B. Stucker. 2013. Defect morphology in Ti-6Al4V parts fabricated by selective laser melting and electron beam melting. Pp. 440–453 in Proceedings of the Solid Freeform Fabrication Symposium. 24th Annual International Solid Freeform Fabrication Symposium, Austin, Tex.

Grasso, M., and B.M. Colosimo. 2016. An automated approach to enhance multi-scale signal monitoring of manufacturing processes. Journal of Manufacturing Science and Engineering 138(5):051003–051003-16.

Grasso, M., and B.M. Colosimo. 2017. Process defects and in-situ monitoring methods in metal powder bed fusion: A review. Measurement Science and Technology 28(4):1–25.

Huang, Q., N. Hu, X. Yang, R. Zhang, and Q. Feng. 2016. Microstructure and inclusion of Ti-Al-4V fabricated by selective laser melting. Frontiers of Materials Science 10(4):428−431.

Kempen, K., L. Thijs, B. Vrancken, S. Buls, J. Van Humbeeck, and J.-P. Kruth. 2013. Producing crack-free, high density M2 HSS parts by selective laser melting: Pre-heating the baseplate. Pp. 131–139 in 2013 Proceedings of the Solid Freeform Fabrication Symposium. 24th Annual International Solid Freeform Fabrication Symposium, Austin, Tex.

Khairallah, S.A., A.T. Anderson, A. Rubenchik, and W.E. King. 2016. Laser powder-bed fusion additive manufacturing: Physics of complex melt flow and formation mechanisms of pores, spatter, and denudation zones. Acta Materialia 108:36–45.

Kleszczynski, S., J. zur Jacobsmühlen, J.T. Sehrt, and G. Witt. 2012. Error detection in laser beam melting systems by high resolution imaging. Pp. 975–987 in 2012 Proceedings of the Solid Freeform Fabrication Symposium. 24th Annual International Solid Freeform Fabrication Symposium, Austin, Tex.

Krauss, H., T. Zeugner, and M.F. Zaeh. 2014. Layerwise monitoring of the selective laser melting process by thermography. Physics Procedia 56:64–71.

Kruth, J.P., L. Froyen, J. Van Vaerenbergh, P. Mercelis, M. Rombouts, and B. Lauwers. 2004. Selective laser melting on iron-based powder. Journal of Materials Processing Technology 149:616−622.

Kruth, J.-P., G. Levy, F. Klocke, and T.H.C. Childs. 2007. Consolidation phenomena in laser and powder-bed based layered manufacturing. CIRP Annals—Manufacturing Technology 56(2):730–759.

Land, W.S., B. Zhang, J. Ziegert, and A. Davies. 2015. In-situ metrology system for laser powder-bed fusion additive process. Procedia Manufacturing 1:393−403.

Li, R., J. Liu, Y. Shi, L. Wang, and W. Jiang. 2012. Balling behavior of stainless steel and nickel powder during selective laser melting process. The International Journal of Advanced Manufacturing Technology 59(9–12):1025–1035.

Liu, Y., Y. Yang, S. Mai, D. Wang, and C. Song. 2015. Investigation into spatter behavior during selective laser melting of AISI 316L stainless steel powder. Materials and Design 87:797–806.

Ly, S., A.M. Rubenchik, S.A. Khairallah, G. Guss, and M.J. Matthews. 2017. Metal vapor micro-jet controls material redistribution in laser powder bed fusion additive manufacturing. Nature Scientific Reports 7(1):4085.

Mani, M., B.M. Lane, M.A. Donmez, S.C. Feng, and S.P. Moylan. 2017. A review on measurement science needs for real-time control of additive manufacturing metal powder bed fusion processes. International Journal of Production Research 55(5):1400–1418.

Matthews, M.J., G. Guss, S.A. Khairallah, A.M. Rubenchik, P.J. Depond, and W.E. King. 2016. Denudation of metal powder layers in laser powder bed fusion processes. Acta Materialia 114:33–42.

Meier, H., and C. Haberland. 2008. Experimental studies on selective laser melting of metallic parts. Materialwissenschaft und Werkstofftechnik 39(9):665–670.

Mercelis, P., and J.-P. Kruth. 2006. Residual stresses in selective laser sintering and selective laser melting. Rapid Prototyping Journal 12(5):254–265.

Mousa, A.A. 2016. Experimental investigations of curling phenomenon in selective laser sintering process. Rapid Prototyping Journal 22(2):405−415.

Moylan, S., E. Whitenton, B. Lane, and J. Slotwinski. 2014. Infrared thermography for laser-based powder bed fusion additive manufacturing processes. AIP Conference Proceedings 1581(1):1191–1196.

NIST (National Institute of Standards and Technology). 2019. Measurement Science for Additive Manufacturing Program—2018. https://www.nist.gov/programs-projects/measurement-science-additive-manufacturing-program-2018.

Niu, H.J., and I.T.H. Chang. 1999. Instability of scan tracks of selective laser sintering of high speed steel powder. Scripta Materiala 41(11):1229–1234.

Parry, L., I. Ashcroft, and R.D. Wildman. 2016. Understanding the effect of laser scan strategy on residual stress in selective laser melting through thermo-mechanical simulation. Additive Manufacturing 12(A):1–15.

Prabhakar, P., W.J. Sames, R. Dehoff, and S.S. Babu. 2015. Computational modeling of residual stress formation during the electron beam melting process for Inconel 718. Additive Manufacturing 7:83–91.

Puebla, K., L.E. Murr, S.M. Gaytan, E. Martinez, F. Medina, and R.B. Wicker. 2012. Effect of melt scan rate on microstructure and macrostructure for electron beam melting of Ti-6Al-4V. Materials Sciences and Applications 3(5):259–264.

Read, N., W. Wang, K. Essa, and M. Attallah. 2015. Selective laser melting of AlSi10Mg alloy: Process optimisation and mechanical properties development. Materials and Design 65:417–424.

Repossini, G., V. Laguzza, M. Grasso, and B.M. Colosimo. 2017. On the use of spatter signature for in-situ monitoring of laser powder bed fusion. Additive Manufacturing 16:35–48.

Scharowsky, T., V. Juechter, R.F. Singer, and C. Körner. 2015. Influence of the scanning strategy on the microstructure and mechanical properties in selective electron beam melting of Ti–6Al–4V. Advanced Engineering Materials 17(11).

Seyda, V., N. Kaufmann, and C. Emmelmann. 2012. Investigation of aging processes of Ti-6Al-4 V powder material in laser melting. Physics Procedia 39:425–431.

Spears, T.G., and S.A. Gold. 2016. In-process sensing in selective laser melting (SLM) additive manufacturing. Integrating Materials and Manufacturing Innovation 5(1):1–25.

Strano, G., L. Hao, R.M. Everson, and K.E. Evans. 2013. A new approach to the design and optimisation of support structures in additive manufacturing. The International Journal of Advanced Manufacturing Technology 66(9–12):1247–1254.

Tammas-Williams, S., H. Zhao, F. Léonard, F. Derguti, I. Todd, and P.B. Prangnell. 2015. XCT analysis of the influence of melt strategies on defect population in Ti–6Al–4V components manufactured by selective electron beam melting. Materials Characterization 102:47–61.

Tapia, G., and A. Elwany. 2014. A review on process monitoring and control in metal-based additive manufacturing. Journal of Manufacturing Science and Engineering 136(6):060801.

Thijs, L., F. Verhaeghe, T. Craeghs, J. Humbeeck, and J.-P. Kruth. 2010. A study of the microstructural evolution during selective laser melting of Ti–6Al–4V. Acta Materialia 58(9):3303–3312.

Thomas, D. 2009. The development of design rules for selective laser melting [Thesis]. University of Wales, Cardiff, Wales, United Kingdom.

Tolochko, N.K., S.E. Mozzharov, I.A. Yadroitsev, T. Laoui, L. Froyen, V.I. Titov, and M.B. Ignatiev. 2004. Balling processes during selective laser treatment of powders. Rapid Prototyping Journal 10(2):78–87.

Trushnikov, D., L.N. Krotov, E.L. Krotova, and N.A. Musikhin. 2016. Reconstructing the melting channel shape in electron-beam welding from the parameters of penetrating X-ray radiation. Part I. Posing inverse geometrical problem. Russian Journal of Nondestructive Testing 52(10):576–582.

Van Belle, L., G. Vansteenkiste, and J.C. Boyer. 2013. Investigation of residual stresses induced during the selective laser melting process. Key Engineering Materials 554–557:1828–1834.

Van Elsen, M. 2007. Complexity of selective laser melting [Ph.D. thesis]. Katholieke Universiteit Leuven.

Weingarten, C., D. Buchbinder, N. Pirch, W. Meiners, K. Wissenbach, and R. Poprawe. 2015. Formation and reduction of hydrogen porosity during selective laser melting of AlSi10Mg. Journal of Materials Processing Technology 221:112–120.

Yasa, E., J. Deckers, T. Craeghs, M. Badrossamay, and J.-P. Kruth. 2009. Investigation on occurrence of elevated edges in selective laser melting. Pp. 180–192 in 2009 Proceedings of the Solid Freeform Fabrication Symposium. 20th Annual International Solid Freeform Fabrication Symposium, Austin, Tex., August 3–5.

Zaeh, M.F., and G. Branner. 2010. Investigations on residual stresses and deformations in selective laser melting. Production Engineering 4(1):35–45.

Zaeh, M.F., and M. Kanhert. 2009. The effect of scanning strategies on electron beam sintering. Production Engineering 3(3):217–224.

Zäh, M.F., and S. Lutzmann. 2010. Modelling and simulation of electron beam melting. Production Engineering 4(1):15–23.

Zeng, K. 2015. Optimization of support structures for selective laser melting [Ph.D. thesis]. University of Louisville, Louisville, Ky.

Zhou, X., X. Liu, D. Zhang, Z. Shen, and W. Liu. 2015. Balling phenomena in selective laser melted tungsten. Journal of Materials Processing Technology 222:33–42.