11th European Conference on Non-Destructive Testing (ECNDT 2014), October 6-10, 2014, Prague, Czech Republic

Online Monitoring of Additive Manufacturing Processes Using Ultrasound

More Info at Open Access Database www.ndt.net/?id=16533

Hans RIEDER1, Alexander DILLHÖFER1, Martin SPIES1,

Joachim BAMBERG2, Thomas HESS2
1
Ultrasonic Imaging Group, Fraunhofer Institute for Industrial Mathematics ITWM; Kaiserslautern, Germany
Phone: +49 631 316004543, Fax: +49 631 316005543; e-mail: [email protected]
2
MTU Aero Engines; Munich, Germany; E-mail: [email protected]

Abstract
Additive manufacturing processes have become commercially available and are particularly interesting for the
production of free-formed parts. With Selective Laser Melting (SLM), components can be produced by localized
melting of successive layers of metal powder. In order to be able to describe and to understand the complex
dynamics of the SLM processes more accurately, online ultrasonic measurements have been performed for the
first time. In this contribution, we report on the integration of the measurement technique into the manufacturing
facility and on a variety of promising monitoring results.

Keywords: ultrasound, welding, aerospace, Inconel, monitoring

1. Introduction
Additive manufacturing processes have been investigated and some of them developed
commercially since the late 1980s. With Selective Laser Melting (SLM), components can be
produced by localized melting of successive layers of metal powder. In comparison with
today’s conventional techniques, this way of manufacturing allows for considerably more
freedom in designing and has a tremendous economic potential. Thus, it is particularly
interesting for the production of geometrically complex aero engine components.

By local melting with a laser beam such engine components have already been manufactured
from the heat-resistant nickel alloy Inconel 718. To ensure the quality, the starting powder
and the manufacturing parameters are supervised; also, the built-up components are inspected
using nondestructive as well as destructive techniques.

In order to be able to describe and to understand the complex dynamics of the SLM processes
more accurately, online ultrasonic measurements have been performed for the first time. In
this contribution, we report on the integration of the measurement technique into the
manufacturing facility. We present first results based on generalized B-scans which illustrate
the build-up of test specimens based on single layers of 40 µm thickness. The analysis of the
ultrasonic signals allows to infer information about the fusion of the single layers and about
the temporal formation of material defects. We also discuss the further potential of ultrasonic
measurements.

2. Additive Manufacturing and Quality Assurance Considerations

Powder bed fusion is the most frequently applied technique for layer manufacturing of
metallic objects. Figure 1 schematically illustrates the basic setup, where a laser beam melts
the metallic powder according to a CAD-file of the component to be build. At MTU Aero
Engines, Selective Laser Melting is used for components made of Inconel 718. Additional to a
specifically installed SLM technology machine, on which the results presented in this
contribution have been obtained, six SLM production machines (EOSINT M 270, [1]) are
employed to investigate and enhance the manufacturing of high-quality aero engine parts.
 Figure 1. Schematic representation of the powder bed fusion process.

The usual procedures for quality assurance (QA) include the control of the machine
parameters in view of process stability and repeatability as well as metal powder control with
respect to chemical composition, particle shapes, sizes and size distributions [2]. However,
due to the high quality standards pursued at MTU Aero Engines, additional measures for
quality assurance have been developed and implemented to monitor each layer surface during
the build-up process. Here, optical tomography (OT) is applied to monitor hot-spots, which
might indicate the formation of possible defects [3,4]. After production, the manufactured
parts are examined in view of deviations from the CAD-geometry using e.g. fringe projection
methods, while X-ray and fluorescent penetrant inspection (FPI) methods supply information
about porosity and the presence of surface defects. The set of MTU’s QA-measures is
complemented by materials testing where e.g. tensile tests and low-/high-cycle fatigue tests
are employed for material characterization [2]. Further promising monitoring techniques,
currently not pursued at MTU, are based on the observation of the welding process via
thermography [5,6] or laser-generated ultrasound [7].

To complement the already implemented OT-monitoring of the welding process, an ultrasonic

monitoring system has been integrated into the EOS technology machine to infer additional
information about the manufacturing process and the state of the built-up component, as
described in the following.

3. Ultrasonic Process Monitoring

3.1 Objectives and Chosen Setup

The aim of monitoring the layer build-up during the manufacturing process using ultrasound
is the observation and/or surveillance of parameters which are only accessible online. These
are the dynamics of the layer build-up, the interface coupling, the local material properties as
well as the formation of residual stresses, distortions and porosity. For the installation of
additional components in the build-chamber, restrictions have to be obeyed which are due to
the specific environmental conditions, mainly the inert gas atmosphere and the elevated
temperature of about 80° C.
In our approach, we have decided to fix the ultrasonic transducer at the lower side of the
build-platform. Figure 2 shows the setup with the fixed probe and the sealed installation of the
RF-cabling underneath the build-platform. We use an unfocused 10 MHz normal incidence
probe of 6.3 mm (¼ inch) diameter generating longitudinal waves, which has been glued to
the platform after grease coupling had been applied.

Figure 2. Ultrasonic probe arrangement underneath the build-platform.

3.2 Monitoring Inspection System

In the course of this study, a PC-based monitoring and inspection system especially designed
for the integration in such production systems has been developed and set-up. The main
characteristics of the four-channel ultrasound transmit-and-receive system are as follows:
bandwidth ranging from 400 kHz up to 30 MHz, data acquisition rate of 250 Mega-samples
per second, 14 bit resolution, process-induced triggering, time and event/welding controlled
monitoring with up to 1000 A-scans per second, integrated DSP-functionality and on-line/off-
line visualisation capabilities.

The system allows to record ultrasonic signals with a temporal resolution of up to 4 ns. Data
acquisition is performed layer by layer, as it is synchronized with the start of each layer-wise
welding process triggered by the EOS machine. During welding, the ultrasonic signals are
recorded within an adjustable time window and the RF-signals are simultaneously visualised.
The signals are stored and can be further processed, which is currently done off-line. The
established inspection system complies with the required specification of being able to
acquire data during build-jobs of up to 8 hours duration. Recording up to 1000 data A-scans
per second results in several Giga-bytes of data to be stored, further processed and evaluated.

4. Demo Build-Job and Ultrasonic Signals

The final objective is to monitor and control the component build-up by giving feedback to
the process if indications or deviations from the reference state or reference parameters are
detected. However, due to the sophisticated manufacturing process, the relationship between
the build-up parameters and the acquired ultrasonic signals has to be investigated gradually in
respective studies with reference to well-defined build-jobs. For illustration of the first steps
of these extensive investigations, we present preliminary results obtained on a cylinder of 20
mm in diameter and 10 mm in height supplied with a void as an internal defect. During the
build-up, illustrated in Figure 3, a sphere of 2 mm in diameter has been created, filled with
non-melted metallic powder. The characteristics of the build-job are: 40 µm layer thickness,
250 layers and approximately 90 minutes build-up time. The inspection procedure consists of
 Figure 3. Illustration of the welding process and the build-up of the cylindrical specimen.

three steps: (i) warm-up of the build-chamber and recording of a sequence of various
backwall echoes within the platform until the ultrasonic signals remain stable at about 80° C,
(ii) start of the SLM-process and monitoring of the welding/build-up phase and finally (iii)
off-line evaluation once the build-job has been completed. Figure 4 shows a detailed view of
the backwall echo with a short time change after the build-up of the 40 µm layer.

By plotting the A-scans acquired after the end of each layer-wise welding, the build-up
process becomes ‚observable‘. The representation of such a plot in Figure 5 shows some
interrupts in the recorded data due to gain adjustments. Also, a small change in interface echo
(IE) at a build-up height of 1.5 mm can be seen.

Figure 4. Detailed view of the backwall echo with a short time change after the build-up of another 40 µm layer.
Figure 5. Visualisation of the ultrasound signals with the bottom plate interface echo (IE) and the backwall echo
(BE); as time progresses the build-up height increases. Right: detailed view of the signals at a height of 1.5 mm.

With increasing build-up height, the recorded A-scans get more and more noisy. Therefore, a
specific A-Scan representation has been chosen (Figure 6, left): the A*-scan constitutes the
average of all A-scans recorded during the build-job. Due to the averaging, the incoherent
signals, i.e. the noise signals are reduced and only static objects become visible such as the
interface and defects, while the backwall echoes cancel out. However, as indicated in Fig. 6,
the void in the built-up cylinder is only poorly visible, due to a misalignment of the probe
relative to the void position. To validate the successful generation of the void as a model
defect, we have performed immersion testing of the finished cylinder. The acquired C-scan is
shown in Figure 6 (right), where we have used a 10 MHz focused transducer (focus 3 inches
in water). The defect is clearly imaged and no further indications are present in the specimen.

Figure 6. Left: the void is poorly visible because of misalignment of the probe relative to the void position;
right: C-scan image of the void by immersion technique.

5. Further Processing Approaches and Evaluations

5.1 Determination of Ultrasonic Velocity

We have evaluated the longitudinal wave velocity after each layer build-up, assuming an
average layer thickness of 40 µm. In order to obtain more stable results, the sound velocity
was calculated using an average of 40 ultrasonic signals recorded for each layer. The results
are presented in Figure 7 for increasing build-up height, showing a nonlinear, monotonous
increase of the ultrasound velocity, which reaches an asymptotic value once a certain height
has been built-up. It is assumed, that this effect is due to the heat influence induced by the
melting laser as the heat penetrates through several layers. This effect is larger in the early
stage of the build process due to the smaller volume which is present to absorb the laser
energy. Additional experiments and evaluations are planned to validate this hypothesis.

Figure 7: Ultrasonic velocity plotted as a function of build-up height.

5.2 Frequency Behaviour

Additionally we have evaluated the frequency behaviour on successive layer build-ups using
Fast-Fourier-Transformation. A representative result is displayed in Figure 8. On the left, the
colour-coded spectral distribution of the ultrasonic signals is shown before, during and after
welding. Here, a clear ‘distortion’ during welding can be seen, which might be used to
characterize the quality of the welding process. On the right of Fig. 8, the spectral distribution
of different build-up layers are compared, where the different colours belong to different
build-up heights. As can be seen, the spectrum changes with the higher frequency components
decreasing with increasing build-up height.

Figure 8: Representation of the frequency dependence during the welding process of one layer (left); spectral
distribution as a function of build-up height (right).

5.3 Build-Job with Variation of Laser Power

In order to gain preliminary information about the influence of varying power of the melting
laser beam on the quality of the manufactured part and the ultrasonic signals, a respective
specimen has been built-up and monitored. The thicknesses of the different sections built with
varying laser power are schematically shown in Figure 9, left. From the recorded A-scans
(Figure 9, right) the start and the end of the build-up phase with 50 % power can be identified.
However, the structure built with only 25 % power has no sufficient coupling to the preceding
layer, so ultrasound cannot be propagated anymore which is indicated by the backwall signal
which remains at the same position once the build-up height exceeds 12 mm (the respective
A-scan is not shown in Fig. 9).

Figure 9. Schematic representation of the variation of laser power (left); monitored signals as a function of the
number of welded layers.

6. Conclusion and Outlook

The results of our first investigations have shown that online monitoring of the complex AM
processes using ultrasound through the build-platform is feasible. Observation of the surface
dynamics during the build-up is possible by evaluating the backwall signals. The signals bear
indications for qualitative evaluation of residual stresses. Also, statements concerning
porosity seem to be possible by additionally evaluating the ultrasonic velocities.

In order to properly exploit the potential of ultrasonic monitoring, further research has to be
performed. It should focus on the stabilization of the inspection conditions, as e.g. the probe
coupling has proven to be critical for efficient data acquisition. The examination of various
defect types in dependence of their position in the component and the evaluation of the
ultrasonic signals in view of porosity are further issues to be pursued. As residual stresses are
critical with respect to distortions of the built-up components we plan to investigate the online
monitoring of respective reference specimens. The presented approach is currently limited to
parts with non-complex geometries, but is as such particularly interesting for e.g. build-
chamber control using test artifacts.

Acknowledgment

This project has received funding from the European Union's Seventh Framework
Programme for Research, Technological Development and Demonstration under Grant
Agreement No. 266271, MERLIN.
References

1. www.eos.info
2. J. Bamberg, K.H. Dusel, W. Satzger; Overview of Additive Manufacturing Activities at
MTU Aero Engines; in: Proceedings of the 41th Annual Review of Progress in QNDE,
Boise, IA, USA, July 20-25, 2014; to be published
3. G. Zenzinger, J. Bamberg, B. Henkel, Th. Hess, A. Ladewig, W. Satzger; Process
Monitoring of Additive Manufacturing by Using Optical Tomography; in: Proceedings
of the 41th Annual Review of Progress in QNDE, Boise, IA, USA, July 20-25, 2014; to
be published
4. V. Carl; Monitoring System for the Quality Assessment in Additive Manufacturing; in:
Proceedings of the 41th Annual Review of Progress in QNDE, Boise, IA, USA, July 20-
25, 2014; to be published
5. H. Krauss; Thermographic Process Monitoring in Powder Bed Based Additive
Manufacturing; in: Proceedings of the 41th Annual Review of Progress in QNDE,
Boise, IA, USA, July 20-25, 2014; to be published
6. S. Moylan, E. Whitenton, B. Lane, J. Slotwinski; Infrared Thermography for Laser-
Based Powder Bed Fusion Additive Manufacturing Processes; AIP Conf. Proc. 1581,
1191 (2014); http://dx.doi.org/10.1063/1.4864956
7. C. Hauser, J. Allen; Development of Aero Engine Component Manufacturing Using
Laser Additive Manufacturing; 3rd EASN Association International Workshop on
AeroStructures, October 9-11 2013, Milan, Italy.