Industrial applications of computed tomography · tomography, sometimes abbreviated to XCT or iCT (i for industrial), isthemethodofusingX-rayradiationtotakeanumber of two dimensional

berues.ntlys ofingefly

n ofs. In thenly

etry. Asl ofpo-rials orird

theMs)s.

idlyveyous

CIRP Annals - Manufacturing Technology 63 (2014) 655–677

fter

gies,

ns a

e.g.,

s as

also

IRP.

Industrial applications of computed tomography

L. De Chiffre (1)a,*, S. Carmignato (2)b, J.-P. Kruth (1)c,R. Schmitt (2)d, A. Weckenmann (1)e

a Technical University of Denmark (DTU), Department of Mechanical Engineering, Denmarkb University of Padova, DTG, Department of Management and Engineering, Italyc Katholieke Universiteit Leuven (KU Leuven), Department of Mechanical Engineering, Division PMA, Belgiumd RWTH Aachen University, WZL, Chair of Metrology and Quality Management, Germanye University Erlangen-Nuremberg, Chair Quality Management and Manufacturing Metrology (QFM), Germany

1. Introduction

This paper gives an overview of the large and increasingnumber of industrial applications of X-ray Computed Tomography(CT) in the manufacturing industry as well as in other industries.The paper can be read as a natural continuation of the CIRP 2011keynote paper describing the use of CT for dimensional qualitycontrol purposes: i.e. for traceable measurement and toleranceverification of dimensions on mechanical components [69]. X-raycomputed tomography, sometimes abbreviated to XCT or iCT (i forindustrial), is the method of using X-ray radiation to take a numberof two dimensional (2D) images of an object in many positionsaround an axis of rotation. From these images, using software, athree dimensional (3D) model of the object’s external as well asinternal structure is reconstructed and can be analyzed. Likewise in[69], X-ray CT is called CT in this paper, the main focus still being ondimensional metrology applications.

As reported in [69], the first CT scanner was built for medicalimaging by Nobel Prize winner Hounsfield in 1969. Since 1980, CTbecame popular for material analysis and non-destructive testing(NDT) and detecting material defects. More recently, in 2005, CTtechnology entered the application field of dimensional metrology,as alternative to tactile or optical 3D coordinate measuringsystems. CT is similar to magnetic resonance imaging (MRI), but

useful tool for examining materials with high atomic numwhile MRI is extremely useful in examining soft biological tissOther 3D imaging techniques, using neutron sources, are currebeing developed for soft materials, and a certain overlap in termmethodology and applications can be expected. MRI imagtechniques are not considered and neutron source is only brimentioned in this paper.

Typical areas of use for CT in industry are in the detectioflaws such as voids and cracks, and particle analysis in materialmetrology, CT allows measurements of the external as well asinternal geometry of complex parts. So far, CT metrology is the otechnology able to measure as well the inner as the outer geomof a component without need to cut it through and destroy itsuch, it is the only technology for industrial quality controworkpieces having non-accessible internal features (e.g. comnents produced by additive manufacturing) or multi-matecomponents (e.g. two-component injection molded plastic partplastic parts with metallic inserts). CT can be considered as a threvolutionary development in coordinate metrology, followingintroduction of tactile 3D coordinate measuring machines (CMin the seventies and that of optical 3D scanners in the eightie

The number of industrial applications of CT is large and rapincreasing. After a brief market overview, the paper gives a surof state of the art and upcoming CT technologies, covering vari

A R T I C L E I N F O

Keywords:

Industrial application

Quality control

Dimensional metrology

X-ray Computed Tomography (CT)

A B S T R A C T

The number of industrial applications of Computed Tomography (CT) is large and rapidly increasing. A

a brief market overview, the paper gives a survey of state of the art and upcoming CT technolo

covering types of CT systems, scanning capabilities, and technological advances. The paper contai

survey of application examples from the manufacturing industry as well as from other industries,

electrical and electronic devices, inhomogeneous materials, and from the food industry. Challenge

well as major national and international coordinated activities in the field of industrial CT are

presented.

� 2014 C

Contents lists available at ScienceDirect

CIRP Annals - Manufacturing Technology

journal homepage: http: / /ees.elsevier.com/cirp/default .asp

icalplesies,

ials,

where MRI uses non-ionizing radio frequency radiation to detectthe magnetic resonance of hydrogen molecules, CT uses ionizingradiation and measures the absorption of X-rays. Consequently,the two techniques have different areas of application. CT is a

iled CTtion* Corresponding author.

http://dx.doi.org/10.1016/j.cirp.2014.05.011

0007-8506/� 2014 CIRP.

types of CT systems, scanning capabilities, and technologadvances. The paper contains a survey of application examfrom the manufacturing industry as well as from other industre.g. electrical and electronic devices, inhomogeneous materand from the food industry. The paper aims at giving a detaoverview of the numerous and different applications offor industrial purposes in the light of industrial applica

http://crossmark.crossref.org/dialog/?doi=10.1016/j.cirp.2014.05.011&domain=pdf

http://crossmark.crossref.org/dialog/?doi=10.1016/j.cirp.2014.05.011&domain=pdf


http://www.sciencedirect.com/science/journal/00078506


requof sintealso

Afor d

2. C

Amedresecomimpincrinspopencontthe

base2D Xtwoin elsurfsignCT

failemillovercoorand

withX-rapres(Stuof msevewermetinnoof iNowdealTodationnascseenNevmetaccu

Fig. 1Adap

L. De Chiffre et al. / CIRP Annals - Manufacturing Technology 63 (2014) 655–677656

irements. The paper also addresses limitations and problemstate of the art CT. Challenges as well as major national andrnational coordinated activities in the field of industrial CT are presented.n overview of the advantages and disadvantages of using CTimensional metrology is shown in Fig. 1.

T market overview

fter developing the first CT (called EMI scan) for human-ical purposes in the early seventies of the last century by thearch branch of the British EMI (a British multinational musicpany) the usage increased rapidly. The technology hasroved enormously and the number of implemented systemseased vastly. Application in industry like NDT for theection of technical objects for pores and inner defects wereed and developed in the eighties, but brought only smaller

ributions to the number of implemented systems. Moreoverdisadvantage of the conventional 2D X-ray systems being filmd could be overcome by highly advanced digital systems with-ray detection pannels and computer evaluation. The use in

-dimensional defects (flaws) testing and completeness checkectronic production, i.e. printed circuit boards (PCBs) as well asace mount technology (SMT), was developed and has led to aificant increase in application of CT systems. The application offor three-dimensional measurements was investigated butd due to lack of accuracy. In the early years after theennium, the 3D-accuracy and traceability problem wascome with a keen solution applying a conventional 3D-dinate measurement system for calibration and traceability,a new method for performing and evaluating measurements

good accuracy. The first coordinate measuring machine withy sensor facility, developed by Werth Messtechnik, wasented to the market during the international fair ‘‘Control’’ttgart, Germany) in 2005 [140]. Fast and accurate (in the rangeicrons) holistic measurements of the entire workpiece withral hundreds of tolerances (even inside of hollow workpieces)e now possible. That has incentivised several vendors ofrology systems to develop new CT systems and led to new and

industrial digital X-ray inspection systems was estimated by Frostand Sullivan [34] for 2009 to be 344.2 million US$, for 2011 US$309.5 million [87]. The forecast for 2014 is 450.6 million US$ [34]and for 2016 is US$ 509.4 million [87]. According to other sources[34] the industrial X-ray inspection systems market is predicted for2017 to be US$ 591.9 million. That is a Compound Annual GrowthRate (CAGR) from 2009 to 2017 of about 7%, what is significantlyhigher than the CAGR in general.

The global geographical distribution of installed systems in2009 is shown in Fig. 2b. Most installations are in North America,Europe and Japan. Future market development will be drivenmainly by India and China, not to forget Russia, Brazil and SouthAfrica (BRIC-states). The trends in the market for industrial CT canbe seen in two regards:

Applications in existing branches are expanding due to newrequirements in the technological leading areas like aerospace,

automotive and transport industry. The new drivers are productsafety and economics. In production the overall productivity andefficiency of manufacturing processes are to be improved. Safety ofcomponents must be increased as recent disasters caused byfailures in materials and parts are demonstrating. Economics ofproduction processes can be improved by replacing conventionalCMMs by CTs. New more comprehensive and more flexiblesolutions will support increasing the market with such applica-tions. In electronics and microelectronics we are faced with newchallenges in increasing economics and inspection comprehen-siveness. Innovative systems will be able to check the complete-ness of devices, of circuitries and – with suitable systems – also ofeach soldering point, even of new devices with hidden solderingpoints like ball grid arrays (BGAs). A considerably high number ofCT systems will be necessary to supply the demand of adaptablesolutions.

New markets can be identified. An emerging new large marketis the food industry. Integration of CT in packaging lines can checkthe content of – vacuum – sealed packages, cans or preserving jarsjust before delivery. Inclusions of contaminants like glass, metal,stone and other can be detected. If tightly focused they can even beeliminated. Moreover, in butcheries and meat processing factories,each piece of meat can be tested for the content of hidden fat andbones and the price can be calculated accordingly. Counting thenumber of such companies or maybe even butcher shopsdemonstrates the volume of a market referring to this. Anotheremerging market will be in security related utilizations. Against thebackground of the fear of terrorism attacks checking cargo, luggage

. Advantages and disadvantages of CT.

ted from [81].

Fig. 2. Market data. (a) Trend in annual revenue for industrial X-ray inspection

systems market [34,79,87]; (b) global distribution of totally installed CT systems.

Based on [121] and referred to Frost & Sullivan 2011-03-24.

vative investigations of research institutes, while the numbermplemented CTs for dimensional measurements blew up.adays the application of CT in industry covers quality controling primarily with dimensional metrology and flaw detection.y digital X-ray systems are common in the medical applica-

s, but industrial digital X-ray inspection systems are still aent market [103]. The development of CTs in medicine can be

as precursor for X-ray systems in industrial market.ertheless the needs and requirements in industry for CTrology are quite different to those in medical fields (mainlyracy and traceability). Fig. 2a shows that the market for

and even complete containers will get more importance. Detectingexplosives and other dangerous materials and devices will berequired not only for airports and ship terminals but also forrailway stations, public buildings like court houses, schools andattractive touristic sightseeing sites as well as factories andcompany sites. A wide upcoming market can be seen in the area

new materials like metal foam and CFRP (carbon fiber reinforcedplastic) and other composite materials, where CT will allow newtechnical solutions. Testing procedures for those materials aswell as for testing components have to follow the technologyleading to a significant market share of industrial CTs.

of ager)f a

andtem

the CTg tofor-

anthe

CTasis

e toowsime

base,

ally

L. De Chiffre et al. / CIRP Annals - Manufacturing Technology 63 (2014) 655–677 657

Altogether it can be forecasted that the CAGR of this technologyalso in future will be increasing faster than average CAGR ofeconomics.

3. CT scanning technologies

CT scanning technologies were reviewed in [69], with details ontechnical systems and components, including X-ray sources, X-raydetectors, kinematic systems, and other hardware and softwaretechnologies. In this section, an up-to-date overview is given, withfocus on: (i) types of CT systems, (ii) scanning capabilities and (iii)technological advances.

3.1. Types of CT systems

Many different types of CT systems are available today: fromsmall devices which can be attached inside a Scanning ElectronMicroscope (SEM) to large machines used for CT scanning of talland heavy parts. The most relevant categories of CT systems arelisted in the following.

3.1.1. Clinical CT

In clinical CT scanners, the X-ray unit (which carries source anddetector) is continuously rotated around the object or patient(which remains stationary or is slowly translated horizontallyalong the axis of the rotating unit) to obtain tomographic imagesrepresenting slices of the scanned body. In over four decades ofmedical applications, several generations of clinical CT scannershave been developed, ensuring a continuous increase of perfor-mances [51].

3.1.2. Material analysis and industrial CT

CT systems for material analysis and other industrial applica-tions such as non-destructive testing are fundamentally differentfrom clinical scanners. In these systems the object is rotated in theX-ray beam, while the X-ray source and the detector remainstationary. In addition, since the dose of radiation transpiercing theobject is normally not critical in industrial CT, greater radiationintensities can typically be used than those applied in clinical CT.Furthermore, because resolution and accuracy requirements aredifferent, scanning parameters usually differ significantly fromclinical CT [57,61]. Resolution and accuracy can also be adjusted bymoving the axis of rotation supporting the object either closer tothe source (higher image magnification and pixel resolution, butmore blurring) or closer to the detector (sharper images, but lessresolution) [69]. This is normally not possible with clinicalscanners where the rotation axis is centered between sourceand detector. Unlike clinical scanners, most CT systems formaterial analysis or industrial use apply cone beam geometryand flat panel detectors as it yields hundredsfold reduction inscanning time (multiple slices measured in one rotation) and goodimage quality [40]. However, systems with fan beam geometry andlinear detectors are also used, especially for reducing scatter effectswhen large thicknesses need to be penetrated using high voltagetubes.

3.1.3. Dimensional metrology CT

In conventional metrology (e.g. tactile coordinate measure-ment) the measurement must be planned taking into account

the flexibility and applicability [27]. Fig. 3 shows an example

multi-sensor CMM including CT sensor, tactile (touch-trigprobe and stable granite base. Fig. 4 shows an example ometrological CT system with liquid cooled micro-focus source

thermally controlled cabinet. The thermal stability of the sysand the cooling of both source and detector may greatly reducethermal effects on measurement accuracy [88,93,108,113].systems for dimensional metrology have to be tested accordinstandard procedures and guidelines in order to ensure conmance with metrological performance specifications. The Germguideline VDI/VDE 2617-13 is currently considered as

fundamental document for specification and verification ofsystems used for coordinate metrology and it forms the main bfor future development of ISO standards [129].

3.1.4. Robot operated CT

In order to automate loading and unloading of the workpiecbe scanned, CT systems can be integrated with robots. Fig. 5 shan example of robot operated CT. Other solutions for reducing tof workpiece handling are discussed in Section 3.3.1.

Fig. 3. Multi-sensor CMM including tactile and CT sensors and stable granite

shown without its X-ray protection housing [140].

Fig. 4. Metrological CT system with liquid cooled micro-focus source and therm

controlled cabinet [89].

ingplesthe

Fig. 5. CT system integrated into the manufacturing line by means of automated

robotized loading [14].

every feature. CT introduces a paradigm shift toward holisticmeasurements of workpieces. Since the dimensional measure-ments are performed on a virtual model, the data acquisition (CTscan) and evaluation can be done at different times and differentlocations. In dimensional metrology applications, special attentionis paid to accuracy and traceability of measurement results [69]. Inorder to enhance the accuracy of CT measurements, metrologicalCT systems are designed involving principles and technologies forCMMs. For example, metrological CT machines can be constructedusing high precision mechanical guideways and thermally stablestructures. CT can be utilized in multisensor CMMs, enlarging here

3.1.5. SEM CT

Specific CT devices can be attached in a standard ScannElectron Microscope (SEM), allowing 3D imaging of small samwith resolution down to 500 nm, without compromising

reguthe

elecdedispecbeamhold

3.1.6

Twithmaxas dpene

Cspecener3.1.7scanstitctionmetlam3.1.8

3.1.7

Eelecagaisyncoffe

Lelecof ostheiverymetdestgenehostSyncfacilmonand

wav

Fi


lar imaging capability of the SEM [10]. These devices utilizeX-ray radiation which is produced by a metal target hit by thetron beam in the SEM. The X-ray radiation is then acquired by acated camera mounted on the side window of the SEMimen chamber (Fig. 6). The object is rotated within the X-ray

using a rotation stage installed in place of standard specimeners.

. Large scale CT

he main restriction in CT scanning of large parts is connected the material attenuation coefficient, which limits theimum accumulated material thickness that can be penetrated,iscussed in Section 3.2.3. In order to increase the maximumtrable thickness, X-ray tubes with high voltage are employed.ommercial standard tubes are typically limited to 450 kV, butial tubes up to 800 kV are also available today [118]. For highergies, linear accelerator X-ray sources are also used (Section). Another important issue to be taken into account whenning large parts is the need of large detectors. In alternative,hing of multiple X-ray projections or multiple CT reconstruc-s can be performed, using various procedures [76]. Otherhods for measuring large parts are for example digitalinography and other techniques briefly discussed in Section.

. Linear accelerators and synchrotron CT

ven though the most common X-ray sources are based ontron guns [69], other possibilities exist to accelerate electronsnst a target to produce X-rays. Linear accelerators andhrotrons in particular, despite being very expensive sources,

r specific advantages with respect to common electron guns.inear accelerators (LINAC) greatly increase the velocity oftrons or other charged particles by subjecting them to a seriescillating electric potentials along a linear beamline. Thanks to

3.1.8. Digital laminography and other techniques

In CT the object have to be irradiated from all angular directions(3608). This is not always possible, for example in case of limitedaccess to the component or in case of large flat objects, due to highabsorption of the object in at least one direction. In order toovercome this problem, different techniques can be used, such aslaminography and tomosynthesis [60].

Digital laminography is a valid solution for scanning flat parts.In this technique, images of planes above and below the plane ofinterest are blurred out by reciprocal movement of the X-raysource and detector, to show a specific layer more clearly. Sourceand detector can be moved for example on circular orbits aroundthe same axis, so that only the points in one plane of the object, thefocal plane, project to the same position on the detector while thepoints in all other planes are blurred. The result of the inspection isthe image of the focal plane, which represent a section of theobject, with superimposed burred images of the other planes. Thistechnique can be applied for example to the inspection of

g. 6. CT device attached in a standard Scanning Electron Microscope [10].

Fig. 7. A 7000 kg steel casting (valve body) undergoing X-ray radiography using a

8.5 MeV linear accelerator [43].

Fig. 8. Aerial view of the Synchrotron ELETTRA in Trieste, Italy.

r high energy, linear accelerators can be used for penetrating thick and/or high absorbing parts, up to meters of concrete oral. Fig. 7, for instance, shows a linear accelerator used for non-ructive testing of large steel castings. X-rays can also berated by synchrotrons, which are special particle accelerators,ed in large facilities, as shown for example in Fig. 8.hrotron CT devices are available at nearly all synchrotronities. Their radiation has the following relevant properties:ochromatic, high coherence, high collimation, high brightnessintensity, low emittance, and wide tunability in energy/

elength [17,95].

multilayer printed circuit boards, welding seams on large partsand aerospace components [42,70]. Motionless laminography is avariant of digital laminography, where multiple detectors are usedto avoid the movement of source and detector. Fig. 9 shows anexample of motionless laminography applied to the inspection ofaircraft wings.

3.2. Scanning capabilities

The key capabilities of CT for industrial applications are brieflydiscussed in the following.

rcee toick-

torest(see

betion

forandthetedizederons

arlyureCT.oidsed

olors is5].

ntlytely, as

atic

trial

m

m

m

l-to-


3.2.1. Resolution

Many factors influence the spatial resolution of CT reconstruc-tions, including: focal spot size of the X-ray source, performance ofthe detector, magnification, number of projections, reconstructionalgorithms, and data post-processing. The focal spot size isparticularly important in determining the image quality. Systemswith focal spot size larger than 0.1 mm are typically referred to asconventional CT or macro CT. Microfocus systems (mCT) have aspot size down to one or few micrometers. Nanofocus systems(nanoCT) may reach sub-micrometric spot size, currently down to0.4 mm [48,60]. Synchrotron CT (sCT) can reach 0.2 mm resolution,and by applying Kirkpatrick–Baez optics (sCT + KB) can currentlygo down to 0.04 mm resolution [106]. Fig. 10 shows typical rangesof spatial resolution for the tomographic systems consideredabove.

3.2.2. Scanning speed

Contrary to coordinate measuring machines, in CT systems thescanning time is independent from the number of features to bemeasured on the object (see Fig. 11). On the other hand, the

these systems is limited by the measuring volume between souand detector, and depends also on: applied magnification duthe cone beam geometry, maximum penetrable material thnesses (see Section 3.2.4) and the capability of the CT systemapply advanced scanning procedures, such as region of intescanning (see Section 3.3.2), extended field of view scanning

Section 3.3.3) and helical scanning (see Section 3.3.4).

3.2.4. Maximum penetrable material thicknesses

The maximum accumulated material thickness that canpenetrated by X-rays depends on the material attenuacoefficient and the X-ray photon energy. Typical values

common materials are given in Table 1; complete tables

graphs are available in literature [22,90]. Before scanning,

object should be oriented as to reduce the maximum penetramaterial thickness. Optimal part orientation should also minimthe variation of penetration depth during object rotation, in orto avoid pixels saturation or extinguishment in X-ray projecti[81,137].

3.2.5. X-ray dose

While reducing the radiation dose per scan is particulrelevant in clinical and biological CT, the effects of X-rays exposon the scanned object are often negligible in industrial

However, in some cases the X-ray dose must be limited to avdegradation of materials, for example in the case of highly expopolymers [13], and to prevent specific effects, such as cmodifications in gem stones [102]. Safety of CT operatorensured through appropriate shielding of the X-ray system [8

3.2.6. Multi-material scanning capabilities

The capability of analyzing multi-material objects is frequedemanded in several industrial applications. CT is definicapable of providing interesting solutions to this demand

Fig. 9. Laminography system inspecting an aircraft wing [31].

Fig. 10. Typical spatial resolutions and object sizes (diameter) for macro CT, micro

CT, nano CT, synchrotron CT (sCT) and synchrotron CT with KB mirrors (sCT + KB).

Adapted from [60].

Fig. 11. Scanning time versus number of features to be measured: schem

comparison of (a) tactile CMM, (b) video CMM and (c) CT system.

Table 1Typical maximum penetrable material thicknesses for common indus

materials.

X-ray voltage 130 kV 150 kV 190 kV 225 kV 450 kV

Steel/ceramic 5 mm <8 mm <25 mm <40 mm <70 m

Aluminum <30 mm <50 mm <90 mm <150 mm <250 m

Plastic <90 mm <130 mm <200 mm <250 mm <450 m

Source: [26].

Note: The values in this table are maximum thicknesses producing low signa

noise ratios; with smaller thicknesses the transmitted intensity increases.

thents

by CTnts

ing:ble

torsrgyedson

scanning time depends on a number of parameters, including:exposure time, number of projections and performance of dataprocessing [69]. Typical scanning time for industrial CT systemswith cone beam currently ranges from few minutes to one or fewhours [26]. However, innovative techniques are available for fasterscanning, as discussed in Section 3.3.1.

3.2.3. Measuring range

As discussed in Section 3.1, very different types of CT systemsexist today: from desktop CT with reduced measuring volume tolarge scale CT. The dimensions of the object that can be scanned in

clearly demonstrated since the earliest clinical CT scans in

1970s. However, CT scanning of multi-material objects presealso significant difficulties, due to different X-ray attenuationdifferent materials and to specific image artifacts [132].manufacturers offer different solutions to facilitate measuremeof multi-material objects through multi-spectra scans, includmulti-material targets (e.g. different materials on an indexahead), dual-source CT and energy-sensitive sandwich detec[36]. Further details on advanced solutions for multi-enescanning are given in Section 3.3.5. Another issue that neparticular attention when performing CT measurements

mulcorrcons3.2.6asse

3.2.7

Ameato bmeaaccustudaccu[19,

Tisonconcdefitionmeareacare

achiinvein Seof CTis stbe pillusasseobta

3.2.8

Dincrcapawerspecdata

TmarexamevalCADa geilluscom

Fig. 1CT, c


ti-material parts is the identification of adequate thresholds forect surface determination. This issue as well as furtheriderations on multi-material scanning are discussed in Section, along with specific examples concerning multi-material

mblies.

. Accuracy

s for other coordinate measuring systems [21,141], thesurement uncertainty of CT depends on the specific objecte measured and the specific parameters chosen for thesurement process. The factors that influence the measurementracy are listed and discussed in [69]. Recently several newies have been conducted on uncertainty evaluation andracy enhancement of CT dimensional measurements

30,50,54,82,86,135].he results of the first international inter-laboratory compar-

of CT systems used for dimensional metrology, which wasluded in 2011 [15], show that sub-voxel accuracy is

nitely possible for CT dimensional measurements on calibra- artifacts. In particular, the comparison showed thatsurement errors in the order of 1/10 of the voxel size arehable for size measurements, while measurements of formmore affected by the influence of CT data noise [15,18]. Theevable accuracy using CT on real industrial parts isstigated in the CIA-CT international comparison describedction 6 of this paper [4]. A general description of the accuracy

in achieving traceable measurements is a complex issue thatill a matter of investigation [69,121] but a first indication canrovided here based on specific investigations [4,15]. Fig. 12trates the expanded uncertainty of CT measurementsssed from comparisons with reference measurementsined on CMMs.

. Software capabilities

ata processing plays an important role in CT technology, witheasing demands on high-performance computing. The generalbilities of software for CT data reconstruction and analysis

3.3. Technological advances

Industrial CT is rapidly evolving, with continuous improve-ments in both hardware and software components. Some of themain technological advances are presented in the following.

3.3.1. Fast CT scanning

Thanks to recent advances in CT components and computingpower, several manufacturers of industrial CT systems haverecently proposed fast CT scanning solutions, capable of inspectingindustrial parts in few seconds [9,11,92]. Due to their scanningspeed, these systems can be used for in-line inspection of products.For instance, Fig. 14 shows a fast in-line CT system for inspection ofcastings, with typical scanning and inspection speed of 5–10 mmcross-section per second, allowing a complete scan and analysis in10 s for small automotive parts (e.g. small pistons or chassis

2. Examples of measuring uncertainty on comparison items vs. dimension for

ompared to general CMM measuring capability.

Fig. 13. Example of graphical result of a commercial software solution for CT data

analysis of fibers orientation in composite materials. Different colors correspond to

different orientations of fibers.

Courtesy of Volume Graphics GmbH.

Table 2Comparison of measurements allowed by tactile CMM, optical CMM and CT.

Possible measurements are indicated by U.

Tactile Optical CT

Freeform geometry U U U

High aspect ratio U U

Soft material U U

Micrometric detail U See Fig. 10

Fast measurement U See Fig. 11

Internal detail U

Fig. 14. Example of fast in-line CT inspection system; workpieces are forwarded on a

conveyor belt through the rotating CT unit [11].

e described in [69]. In specific industrial applications of CT,ific software capabilities may be requested, especially for CT

analysis.o this end, several software solutions are now available on theket, allowing task-specific analysis of CT data, including for

ple: fiber composite material analysis (Fig. 13), wall thicknessuation, porosity and inclusion analysis, comparison to nominal

geometry, coordinate metrology, etc.: see examples in [69]. Asneral conclusion concerning scanning capabilities Table 2trates the possibility of performing measurements using CTpared to tactile and optical CMMs.

dedme

also

g’’)ongrcehas69]

begle

ted cant tovederal

theven

butent

rces36].tingows. CT

as

lizethatfer-fer-

allsesg aresredrastsoft


components) and in 80–90 s for larger engine components (e.g.cylinder heads) [11].

Another advanced solution in fast CT scanning is the possibilityof scanning moving parts, obtaining three-dimensional CTreconstructions that capture the movement over a period of time.The result is a dynamic volumetric dataset (i.e. a CT volume modelthat includes time and motion), called 4D CT scan that can bereproduced like a movie (see Fig. 15). The University of Leuvendeveloped a rotating in situ compression unit that was integratedin a CT scanner. It allows 3D visualization of the progressivecompression and collapse of e.g. scaffold or lattice structures(beam-like 3D structures): see Section 4.2.3 [64].

3.3.2. Region of interest scanning

When a specific detail of a larger object needs to bereconstructed at a higher resolution, then region of interest(ROI) CT scanning can be performed. ROI scanning comprises arange of different methods that allow measuring specific regions(volume portions) without reconstructing the whole object[60,73,76]. Fig. 16 shows an example of a particular ROI scanning,where the object is first scanned entirely at a lower magnification(coarse resolution) and then the region of interest is scanned at ahigher magnification (fine resolution), so that the higher magnifi-cation region is reconstructed taking into account also the voxelinformation from the coarse voxel volume [26].

the SFOV in the direction of the rotary axis, the extenreconstruction may be obtained by stitching several CT volureconstructions [26]. Extended field of view scanning is useful

for enhancing the obtained resolution.

3.3.4. Helical scanning

Helical scanning (also improperly called ‘‘spiral scannininvolves simultaneous rotation and translation of the object althe rotation axis, so that the relative movement of the X-ray souand detector to the object describes a helix. This procedure

several advantages such as elimination of Feldkamp artifacts [and increase of resolution along the axis of rotation. It can alsoused for scanning elongated objects that cannot fit into a sinexposure [76].

3.3.5. Multi-energy scanning and color CT

In multi-energy CT, the analyzed materials can be differentiaby using several advanced scanning methods, which basicallybe classified in two main approaches: (i) exposing the objecdifferent X-ray spectra and (ii) employing energy-resoldetectors. The first approach may be implemented in sevways: different spectra can be obtained for example by varyingscanning parameters (e.g. the source voltage, or the power, or ejust the current – which actually does not affect the spectrumcan help to achieve better signal-to-noise ratios [67]), or differtarget materials (e.g. using multi-target heads), or different sou(e.g. using dual-sources CT systems, with two different tubes) [For the second approach, energy-sensitive photon-coundetectors are needed. Their ability to resolve energies allenergy-selective imaging with a single X-ray exposure [136]systems equipped with such detectors are often referred to‘‘Spectral CT’’ or ‘‘Color CT’’ [1].

3.3.6. Phase contrast and dark field imaging

Phase contrast and dark field imaging can be used to visuasmall details and increase image contrast within structures

otherwise would appear uniform. Phase contrast imaging difentiates between structures under analysis by exploiting difences in the refractive index of materials and highlighting smdetails of differing refractive index [99]. Dark field imaging uscattering from sub-micron structures in the sample, offerinpowerful contrast mechanism to reveal subtle structural featuof an object [98]. The main advantage of these methods compato normal absorption-contrast X-ray imaging is enhanced contthat makes it possible to see smaller details and improved

tissue contrast [119], see Fig. 17.

Fig. 15. 4D CT reproducing the movement of a screw [91].

Fig. 16. Example of ROI scanning on electric razor blades: (a) CT acquisition of the

whole object, (b) CT acquisition of a portion with better resolution (higher

magnification), (c) CT result for the whole object, (d) CT result for the high

are in

caleichory

resolution region [26].

Fig. 17. Conventional X-ray radiograph of a spectacle case (top) compared with

dark-field X-ray radiograph [126].

3.3.3. Extended field of view scanning

Several methods are available for extending the CT scan field ofview [52,76]. They may be needed when a portion of the scannedobject is positioned outside the scan field-of-view (SFOV) andhence the line integrals corresponding to those regions are notprojected into the detector. If the object is larger than the SFOV,methods are available for extending the SFOV by combiningmultiple projection images taken at different positions of theobject, which is translated laterally [76]. If the object is longer than

Other 3D imaging techniques, using neutron sources,

currently being developed for soft materials, as illustratedFig. 18. Other 3D imaging techniques require the use of large-sfacilities such as neutron sources or synchrotron sources whprovides opportunities beyond those achievable from laborat

X-ramenespepen[58,inte[8,1matadvaindue.g.

Ima

4. In

4.1.

Nof shdemprod(e.g.withwhisuchmentestiisticmeacostComthe evoluinsprevenonwor

Fig. 1using

Fig. 1synch

deno

Adap


y sources. Neutron imaging provides unique, and compli-tary opportunities compared to X-ray imaging and arecially well suited for biological materials, or when very high

etration power is required (i.e. for imaging of engine blocks)59]. When it comes to synchrotron X-ray sources, the highnsity of such X-ray beams allow e.g. for ultrafast x-ray imaging09], Fig. 19, and multi-modal imaging of polycrystallineerials [72]. In order to facilitate industry access to suchnced imaging facilities, dedicated imaging centers linkingstry with major facilities are starting to emerge. Examples arethe Manchester X-ray Imaging Facility [127] and the DTUging Industry Portal [123].

dustrial application fields and requirements

Introduction

owadays, the manufacturing industry is facing the challengeorter product life cycles and growing product varieties. Thisands highly cost and time efficient product development anduction processes. Improvements in production technologies

injection molding) allow the manufacturing of complex parts freeform surfaces and a huge amount of different features,

ch have to be inspected. On the one hand advanced materials, as fiber reinforced plastics, enable new product develop-ts, but on the other hand require new measurement and

Fig. 20 presents an overview of different fields of application forthe CT technology in the industrial domain. CT measurements canbe evaluated voxel based or surface based. Voxel based, qualitativeevaluation can be classified in visualization and non-destructivetesting (NDT). Quantitative, surface based evaluation comprisesdigitization and dimensional metrology. The need for traceabilityincreases with the complexity of the task. Simple visualizations donot require absolute measures. For metrology applicationstraceability is crucial. Fig. 21 shows a flow chart of a typicaldimensional CT measurement process while Fig. 22 depicts use ofthe substitution method to achieve traceability of CT measure-ments [53,69].

8. Conventional X-ray radiograph of a camera (left) compared with radiograph

neutrons [96].

9. Historical development of fast X-ray tomography. Open symbols denote

rotron sources, while filled ones represent laboratory sources, red squares

te white beam and black circles monochromatic beam scanners.

ted from [77].

Fig. 20. Fields of application for CT.

Fig. 21. Flow chart of a typical dimensional CT measurement process [81].

Fig. 22. Establishment of traceability of CT measurements using ISO 15530-3 [53].

ng methods. Testing the conformity of the product character-s in every production stage with accurate and time efficientsurement technologies can contribute to reduce waste ands during the manufacturing. In this context industrial X-Rayputed Tomography (CT) offers a large variety of applications inntire development and production chain. CT delivers a holisticmetric model of the workpiece that can be used for versatileection tasks and dimensional measurements as well as forrse engineering applications. The CT has the capability to-destructively determine the inner and outer geometry ofkpieces.

on-ing,

avernalandring

queoneion)tc.)

lityandtrol

o-ere

ofstryeast

res)24. toro-

20].venplit

be

lysis


Visualization refers to the function testing of assembledworkpieces. It is a qualitative, visual conformity inspection ofthe interplay or the existence of components. It allows the analysisof function or even malfunction under real working conditions.Non-destructive testing is a qualitative testing method. Itencompasses the issues of defect analysis and material characteri-zation. Defect analysis focuses on the visual inspection ofmanufactured workpieces in terms of pores/blisters, voids,inclusions and cracks. Defect analysis is widely applied in castingand forming industry and injection molding industry to assure thequality of workpieces. Weak spots (e.g. regions of high poredensity) can be identified easily. Material characterization focuseson the determination of intended material properties. Typicalapplications are the analysis of fiber orientation, alignment anddensity in compound materials and pore size and pore density infoams (here pores are a defined material characteristic). In bothcases the results can be used for instance to improve themanufacturing processes concerning molding die geometry andmolding process parameters or to distinguish between good andbad parts (see Section 4.2.5, Fig. 42). Digitization describes thecapability to generate and evaluate virtual models of workpiecesfrom CT measurements. It can be structured in simulation andreverse engineering. Simulation is a widely used and powerfultechnique to predict or investigate workpiece properties andbehaviors. The object of interest needs to be modeled to performthe simulation. Instead of using a simplified model (e.g. CAD data),the CT provides detailed models of real workpieces. For examplefinite element simulations (FEM) can be used to investigate heattransport or mechanical stress. In reverse engineering applicationssurfaces of the virtual model are extracted and reconstructed forCAD applications. Reverse engineering enables to recover lostdrawings, capture prototypes or to build spare parts. The CT opensup new possibilities in the product development. Metrologycomprises any dimensional measurement on CT data. According tothe German guideline VDI/VDE 2630 Part 1.2 [130] the dimen-sional metrology aspect is differentiated in nominal/actualcomparison, tolerance analysis and wall thickness analysis.Nominal/actual comparison is an analysis and color codeddeviations visualization of the geometric deviations between aCAD model or a reference workpiece (nominal data) and ameasured workpiece (actual data). It allows a holistic, non-featurebased evaluation. Tolerance analysis comprises dimensional,shape, form and position tolerances, determination of compensat-ing elements, regular geometry and sculptured surfaces. Incontrast to conventional coordinate measurement, the measure-ments are performed on the virtual model. Wall thickness analysisdetermines a characteristic number in the volumetric model.Dimensional CT measurements can be used for quality controlthroughout the whole product development and manufacturingcycle.

4.2. Manufacturing industry

Since about 2005 the manufacturing industry started showinggreat interest in CT technology for quality control purposes:

� The rising complexity of components and products, with highfunction integration, has led to components with more complex,often internal, features that cannot be controlled on dimensions

yielding multi-material components also requires CT for ndestructive quality inspections (geometry, dimensions, fittporosity).� New production methods, in particular additive techniques, h

favored designing and producing components with intecavities that cannot be produced with traditional methods

cannot be measured on tactile or optical coordinate measumachines.� Not to the least, people start recognizing that CT has the uni

possibility to apply for a double quality check in one step and

measurement task: e.g. checking dimensional quality (precisand checking material quality (e.g. porosity, weld quality, ecan be done using the same CT measuring data.

The latter is illustrated in Fig. 23 [69] showing three quachecks done with the same CT data: (i) control of form

geometrical deviations, (ii) thickness verification and (iii) conof material density/porosity. This and other examples (e.g. thermforming, folding and welding of plastic honeycomb panel) walready given in [69].

4.2.1. Casting and forming industry

The aforementioned benefits of CT quality control areparticular importance to the foundry or metal forming induwhere checking for material flaws (pores, inclusions, etc.) is at las important as checking for dimensional quality.

Castings often depict internal cavities (produced with cothat also call for dimensional CT measurements: see Fig.

Forming of hollow components is also coming up as a wayreduce mass and weight in the transportation industry: e.g. hydformed hollow camshafts or hollow constructed crankshafts [1An example of an aluminum casting measured with CT was giin the CIRP keynote paper of 2011 [69]. That casting had been sinto several segments, provided with reference spheres, as to

Fig. 23. Reconstructed CT model (a), control of geometry (b), wall thickness ana

(c), and porosity inspection (d) of car inlet fan.

nallityn aray

eraln of

inethe

forr to

and tolerances in a non-destructive way with tactile or opticalCMMs. Control of complex assembled products also call for CT:one need to look inside the assembled product to detectmismatches (e.g. undesired gaps or collisions between individualcomponents of the assembly, possibly due to unwanteddeformations of those components enforced by the assembly).Measuring the components of an assembly separately (even ifwithin specs) does not guarantee proper functioning of theassembled system.� Techniques like plastic injection molding with metallic inserts,

two-component injection molding or other production methods

suited as a reference object for accuracy verification of dimensioCT measurements on castings [7]. Fig. 25 shows a CT quacontrol process for turbine blades. The blades were measured oNikon Metrology 450 kV scanner equipped with a linear X-detector. Individual dimensions are checked within sevsections and are used to provide a global Fail/Pass evaluatiothe blades.

CT measurements of a 3-cylinder head of a combustion engare shown in Fig. 26 (notice the steel bushings inserted in

aluminum cylinder head). The CT measurements were aimedreverse engineering and redesigning the cylinder head in orde

Fig. 2sectio

Fig.

chan

right

Cour

Fig. 2sourc


solve a problem of overheating and cracking of the head. Itinvolved following steps:

� Scanning of cylinder head on 10 MeV CT scanner and 2Dreconstruction (BRP-Rotax GmbH & Co).� 3D reconstruction and STL meshing (Materialize Mimics1

software).� Segmentation of different materials and internal structure

(Materialize Mimics1 software).� Re-meshing for FEM and CFD (Computational Fluid Dynamics)

(Materialize Mimics1 software).� CFD calculation (Star-CD1).� (Re)design using STL-based CAD systems (Materialize 3-matic1

CAD software).

4.2.2. Machining (subtractive manufacturing)

Machined parts call for CT for dimensional check of internalfeatures: for such parts, material quality control is less of an issue,while external features or cavities that are fully visible externallycan better be measured using traditional CMMs equipped withtactile or optical probes (e.g. laser stripe scanners in case multiplemeasuring points are desirable as in CT measurements). Internalfeatures of machined parts calling for CT inspection mostlyinvolves small/long internal channels (drilled, bored, tapped, lasermachined, etc.) or small re-entrant cavities (milled, EDMed, etc.)whose dimensions cannot be measured with traditional means:see Figs. 27 and 28. In quite some cases the purpose is not only tocheck for geometry, form and dimension of the cavities, but also tocheck for undesired burrs at intersecting holes: see Fig. 29. Fig. 30gives examples of micro parts that cannot be measured withclassical CMM.

5. Quality control of turbine blades: voxel model (left); dimensional check of

ns (right, green/red = within/out of tolerance).

Fig. 27. Aluminum component (120 mm � 120 mm � 220 mm) measured with CT.

Courtesy of KU Leuven.

4. Casted iron duct (bottom left) measured on a CT scanner with a 450 kV

e (top) and compared to the nominal CAD model (bottom right).

Fig. 28. CT CAD-comparison of hydraulic manifold of race car: scanned part (a),

comparison with nominal CAD outer (b) and inner (c) geometry.

Courtesy of Nikon Metrology/X-Tek, Belgium/UK.

26. Cylinder head showing aluminum block, steel bushings and cooling

nels/jacket (top-left); extracted STL/FEM mesh of water jacket shown top-

; bottom shows single slices for re-design.

tesy of Materialize NV, Belgium.

Material removal machining processes, being still the mostprecise manufacturing processes, are also the most demanding interms of accurate dimensional measurement. The accuracy anduncertainty of dimensional CT measurements is however counter-acted by typical phenomena as X-ray beam hardening, gray-valueedge thresholding, etc. This is illustrated in Fig. 31, where aprecision pin (Ø4 mm) inserted in a hollow stainless steel steppedcylinder (see dimensions Fig. 31a) was measured on a 225 kV CTdevice (Nikon Metrology MCT225). The hollow stepped cylinder

ingtion

nitd.

step

s.


was produced on a precision lathe (Mori Seiki NL2000Y/500) andcalibrated on a Mitutoyo FN904 CMM. The pin was a calibrated pin(tolerance �1 mm) [125]. Fig. 31b demonstrates that optimizing thebeam hardening (BH) correction reduced the non-systematicdimensional errors on the inner pin diameter (i.e. error neglectingthe systematic offset of about 3 mm) from 7 mm to 2 mm. Similarly,Fig. 31c demonstrates a reduction of the average outer cylinder edgeoffset by around 3 mm (typically from 5 mm to �2 mm) and of theinner hole diameter by 8 mm (from �13 mm to �5 mm): see arrows.Proper BH correction reduced the overall dimensional errors (bothsystematic and non-systematic) to within 5 mm.

4.2.3. Additive manufacturing

Additive manufacturing (AM) offers unique possibilities forproducing parts with internal cavities or lattice structures that areimpossible to produce with other manufacturing techniques.Hence, many AM parts feature such internal geometries that allowoptimizing the component’s, weight, shape and strength: Fig. 32.AM definitely calls for CT quality control as it is the only method toperform non-destructive dimensional measurement of innerfeatures and non-destructive density/porosity verification, whichis a critical issue in AM.

An example of an AM nozzle with conformal helix coolingchannel is given in Fig. 33.

Fig. 34 gives an example of a component made from to partsproduced by selective laser sintering of nylon powder (whitecomponents) and assembled by gluing (black glue). Blue ink waspoured inside the glued components. It demonstrates that the

compressed, and to reconstruct 3D CT images of the collapsscaffold, while recording compression force versus deformaand calculating local stresses. Fig. 36 depicts the compression uand some CT images and visual picture of a collapsing scaffol

Fig. 29. CT check for undesired burrs at intersecting holes: (a) object front view,

(b) object top view, (c) 3D CT model showing burrs.

Fig. 30. CT measured of micro milling cutter Ø0.4 mm (left) and of micro holes

Ø4.00 mm down to 0.25 mm (right).

Fig. 31. CT measurements of calibrated steel pin inserted in stainless steel

cylinder.

Fig. 32. Examples of AM parts with internal features and lattice structure

dedeseble

37),of adel,eenandults

parts are leaking (see blue area on left picture). CT scanningrevealed that this is due to interlayer porosity (see right picture).To check the performance of AM in producing thin latticestructures for lightweight aerospace components or porousmedical scaffolds for bone regeneration, several such latticestructures were manufactured by selective laser melting andtested with a CT scanner: Fig. 35 [128]. In order to analyze thestiffness, strength and collapsing mode (break, bucking, shear, etc.)of the scaffolds, a rotating compression unit has been designed andintegrated in a CT scanner (Philips HOMX 161) [64]. The rotationallows taking 3608 X-ray pictures of the scaffolds being gradually

4.2.4. Injection molding

CT is particularly suitable to investigate injection molpolymer parts, thanks to the good X-ray penetrability of thmaterials. Scan tasks typically encompass run-in and troushooting by part geometry comparison with CAD model (Fig.

identification of QC measurements, and defect analysis. Check

molded part geometry is relevant with respect to the CAD mobetween single cavities, among different molders, betwdifferent materials, and after exposure to heat treatment

wear [2]. Challenges are often encountered in terms of: res

depeand

variund

Oissu

Fig. 3mode

parti

Fig. 3pore


process run-in is carried out verifying critical features on partsfrom the different cavities using tactile CMMs, but companies areinvestigating the possibility of validating molds from CT measure-ments. Currently, 90% of the use of CT within the Prefilled DeviceQuality Control Department at Novo Nordisk A/S (DK) is related tomold approval, while 10% are special tasks, such as, e.g., scanningof a subassembly in order to check if click springs are assembledcorrectly [116]. A similar example is found at LEGO A/S (DK) wherethe verification department is responsible for measuring andtesting of LEGO elements from new production molds [29]. The testresults are used as a basis for approval and release of new elements

Fig. 33. Injection nozzle (�30 mm height).

4. SLS components glued together; the right picture is a section thru the CT

l, showing the nylon parts (gray color), interlayer porosity (dark spots), small

cle inclusions (white spots) and the glue (white area).

Fig. 36. Rotating compression unit integrated in CT scanner and used for

compression testing of titanium bone scaffolds. Left: scaffolds. Center: test unit.

Right: pictures of scaffold detail (top) and load unit (bottom).

Fig. 37. Comparison with CAD model of an injection molded cartridge holder [2].

Fig. 38. Injection molded part with a flash clearly visible under a microscope (left)

but invisible to CT (right) [116].

5. CT measurements of scaffolds with frequency plot of strut diameter and

size.

nding on fitting method, datum system interpretation, flashmold lines, rough surfaces, rounded corners, and large

ations in wall thickness. As an example, a flash easily observeder the microscope can be invisible to CT, Fig. 38.ften, mold approval, or mold validation, is a major production

e. Typically, validation of injection tools in connection with

– therefore of molds – and are an important contribution in thevalue chain of mold manufacturing. The LEGO elements aremeasured either by the use of Coordinate Measuring Machines ormanual measuring equipment. The idea is to measure as manyelements as possible using measuring machines, but today around75% of the features are still measured manually. However, manyelements feature complex geometries and require CT scanning:these elements feature high aspect ratios and hidden positions,where measurements by common means are not possible. Theverification department has recently invested in a CT scanner,which will further support quality control of the molds.

. (d)


The difference in measuring many parts can be appreciatedconsidering Fig. 39 which compares a set-up with several itemsready for CNC measurements on a tactile CMM with a multiple partholder for use in a CT scanner. An important application areaconcerns insert molding (where plastic is injected around an insert,usually metal), outsert molding: (similar to insert molding but herea metal sheet surrounds some plastic elements), and overmolding,(where another plastic is injected around a plastic product oraround a part of the plastic product). In the case of different similarmaterials (e.g. two plastics) or dissimilar materials (metal andplastic), CT is usually challenging, as addressed later in Section4.2.5.

CT can be used also for dimensional verification of microcomponents produced by micro injection molding [94]. Specificcalibration artifacts can be used in connection with CT scanning ofmicro components [17,78].

Other plastic components that can be measured effectively byCT are polymeric prosthetic joint components, where CT isparticularly useful for measuring the geometry of worn bearingsurfaces and quantifying wear volumes, with reduced uncertaintyin comparison to CMM measurements [16,19,20,117].

Injection molding is not limited to polymers but can also

Fig. 39. Set-up with several parts ready for CNC measurements on a tactile CMM

(top) and multiple part holder for use in a CT scanner (bottom) molds [29].

Fig. 40. CT scan of a metal injection molded part. (a–c) Section views

Reconstructed 3D volume [65].

Fig. 41. CT image of watch assembly.


aimthatsert 44.

Fig. 42. Example of plastic actuator (top left) and insert (top right) to fit together,

and example of fit analysis (bottom), where the insert is colored purple, the actuator

is colored gray and the region of overlap is colored green (overlap 0.04–0.07 mm).

involve metal [104]. Metal injection molding (MIM) is, e.g.,practiced at Grundfos (DK), where CT is used to detectmanufacturing defects, Fig. 40 [65].

4.2.5. Assemblies

X-ray CT is a unique tool for inspection of assemblies (Figs. 41–43). It allows visualizing the various components of the assemblyin the assembled state. This is essential as separate inspection ofthe individual components of an assembly (even if they all arewithin specification) does not guarantee proper functioning of theassembly.

An example of fit assembly analysis is given in Fig. 42. The

was to investigate a number of plastic actuators and inserts

should fit together and to detect why some pairs of actuator-indid not fit while others did. Another example is given in Fig.

It coAs imodthe

wellasse

4.2.5

ent mray

X-ra

� It

mmsmlofoDuit

chintecreaexwiex

Fig.

(insu

Cour

Fig. 4(c) re

invis

Cour

Fig. 4and i

X-ray


ncerns the assembly of two bolds made of the same material.t can be seen from the cross section thru the reconstructedel (Fig. 44d), the CT measurements does not allow to identify

horizontal fitting plane between the two bolts as they fit quite together. This might be a problem when measuring anmbled system.

.1. Multi-material assemblies. Assemblies consisting of differ-aterials are often encountered in measurement tasks using X-

CT. Multi-material assemblies are inherently problematic fory CT:

is often a dilemma when setting up the scanning parameters ifulti-material assemblies consist of both light plastic and denseetal parts. Low energy X-ray beams are essential for detectingall details and creating sufficient contrast especially for the

wer density material; High energy X-ray beams are necessaryr penetrating and revealing internal structures of denser parts.al energy CT (DECT) is often applied for solving this problem:utilizes different X-ray spectrum in order to optimally

aracterize multi-material components [49]. It is facilitated two different ways: dual detector and dual exposure/sourcechniques [105]. The dual detector technique is capable ofeating two datasets in one scan by using a multi-layer detector;ch layer is sensitive to a different energy band. The dualposure/source method consists of two separate CT scans, oneth high energy X-ray and one with low energy X-ray. Anample is shown in Fig. 45.

A connector with encapsulated metal parts is scanned usingdifferent X-ray energies. In the last step, X-ray projection imagesacquired with different X-ray energies are combined. Fig. 46demonstrates a working pipeline using multi-energy image stackfusion in CT dimensional metrology [67].� Surface determination is undoubtedly another main difficulty for

industrial CT when dealing with multi-material assemblies. Theglobal thresholding technique (defining material surface using asingle gray value: iso-surface) does not work, because using suchmethod for multi-material surface determination would create‘‘imaginary’’ layer of less attenuating materials around the edgeof more attenuating materials. As demonstrated in Fig. 47, an‘‘imaginary’’ layer of aluminum is created between steel and airat the bottom [46].

43. Example of a complex multi-material assembly: drug delivery device

lin pen) including components of different polymeric materials.

tesy of Novo Nordisk A/S [116].

4. Assembly of two threaded components: (a) real parts, (b) CT measurement,

constructed CT model, (d) cross section thru right part of CT model showing

ible contact planes.

tesy of Nikon Metrology/X-Tek, Belgium/UK.

Fig. 46. Example of the typical workflow of multi-energy image stack fusion [67].

Fig. 47. Voxel model containing three materials (steel, aluminum and air) and gray

value profile along the arrow. Due to the limitation of iso-surface technique, an

‘‘imaginary’’ layer of aluminum is created at the bottom interface between steel and

air [46].

5. Example of dual energy CT. Connector with encapsulated metal parts (left)

ts 2D X-ray projection images acquired using high (middle) and low (right)

energies [67].

In order to get more accurate material edges, local adaptivethresholding is necessary. As described in Fig. 48, this techniquestarts from a rough global contour assessment (defined by singlegray value) and searches for the steepest gray value change in itsneighborhood: the final material edge is then defined locally at thehighest gradient. This method accounts for local gray valuevariations caused by e.g. beam hardening artifacts and scatteringnoise.

However, the local adaptive thresholding method also has itslimitation, as shown in Fig. 49, where a set of steel and ZrO2 endgauges has been fitted together. Because the X-ray attenuations of

lies,hy

thems.hy

nts

bly

ROI


both materials are similar, their peaks in the gray value histogrammerge with each other (see Fig. 49b). Fig. 49c depicts the 3D modelgenerated by local adaptive (left) and global (right) thresholding.Neither of these two methods could make clear distinctionbetween steel and zirconia. Moreover, a lot of noise is observedaround the ZrO2 part when the local thresholding method isapplied. In this case, local adaptive thresholding is more easilyaffected by noise (low S/N ratio). Despite of the difficulties inselecting scan settings and in material surface determination,industrial CT has already proven its power in both qualitative (e.g.void detection, porosity or density observation) and quantitative(e.g. CAD comparison, dimensional measurements) assessmentsfor multi-material assemblies. An example of void detection anddefect analysis for a lamp assembly is given in Fig. 50. The aim wasto check for voids in the sealing layer and to find out the reason forair leakage. Another example on inspection a multi-material carheadlamp assembly using industrial CT is given in Fig. 51. The lampfilament is inspected using region of interest scanning (ROI, seeSection 3.3.2).

4.2.5.2. Example of mechanism analysis. Besides static assembCT is an excellent tool to analyze mechanisms and to check wmechanisms may fail (Fig. 52). Here again, analysis of

individual components of mechanisms may not reveal probleCT allows visualizing the mechanism in 3D and visualizes e.g. wit blocked: e.g. where an unwanted collision between componeoccurred.

Fig. 48. Working principle of the local thresholding method demonstrated on a

reconstructed slice [124].

Fig. 49. Problems in segmenting multi-material assemblies which consists of

materials with similar X-ray attenuation.

Fig. 51. Example of CT inspection of car headlamp assembly: complete assem

(top left), selected components (top right), lamp component (bottom left), and

CT of lamp filament (bottom right).

Courtesy of Werth Messtechnik.

Fig. 52. Analysis of watch mechanism.


lextedum

12 CTtter

ofnts.sonth a

Fig. 50. (a) Lamp assembly with glass welded feet plate, (b) 2D X-ray projection

image, (c) reconstructed CT slice, (d) 3D CT image.

4.2.5.3. Examples of complex assemblies. Fig. 53 shows a compelectro-mechanical assembly. It consists of a plastic injeccomponent with many metallic inserts (4 cylindrical aluminscrew holders and 27 copper connector pins) and in whichinduction coils are mounted. The part was measured on thescanners of KU Leuven, with and without the coils. In the lacase, appropriate thresholding allowed a clear segmentationgeometrical of the plastic, aluminum and copper componeComparison of individual dimensions and full model compari(CAD compared) was made between measurements done wi

CMMproj

Fmeamodspri

4.3.

4.3.1

Tindu(pactronboarsemprevvisuacceusedcomgenesembonvoidan e

Fig. 5and r

view

segm

Fig. 5Cour


, a 225 kV CT scanner, a 450 kV CT scanner and a fringeection system.ig. 54 shows a complex camera assembly reconstructed for CTsurements. External and internal views of the reconstructedel clearly show the various components: plastic casing, metal

ngs, electric wires, electronic components, etc.

Other industries

. Electrical and electronic devices

he wide range of applications for the CT in the electronicstry can be subdivided into the areas of componentskaging), connecting processes and complex structures. Elec-ic devices contain electronic components and printed circuitds (PCB). In electronic component manufacturing, for instance,

(2.0 mm � 1.2 mm) is shown in Fig. 55. The left image shows a 2DX-ray projection of the conductor. The different layers overlap. Theright image shows the 3D CT model [107,108]. The front cap ispartially removed and the coil is clearly visible.

Multilayer PCB’s consist of various layers with circuit paths. Amultilayer PCB fitted with SMD components is shown in Fig. 56.The carrier material was removed in the CT image to allow thedetection of short circuits that can be caused by etching defects,broken circuit paths in the different layers, defective plating ofthrough holes or measurement of layer offsets and annular ringwidth. Also the analysis of laser drilled bore holes can beperformed (wall thickness and bore quality). The electronicindustry uses a variety of connecting processes. The most commonprocess for PCB assembly is soldering. During the solder processdefects like missing solder filets, solder bridges or non-wettingdefects can occur. CT allows NDT for hidden defects, for examplethe void detection on mounted ball grid arrays (BGA). Fig. 57 showssolder balls of a BGA. Pores in the solder balls are colored in red.Some thin bond wires are also visible [84]. Voids and blisters on thecontact zone of components and PCB can be analyzed.

3. Electro-mechanical device. (a) Top of assembled component showing coils

ubber sealing; (b) CT with coils and sealing; (c) bottom of component; (d) CT

of top without coils; (e) X-ray image; (f) segmentation of alumina inserts; (g)

entation of copper pins.

4. 3D CT view of camera assembly (left) and internal detail (right).

tesy of Nikon Metrology/X-Tek, Belgium/UK.

Fig. 55. SMD inductor: 2D X-ray image (left), 3D CT model (right, front cap was

removed for better visibility of the coil).

Courtesy of phoenixjX-ray.

Fig. 56. Multilayer PCB. The carrier material in the CT image was removed.

Courtesy of QFM Erlangen [138].

Fig. 57. Ball grid array. Pores in the solder balls are marked in red [84].
i-conductors or inductors are encased in supporting cases toent physical damage and corrosion. Since CT enables toalize inner structures in assembled state which are notssible for conventional measurement technologies, it can be
for failure analysis and quality assurance. Due to the size ofponents, CT machines with nano or micro focus tubes arerally used to reach sufficient resolution. Typical tasks on

iconductors are detection of broken bonds or inspection of ballds. The detection of void size and void distribution in IC’s or

detection on solder surfaces of power transistors is done. Asxample a surface-mount device (SMD) inductor type 0805

Another widely used connecting process is wire crimping.Fig. 58 shows the 3D model of a crimp connection. In this case thetask is to determine the number of ingoing and outgoing strands.One can easily count 19 ingoing and 17 outgoing strands in thethree slices on the right side. Due to overlaps, this would not bevisible on 2D X-ray projections. The density in the crimp zone canalso be analyzed [108].

CT is the ideal tool for inspections of complex structures in theelectronic industry. Fig. 59 shows a slice of an epilator which can beequipped with different heads. It shows an optional head on the

thecay,ring

beod/s in

anith

oryperone. CTandwasre-

lingand

aterlice.

lly.


left and the standard head on the right side. The marked regionshows the lock mechanism. The standard head fits well. The CTimage reveals that the lock mechanism on the optional head doesnot catch due to dimensional deviations.

Fig. 60 shows a charger for an electric tooth brush. The markedregion shows the position of the fuse. On the left side the charger isassembled completely. On the right side the fuse is missing.

4.3.2. Inhomogeneous materials

Fiber reinforced plastics are widely used in lightweightconstruction. The mechanical properties depend significantly onthe fiber orientation and the fiber/matrix composition. The CT is aperfect tool to determine fiber orientation and distribution (seeFig. 13), detection of agglomerations of filler, voids and blisters in

recovery could be improved by 6–10% [71]. A CT scan can reveallocation of internal log defects like branch knots, rot or deinsect damage and cracks. Density, moisture distribution and

width are other quality characteristics of wood that candetermined with CT [45]. Also the identification of heartwosapwood boundary or anomaly detection at glue line interfaceengineered wood products is applications for the CT. Asexample Fig. 62 shows a slice view of a subalpine fir log wseveral defects like branch knots.

CT is also used on fiber based natural materials. This categcomprises wooden products like MDF, OSB, flake boards and paproducts e.g. laminates. To improve or develop new materials,

has to understand the characteristics and the micro structureenables to visualize complex structures of fibers, splints and strorientation. Correlation of density and pore distribution

analyzed by Standfest et al. [112]. Glatt et al. used CT measuments for the calculation of material properties and the modeof material geometry. Their work focuses on meshes, fabrics

felts [39].Fig. 63 shows a rolled up sample of wet knit ware. The w

film can be clearly distinguished from the knit wear in the sview on the left side. The right side shows the 3D view [115]

Fig. 58. Crimp connector. The right side shows three slices.

Courtesy of phoenixjX-ray.

Fig. 59. 2D Slice of an epilator with 2 different heads. The locking mechanism on the

left side does not catch.

Courtesy of Braun GmbH, Procter & Gamble.

Fig. 60. Tooth brush charger. The fuse on the right image is missing.

Courtesy of Braun GmbH, Procter & Gamble.

Fig. 61. Glass fiber composite. On the left side the resin was removed virtua

Courtesy of phoenixjX-ray [108].

Fig. 62. 2D Slice of a subalpine fir log [71].

Fig. 63. CT images of wet knit ware [115].

resin [62,80,97]. Fig. 61 shows a sample of a glass fiber compoundmaterial [108]. On the right side of the image the fiber mats and theresin filler are shown. Voids and blisters in the resin are clearlyvisible. On the left side of the image the resin is faded out. Alleleven layers can be inspected separately. Compared to othermeasurement methods (e.g. 2D X-ray images) defects can beexactly located in the workpiece. This is a big advantage of CTespecially for layer wise built material.

The CT technology offers a high potential for quality control andoptimization in the wood industry. The knowledge about defectscan be used to optimize the cutting process. The lumber value

Athis

wermicr[114

Sing aof cdistrprocorie

4.3.3

Tprod(DMDenprodlast

autocounmenin G107,devedistrthe dto adLeanresu

weig

whebontialscom

Fig


nother example of this type of product is shown in Fig. 64. Onsample of paper pore size, pore distribution and delaminatione inspected. CT even offers a new possibility to examine theo structure of paper and enables to design new materials].

teel fiber reinforced concrete is widely used in civil engineer-pplications. Steel fibers improve the energy absorption abilityoncrete. A substantial quality criterion is the uniformlyibution and orientation of fibers. CT and associated postessing enables to explore the fibers with respect to their mainntation, to classify and to statistically evaluate them [101].

. Food industry

he food industry is also characterized by large volumeuction. As an example, the Danish Meat Research InstituteRI) which is a division of the Danish Technological Institute inmark supports with cutting edge technologies a yearly nationaluction of over 20 million pigs and 15 million poultry. During the

years, the need for advanced measuring technology for flexible,mated production has increased in Denmark and other high costtries. At the same time, the necessary technological develop-t has become available with CT. Similarly to research institutesermany and other European countries [12,32,35,55,56,100,111,122], DMRI has been using CT for a number of yearsloping an on-line X-ray system for measuring meat-fat-ibution in pig carcasses or pieces [24,25,133]. By measuringistribution of meat, fat and bones in pork middles, it is possibleapt the cutting according to current market prices. The weight

Meat Content (LMC) assessment is performed with the use oflts from CT according to the formula:

ht ¼ V fat � bfat þ Vmeat � bmeat þ Vbone � bbone

re V indicates the volume and b the density of fat, meat, ande, respectively. Preliminary studies have demonstrated poten-

and realistic solutions, but have also clarified the extent andplexity of the challenges and risks. Fig. 65 shows a new CT

scanner truck for online use in production and makes a decisivebreak with the normal understanding of CT technology, its use,speed and limitations.

The CT scanner enhances the existing facilities allowingoptimizing and adjusting the cutting of each individual carcassaccording to current prices, quality demands and orders. Measure-ments result in the spatial distribution of meat, fat and bones anddeliver an optimal recipe for automatic cutting of pork middlesresulting in a return of investment for the slaughterhouse of lessthan 12 month. The CT scanners self-diagnostic and reportingcapability generate a detailed operational status including functioncontrol. The on-line CT-scanner is characterized by:

� High capacity i.e. ability to measure 700 pork middles/hour.� Cost effectiveness corresponding to 12 month return on

investment for the slaughterhouse.� A flexible platform for a broad range of applications.

A complete system coined ‘‘PigClassWeb’’, for handling thelarge amounts of CT-scans acquired by the DMRI in R&D projectshas been developed and implemented. Through advanced imageanalysis PigClassWeb enables DMRI and the slaughterhouses toperform virtual cuts in a reference pig. These cuts are automaticallypropagated to the whole population of pigs that are scanned, insuch a way that the virtual cuts are anatomically similar for eachcarcass, irrespectively of size, weight and proportions.

The ability to very accurately estimate the weight of arbitrarycuts enables to report the yields of the cuts on the population as awhole, as well as on each scanned carcass. The user can access theapplication through a simple web-browser and adjust the settingsof a specific cut through a view of the scanned reference carcass.For the simpler type of cuts the results are ready within seconds,when applying the cut on the whole population. PigClassWeb isscalable in the sense that future scans automatically are processedand included in the ‘‘Population of Virtual Pigs’’. New types of cutsare currently in the pipeline to be included in PigClassWeb keepingit up-to-date and making it a very flexible tool for DMRI in R&Dand for the slaughterhouses in the planning of their production.Fig. 66 shows the principle of an automated cutting line for pigcarcasses, Fig. 67 shows an automated middle cutting line, andFig. 68 illustrates the concept of virtual cutting of pig carcasses[25,33,47,133,134].

Besides the pork and poultry meat production, also the fishindustry and, more recently, other food industries are providing CTapplications. Here, new imaging techniques for on-line inspectionof food products as well as application of X-ray CT in high-resolution studies of food products for developing high quality

. 64. 3D view of a paper sample to inspect pores and delamination [114].

Fig. 66. Example of an automated cutting line of pig carcasses. From left to right: The

pig carcass is scanned, analysis of data with optimal cutting calculation, and robot

performs automated cutting of the pig carcass [41].

Fig. 65. DMRI’s Scannerborg [24].

food are of interest. One of the promising aspects is to use new X-ray imaging modalities for detection of hard-to-find foreign bodiesin food products. Refraction- and scattering-based contrastprinciples can be used to detect pieces of paper or insects in foodas complementary to conventional absorption contrast which isused already today for detection of glass, metal or stones. Anotherexample is determination of the distribution of ingredients in theproduct. These can be for instance nuts or berries in chocolate barsor holes in cheese, which are important quality parameters for theconsumers, Fig. 69. The idea is to apply X-ray imaging using severalmodalities to obtain a quantitative measure for the distribution.

imean

forandn a

weate

inees)

ust

anller

ller of

tion

blewerraicdgeAD-ore

74]. oft of10]d toT in

ingand

re-

the

Theent

m.sur-on-

nic

ay

ing

tru-tru-nts

and


These two kinds of inspections are performed using radiographyyielding 2D images but efforts concern also CT studies usingsynchrotron radiation facilities for inspecting food products at highresolution.

5. Challenges

- The accuracy must be significantly improved [15]. At the tbeing for normal standard applications one achieves

uncertainty in the range of a micrometer. But particularlymeasurements of features with dimensions in micro-

nanometer range an uncertainty in the range of less thatenth of a micron are necessary. That’s also a requirement ifcompare this technology with optical or tactile coordinmeasurement results.

- Larger workpieces as well as parts of higher density (e.g. machparts, gear boxes, crank cases, engine blocks, even ship enginmust be possible to be inspected [139].

- Also other objects like trucks, containers for ocean freight, mbe possible to be checked [139].

- More efficient X-Ray detectors must be developed withimproved signal-to-noise ratio in order to allow a smaexposure integration time and faster measurement.

- X-Ray tubes with more energy (higher current) and with smafocal spots will contribute to the measurement possibilitylarger objects and/or to more precise results at lower integratime [44].

- Improved reconstruction algorithms (maybe with selectaoptions specific to the measurement task) may need feprojections for accurate volume reconstruction: e.g. algebreconstruction, iterative reconstruction, use of prior knowleabout the object (e.g. from the product specification in the Cmodel or from the manufacturing process itself) may lead to maccurate results and/or to shorter reconstruction time [63,73,

- Task-specific measurement setups: selected combinationX-ray tube and detector, special trajectories for movemendetector and tube during data acquisition (robot CT, etc.) [37,1and special correction techniques for industrial CT [6] will leaimprovements in economy and application bandwidth of Cindustry.

- Objects consisting of different materials (material mix CT) bemeasurable by use of different spectra, special imaging

reconstruction techniques [23,67,75].- User assistance for reducing setup time and reducing measu

ment uncertainty [38,137].- Determination of measurement uncertainty [66,68,69,131].

In a nutshell: CT in industrial applications will have to meetmeasurement challenges of the future. Those are:

- Better resolution and accuracy e.g. by smaller focus spot.

resolution should be in the range of 10 nm and the measuremuncertainty for normal standard tasks in the range of 0.1 m

- Larger applicability range, particularly the possibility of meaing multi-component-parts, e.g. assembled electric and electric components.

- Possibility of measuring parts of material mix like electrocomponents and assembled mechanical components.

- Larger measurement ranges for larger objects like freight, railwwagons, containers, diesel engines for ships, etc.

- For in-line measurements in the production line the shieldmust be solved by labyrinths or other solutions.

- Better economy of CT measurements by less expensive insments and by better workman friendly operation of the insments. An unskilled worker should be able to make measuremeafter a one-day-course. The software must guide the operator

Fig. 67. Automated middle cutting in the meat industry.

Courtesy of DMRI.

Fig. 68. Screenshot of the PigClassWeb application for applying virtual cuts and

computing yields on a population of scanned pig carcasses [24].

Fig. 69. Radiograms of a ‘‘Toms’ turtle’’ chocolate with filling using absorption-

(upper left), refraction- (upper right) and scattering-based contrast. The contrast

between filling and chocolate is largest using the scattering contrast.

Courtesy of Niels Bohr Institute, Copenhagen University.

s inre-

ults

in a CT.

Future demands and requirements in general metrologyconcern mainly:

- Faster cycle time for a measurement according to the cycle timeof manufacturing.

- Better economy of the measurements process.- Improved operability by workers.- Enlargement of applicability range.

When considering the special technology of ‘‘CT dimensionalmeasurement’’ its individual characteristics must be regarded:

eliminate the risk of malfunction and maloperation.- Characterizing the measurement uncertainty for the result

graphical representation (as false color or as hedgehog repsentation) will lead to an improved understanding of the resand a faster interpretation of the results.

6. Major coordinated activities in the field of industrial CT

The vast interest by academics and industrialists is reflectednumber of major coordinated activities in the field of industrial

The

innoScie2009for

devecomresecomscaninte

AModunivrunnbud

T– Inton A13 PINTEinduencoer ex(CT-tiveing)modFurtof thprobmetmentech

AEuroatteinteheld

TsystwasinvoexpeLaboPadoreprThe

TinduEngi

Fig. 7carto


‘‘Center for Industrial Application of CT scanning – CIA CT’’ is anvation consortium co-financed by the Danish Ministry of

nce, Technology and Innovation that has run over the period–2013, focusing on the industrial application of CT scanning

advanced 3D measurement, quality assurance and productlopment: see Fig. 70. The consortium has acted as a nationalpetence center for industrial CT scanning through a number ofarch activities and initiatives, aiming to help the participatingpanies and Danish industry with the introduction of CTning as measuring technology along with research atrnational level. Project web site: www.cia-ct.mek.dtu.dk.nother framework research project called ‘‘New X-ray Imagingalities for Safe and High Quality Food (NEXIM)’’ involvingersities and several companies from the food industry ising in Denmark over the period 2012–2015 with a total

get of 4 Ms and involving 4 PhD and 6 postdoc projects [83].

he European Marie Curie Initial Training Network ‘‘INTERAQCTernational Network for the Training of Early stage Researchersdvanced Quality control by Computed Tomography’’ involveshD and 2 postdoc projects over the period 2013–2017. TheRAQCT project has been conceived as a pan-Europeanstrial-academic initiative that will provide the unique andmpassing training environment required, by bringing togeth-pertise from industry and academia in each of these domains

equipment, CT-software, NDT, dimensional metrology, addi- manufacturing, micro-manufacturing, composite manufactur-. The research will develop procedures for fast and accurate CTel acquisition, with special emphasis on multi-material parts.hermore, INTERAQCT targets quantification and improvemente reliability of CT measurement results, by determining theability of detection of material defects as well as by achieving

rological traceability. In addition, CT based quality improve-t loops will be targeted for key emerging manufacturingnologies.

number of conferences are devoted to the field of CT inpe. Four annual CIA-CT conferences with international

ndance have been held so far in Denmark. Since 2008, anrnational Conference on Industrial Computed Tomography is

in Wels, Austria www.3dct.at.he first international interlaboratory comparison of CT

participants from 8 countries. Instead of using reference artifactsfor calibration and verification of CT scanners, this comparison hasinvolved two real industrial parts commonly measured in industry,in terms of materials, dimensions and geometrical properties, asshown in Fig. 72: a plastic Lego brick and a metallic part of amedical device. Parallel circulation of 27 sets with the two itemstook place in spring 2013 and the final report was available inSeptember 2013 [4]. An example of results is shown in Fig. 73. Thecomparison has shown that CT measurements on the smallindustrial parts that have been used can be divided in two groups.Group A with average measurement uncertainties in the range 6–15 mm, and Group B with average uncertainties in the range 14–53 mm, and maximum values up to 158 mm, compared to averageuncertainties in the range 1.5–5.5 mm using CMMs. Fig. 74illustrates the expanded uncertainty of CT measurements fordifferent measurands on the two comparison items compared withreference measurements obtained on CMMs.

0. The vast field of industrial application of CT as illustrated by a CIA-CT

on: pig meat, mobile phone, hearing aid, insulin pen, Lego bricks [28].

Fig. 71. The four CT Audit items enclosed in their cylindrical sealed boxes: (a) CT

tetrahedron, (b) pan flute gauge, (c) PTB Calotte cube, (d) QFM cylinder [15].

Fig. 72. The two items used in the CIA-CT interlaboratory comparison of industrial

CT scanners: plastic Lego brick and metal part from a medical device [4].

Fig. 73. Example of comparison of measurement results using CT with reference

CMM data. Diameter of knob number 4 on Lego brick [4].

ems used for dimensional measurements, called ‘CT Audit’, carried out in the period from March 2010 to March 2011,lving 15 CT systems in Europe, America and Asia, operated byrt users. The project was organized and coordinated by theratory of Industrial and Geometrical Metrology, University ofva, Italy. Four calibrated samples were chosen for circulation,

esenting a variety of dimensions, geometries and materials.samples are shown in Fig. 71.he ‘‘CIA-CT comparison – Inter laboratory comparison onstrial Computed Tomography’’, organized by DTU Mechanicalneering, Denmark, and carried out in 2013, has involved 27

The CIA-CT audit – Inter laboratory comparison of medical CTscanners for industrial applications in the slaughterhouses wasorganized by DTU Mechanical Engineering in 2011–2012. Twosynthetic phantoms were used instead of real pig carcasses. Eachphantom consists of several polymer components made out ofPolymethyl methacrylate (PMMA), Polyethylene (PE) and Polyvi-nyl chloride (PVC). A comparison was made of volume measure-ment performance of different medical CT scanners in Europe. Thephantoms circulated among four participants and a total of sixmedical CT scanners. Volume estimations were compared to

http://www.cia-ct.mek.dtu.dk/

http://www.3dct.at/

ional

n of012,

n.be/

igh-ntrolstrial.peed18th

le by

ea-CIRP

of CTf Int.uary

009)rical

T forional

iCT

oor-

on ofc Hip

ticalnals

ctricientsenceData

on of336.e onersi-

viid CT-

Jeju

logy,.

d ed.

.

.inty

h In-

ssed

gen-wer-

(PhDark.

ext=

eat Re-

h to

itionence


volumes determined by water displacement performed before thecirculation and used as reference values [3].

7. Conclusions

The number of industrial applications of CT is large and rapidlyincreasing, as shown by the numerous and different applicationexamples given in this paper from the manufacturing industry aswell as from other industries, e.g. electrical and electronic devices,inhomogeneous materials, and from the food industry. The paperhas also identified limitations and problems of state of the art CT insolving some of the industrial challenges. A number of majornational and international coordinated activities in the field ofindustrial CT underline the strong industrial and academic interest.

Acknowledgements

The authors extend thanks to the colleagues of CIRP STC-P fortheir continuous support during preparation of the paper. Specialthanks are due to colleagues and researchers who have directlycontributed to the paper: Wim Dewulf, Harald Bosse, RichardLeach, Jose Yague-Fabra, Ye Tan, Frank Welkenhuyzen, SebastianPollmanns, Christopher Isenberg, Lars Bager Christensen, JanLasson Andreasen, Pavel Muller, Thomas Kjærgaard, Erik MejdalLauridsen, Carsten Gundlach, Jais Angel, and Alessandro Stolfi. Theauthors also want to thank a number of companies for theirsupport: Materialize, Nikon Metrology, Werth, and Zeiss. Thanksare also extended to the Danish Ministry of Science Technology andInnovation for funding CT activities through the CIA-CT project.

References

[1] Anderson NG, Butler AP (2010) Spectroscopic (multi-energy) CT Distin-guishes Iodine and Barium Contrast Material. MICE European Radiology20(9):2126–2134.

[2] Andreasen JL (2010) CT Scanning in Medical Device Development. ConferenceApplication of CT Scanning in IndustryTechnical University of Denmark.

[3] Angel J, Christensen LB, De Chiffre L, Cantatore A (2014) CIA-CT InterLaboratory Comparison of Medical CT Scanners for Industrial Applications inthe Slaughterhouses, Final Report, Department of Mechanical Engineering,Technical University of Denmark .

3D Imaging of Haines Jumps in Porous Media Flow. Proceedings of the NatAcademy of Sciences of the United States of America 110:3755–3759.

[9] Benninger R, Katuch P (2012) InLine CT – The Next Step in the EvolutioIndustrial CT. Proc. of Int. Conf. on Industrial Computed Tomography – iCT 2Wels, Austria, 19–21 September 2012.

[10] Bruker SkyScan (2013) Micro-CT Attachment for SEMhttp://www.skyscaproducts/SEM_microCT.htm. Accessed 31.05.13.

[11] Brunke O, Hansen F, Stuke I, Butz F, Jeltsch F (2012) A New Concept for Hspeed Atline and Inlinect for up to 100% Mass Production Process CoAllowing both, 3D Metrology and Failure Analysis. Proc. of Int. Conf. on InduComputed Tomography – iCT 2012, Wels, Austria, 19–21 September 2012

[12] Brunke O, Hansen F, Stuke I, Butz FF (2012) A New Concept for High-sAtline and Inlinect for up to 100% Mass Production Process Control.

World Conference on Nondestructive Testing, Durban, South Africa.[13] Cardoso dos Santos LG, Kawano Y (1994) Degradation of Polyacrylonitri

X-ray Radiation. Polymer Degradation and Stability 44–1:27–32.[14] Carl Zeiss IMT http://www.zeiss.de/imt. Accessed 31.05.13.[15] Carmignato S (2012) Accuracy of Industrial Computed Tomography M

surements: Experimental Results from an International Comparison.

Annals 61(1):491–494.[16] Carmignato S, Balcon M, Zanini F (2014) Investigation on the Accuracy

Measurements for Wear Testing of Prosthetic Joint Components. Proc. oConf. on Industrial Computed Tomography – iCT 2014, Wels, Austria, Febr2014.

[17] Carmignato S, Dreossi D, Mancini L, Marinello F, Tromba G, Savio E (2Testing of X-ray Microtomography Systems Using a Traceable GeometStandard. Measurement Science and Technology 20:084021.

[18] Carmignato S, Pierobon A, Rampazzo P, Parisatto M, Savio E (2012) CIndustrial Metrology – Accuracy and Structural Resolution of CT DimensMeasurements. Proc. of Int. Conf. on Industrial Computed Tomography –2012, Wels, Austria, 19–21 September 2012.

[19] Carmignato S, Savio E (2011) Traceable Volume Measurements Using Cdinate Measuring Systems. CIRP Annals 60(1):519–522.

[20] Carmignato S, Spinelli M, Affatato S, Savio E (2011) Uncertainty EvaluatiVolumetric Wear Assessment from Coordinate Measurements of CeramiJoint Prostheses? Wear 270(9–10):584–590.

[21] Carmignato S, Voltan A, Savio E (2010) Metrological Performance of OpCoordinate Measuring Machines under Industrial Conditions. CIRP An59(1):497–500.

[22] Chantler CT (2000) Detailed Tabulation of Atomic Form Factors, PhotoeleAbsorption and Scattering Cross Section, and Mass Attenuation Coefficin the Vicinity of Absorption Edges in the Soft X-Ray Addressing ConvergIssues of Earlier Work. Journal of Physical and Chemical Reference

29(4):597–1048.[23] Chen SY, Carroll JD (2000) 3-D Coronary Reconstruction and Optimizati

Coronary Interventions. IEEE Transactions on Medical Imaging 19(4):318–[24] Christensen LB (2013) CT in the Food Industry. 4th CIA-CT Conferenc

Applications of Computed Tomography in Industrial Production, 2013, Univty of Copenhagen.

[25] Christensen LB, Erbou SGH, Vester-Christensen M, Hansen MF, Darre M, HM, Olsen EV (2010) Optimized Workflow and Validation of Carcassscanning. 56th International Congress of Meat Science and Technology,Island, South Korea.

[26] Christoph R, Neumann HJ (2011) X-ray Tomography in Industrial MetroSuddeutscher Verlag onpact GmbH, Munich, ISBN: 978-3-86236-020-8

[27] Christoph R, Neumann H-J (2007) Multisensor Coordinate Metrology, 2nSV Corporate Media, Munich.

[28] CIA-CT Project newsletter nr 1, 2009. Technical University of Denmark[29] CIA-CT Project newsletter nr 8, 2013. Technical University of Denmark[30] Dewulf W, Kiekens K, Tan Y, Welkenhuyzen F, Kruth J-P (2013) Uncerta

Determination and Quantification for Dimensional Measurements witdustrial Computed Tomography. CIRP Annals 62(1):535–538.

[31] Digiray Corporation http://www.digiray.com/hangar/index.html. Acce31.05.13.

[32] Dobrowolski A, Romvari R, Allen P, Branscheid W, Horn P (2003) RontComputertomographie als mogliche Referenz fur die Schlachtkorperbetung beim Schwein. Mitteilungsblatt BAFF Kulmbach 161:267–273.

[33] Erbou SGH (2009) Modeling the Biological Diversity of Pig Carcasses,

thesis) DTU Informatics, Technical University of Denmark, Lyngby, Denm[34] Frost & Sullivan http://www.frost.com/srch/catalog-search.do?queryT

computed+tomography. Accessed 18.12.13.[35] Font I, Furnols M, Teran MF, Gispert M (2009) Estimation of Lean M

Content in Pig Carcasses Using X-ray Computed Tomography and PLSgression. Chemometrics and Intelligent Laboratory Systems 98(1):31–37.

[36] Fornaro J, Leschka S, et al (2011) Dual- and Multi-energy CT: ApproacFunctional Imaging. Insights Imaging 2:149–159.

[37] Fuchs T, Schon T, Hanke R (2010) A Translation-based Data AcquisScheme for Industrial Computed Tomography. 10th European Confer

Fig. 74. Expanded uncertainty (k = 2) of measurements on the two industrial items

shown in Fig. 72. D, R and L refer to different diameter, roundness and length

measurands. Group A refers to results in agreement with reference values; Group B

indicates those measurements for which a plausible uncertainty was recalculated [5].

Scane on

rtiesSep-

igital

. 23,

andc. of

[4] Angel J, De Chiffre L (2013) Final Report: CIA-CT Comparison Inter LaboratoryComparison on Industrial Computed Tomography, Department of MechanicalEngineering, Technical University of Denmark.

[5] Angel J, De Chiffre L (2014) Comparison on Computed Tomography UsingIndustrial Items. CIRP Annals – Manufacturing Technology 63(1).

[6] Baer M, Hammer M, Knaup M, Schmidt I, Christoph R, Kachelrieß M (2012)Scatter Correction Methods in Dimensional CT. Conference on IndustrialComputed Tomography (ICT), Wels, Austria, September 2012.

[7] Bartscher M, Hilpert U, Fiedler D (2008) Determination of the MeasurementUncertainty of Computed Tomography Measurements Using a Cylinder Headas an Example. Technisches Messen 75:178–186.

[8] Berg S, Ott H, Klapp SA, Schwing A, Neiteler R, Brussee N, Makurat A, Leu L,Enzmann F, Schwarz J-O, Kersten M, Irvine S, Stampanoni M (2013) Real-time

on Non-Destructive Testing ECNDT, Moscow, Russia 7–11 June 2010.[38] Giedl-Wagner R, Miller T, Sick B (2012) Determination of Optimal CT

Parameters Using Radial Basis Function Neural Networks. ConferencIndustrial Computed Tomography (ICT), Wels, Austria, September 2012.

[39] Glatt E, Becker J, Rief S, Wiegmann A (2010) Calculation of Material Propeand Modelling of Material Geometry. Proc. of iCT 2010, Wels, Austria,

tember 2010.[40] Goebbels J, Zscherpel Z, (Eds.) (2011), Proc. of Int. Symposium on D

Industrial Radiology and Computed Tomography, 978-3-940283-34-4.[41] Godske B (2008) Slagterirobot skal skære lige til benet. Ingeniøren no

Produktion, Automation, 6th June 2008. s. 8–9.[42] Gondrom S, Schropfer S (1999) Digital Computed Laminography

Tomosynthesis – Functional Principles and Industrial Applications. Pro

http://refhub.elsevier.com/S0007-8506(14)00193-0/sbref0005































http://www.skyscan.be/products/SEM_microCT.htm

http://www.skyscan.be/products/SEM_microCT.htm










http://www.zeiss.de/imt












































http://www.digiray.com/hangar/index.html

http://www.digiray.com/hangar/index.html






http://www.frost.com/srch/catalog-search.do%3FqueryText=computed+tomography

http://www.frost.com/srch/catalog-search.do%3FqueryText=computed+tomography





















[43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

[51]

[52]

[53]

[54]

[55]

[56]

[57]

[58]

[59]

[60]

[61]

[62]

[63]

[64]

[65]

[66]

[67]

[68]

[69]

[70]


International Symposium on Computerized Tomography for Industrial Applica-tions and Image Processing in Radiology. DGZfP Proceedings BB 67-CD, Berlin,March, 15–17, 1999.

Goodwin Steel Castings foundry http://www.goodwinsteelcastings.com/en/facilities/foundry. Accessed 31.05.13.

Gruhl T (2010) Technologie der Mikrofokus-Rontgenrohren: Leistungsgren-zen und erzielte Fortschritte. Fraunhofer IPA Workshop F 207: HochauflosendeRontgen-Computertomographie-Messtechnik fur mikro-mechatronische Sys-teme.

Gu H, Yu L, Wang Y (2010) CT Number in Wood Physical Properties PredictionBased on Computed Tomography Technology. Proc. of the 2010 IEEE Interna-tional Conference on Automation and Logistics, August 16–20 2010, Hong Kongand Macau.

Haitham Shammaa M, et al (2010) Segmentation of Multi-material CT Data ofMechanical Parts for Extracting Boundary Surfaces. Computer-Aided Design42(2):118–128.

Hansen MF (2009) The Virtual Knife, (PhD thesis) DTU Informatics, TechnicalUniversity of Denmark, Lyngby, Denmark.

Heinzl C (2009) Analysis and Visualization of Industrial CT Data, (PhD thesis)Institute of Computer Graphics and Algorithms, Vienna University of Tech-nology, Vienna.

Heinzl C, et al (2008) Statistical Analysis of Multi-material Components usingDual Energy CT. Proc. Vision, Modeling, and Visualization, October 8–10, 2008.

Hiller J, Maisl M, Reindl LM (2012) Physical Characterization and Perfor-mance Evaluation of an X-ray Micro-computed Tomography System forDimensional Metrology Applications. Measurement Science and Technology23(8). art. no. 085404.

Hsieh J (2009) Computed Tomography: Principles, Design, Artifacts, and RecentAdvances, 2nd ed. Hoboken, John Wiley & Sons.

Hsieh J, Chao E, Thibault J, Grekowicz B, Horst A, McOlash S, Myers TJ (2004) ANovel Reconstruction Algorithm to Extend the CT Scan Field-of-view. MedicalPhysics 31(9):2385–2391.

ISO 15530-3 (2011) Geometrical Product Specifications (GPS) – CoordinateMeasuring Machines (CMM): Technique for Determining the Uncertainty ofMeasurement – Part 3: Use of Calibrated Workpieces or Measurement Standards.

Jimenez R, Ontiveros S, Carmignato S, Yague-Fabra JA (2013) FundamentalCorrection Strategies for Accuracy Improvement of Dimensional Measure-ments Obtained from a Conventional Micro-CT Cone Beam Machine. CIRPJournal of Manufacturing Science and Technology 6:143–148.

Judas M (2010) CT Calibration for Pig Carcass Classification in Germany. Proc.of the Farm Animal Imaging Congress, Rennes, France.

Judas M, Horeth R, Branscheid W (2010) Rontgen-CT als Referenz fur dieHandelsklasseneinstufung von Schweinehalften, 45. Kulmbacher Woche 4.-5.Mai 2010, Kurzfassungen der Fachvortrage Veranstaltet vom Max Rubner-Insti-tut Bundesforschungsinstitut fur Ernahrung und Lebensmittel, Standort Kulm-bach.

Kalender WA (2011) Computed Tomography: Fundamentals, System Technolo-gy, Image Quality, Applications, John Wiley & Sons.

Kardjilov N, Hilger A, Manke I, Strobl M, Treimer W, Banhart J (2005) IndustrialApplications at the New Cold Neutron Radiography and TomographyFacility of the HMI. Nuclear Instruments and Methods in Physics ResearchSection A: Accelerators Spectrometers Detectors and Associated Equipment542(1–3):16–21.

Kardjilov N, Manke I, Hilger A, Strobl M, Banhart J (2011) Neutron Imaging inMaterials Science. Materials Today 14(6):248–256.

Kastner J (2011) X-ray Computed Tomography for the Development of Materialsand Components, (Habilitation thesis) Vienna University of Technology,Vienna.

Kastner J, (Ed.) (2012), Proc. of Int. Conf. on Industrial Computed Tomography –iCT 2012, ISBN: 978-8440-1281-1.

Kastner J, Plank B, Salaberger D, Sekelja J (2010) Defect and Porosity Deter-mination of Fibre Reinforced Polymers by X-ray Computed Tomography. 2ndInternational Symposium on NDT in Aerospace 2010 – We.1.A.2.

Katsevich A (2003) A General Scheme for Constructing Inversion Algorithmsfor Cone Beam CT. International Journal of Mathematics and MathematicalSciences 21:1305–1321.

Kerckhofs G, et al (2010) The Combined Use of Micro-CT Imaging, In-situLoading and Non-rigid Image Registration for 3D Experimental Local StrainMapping on Porous Bone Tissue Engineering Scaffolds under CompressiveLoading. Proc. 10th Eur. Conf. on Non-Destructive Testing (ECNDT), Moscow, 7–11 June 2010.

Kjærgaard T (2012) CT Scanning in Pump Manufacture. Conference IndustrialApplications of CT Scanning – Possibilities & Challenges in the ManufacturingIndustry, IPU and Technical University of Denmark.

Kramer P, Weckenmann A (2010) Simulative Abschatzung der Messunsi-cherheit von Messungen mit Rontgen-Computertomographie. in Kastner J,(Ed.) Proceedings – Industrielle Computertomografie, Shaker, Aachen243–248.

[71] Lister PF (2001) Applications of X-ray Computed Tomography in the WoodProducts Industry, Forintek Canada Corp, Vancouver, Canada2001.

[72] Ludwig W, King A, Reischig P, Herbig M, Lauridsen EM, Schmidt S, ProudhonH, Forest S, Cloetens P, Rolland du Roscoat S, Buffiere JY, Marrow TJ, PoulsenHF (2009) New Opportunities for 3D Materials Science of PolycrystallineMaterials at the Micrometre Lengthscale by Combined Use of X-rayDiffraction and X-ray Imaging. Materials Science and Engineering A524:69–76.

[73] Maass C, Knaup M, Sawall S, Kachelriess M (2010) ROI-Tomografie (LokaleTomografie). Proc. of iCT 2010, Industrielle Computertomografie Tagung, 27–29September 2010, Wels, Austria.

[74] Maass C, Knauer M, Kachelriess M (2011) New Approaches to Region ofInterest Tomography. Medical Physics 38(6):2868–2878.

[75] Maass C, Meyer E, Kachelriess M (2011) Exact Dual Energy Material Decom-position from Inconsistent Rays (MDIR). Medical Physics 38(2):691–700.

[76] Maass N, Stephan J (2012) Computed Tomography on Large Objects –Theoretical and Practical Solutions with Extended Field of View and ROI-CT. Conference Industrial Applications of CT Scanning at DTU, Lyngby, Denmark,12.06.2012.

[77] Maire E, Withers PJ (2014) Quantitative X-ray Tomography. InternationalMaterials Reviews 59(1):1–43.

[78] Marinello F, Savio E, Carmignato S, De Chiffre L (2008) Calibration Artefact forthe Micro Scale with High Aspect Ratio: the Fiber Gauge. CIRP Annals57(1):497–500.

[79] MarketResearch.com http://www.marketresearch.com/search/results.asp?categoryid=0&qtype=2&title=&publisher=Global+Industry+Analysts&query=computed+tomography&lprice=&uprice=&datepub=&regionid=&submit2=Search. Accessed 18.12.13.

[80] Meneghetti G, Ricotta M, Lucchetta G, Carmignato S (2014) An HysteresisEnergy-based Synthesis of Fully Reversed Axial Fatigue Behaviour of Differ-ent Polypropylene Composites. Composites Part B: Engineering. http://dx.doi.org/10.1016/j.compositesb.2014.01.027.

[81] Muller P (2013) Coordinate Metrology by Traceable Computed Tomography,(PhD thesis) Department of Mechanical Engineering, Technical University ofDenmark, Lyngby, Denmark.

[82] Muller P, Hiller J, Cantatore A, De Chiffre L (2012) A Study on EvaluationStrategies in Dimensional X-ray Computed Tomography by Estimation ofMeasurement Uncertainties. International Journal of Metrology and QualityEngineering 3(2):107–115.

[83] NEXIM. http://www.fi.dk/tilskud/bevillingsoversigter/2011/202-8-mio-til-strategisk-forskning-i-sundhed-foedevarer-og-velfaerd/projektresumeer.Accessed 18.12.13.

[84] Nanoeva – Center for Non-destructive Nano Evalualtion. http://www.nanoe-va.de/fileadmin/images/events/phone.pdf. Accessed 04.12.13.

[85] National Council on Radiation Protection and Measurements (2004) Report147, Structural Shielding Design for Medical X-Ray Imaging Facilities, NCRP,Bethesda, MD.

[86] Nardelli VC, Arenhart FA, Donatelli GD, Porath MC, Niggemann C, Schmitt R(2012) Feature-based Analysis for Quality Assessment of X-ray ComputedTomography Measurements. Measurement Science and Technology 23(10).art. no. 105006.

[87] NDT.net http://www.ndt.net/exhibit/list.php3?globalSearch=1&searchmo-de=AND&Find=computed+tomography+market. Accessed 18.12.13.

[88] Neuser E, Brunke O, Suppes A (2011) High Resolution Industrial CT Systems:Advances and Comparison with Synchrotron-Based CT. Proc. of InternationalSymposium on Digital Industrial Radiology and Computed Tomography, 20–22June 2011, Berlin, Germany.

[89] Nikon Metrology http://www.nikonmetrology.com/. Accessed 31.05.13.[90] NIST. http://www.nist.gov/pml/data/xraycoef/index.cfm. Accessed 31.05.13.[91] North Star Imaging http://4nsi.com/industrial-x-ray-news/taking-3d-

x-ray-inspection-to-a-new-dimension-4d-computed-tomography. Accessed31.05.13.

[92] North Star Imaging http://xviewct.com/. Accessed 31.05.13.[93] Oliveira FB, Porath MC, Nardelli VC, Arenhart FA, Donatelli GD (2013)

Investigation and Minimization of Thermal Drift Effects on TridimensionalCT Measurements. Annals of the 11th International Symposium of MeasurementTechnology and Intelligent Instruments, Aachen, 1–5 July 2013.

[94] Ontiveros S, Yague-Fabra JA, Jimenez R, Tosello G, Gasparin S, Pierobon A,Carmignato S, Hansen HN (2012) Dimensional Measurement of Micro-moulded Parts by Computed Tomography. Measurement Science and Technol-ogy 23(125401):9. http://dx.doi.org/10.1088/0957-0233/23/12/125401.

[95] Pani S, Arfelli F, Bravin A, Cantatore G, Longo R, Olivo A, Poropat P, Rigon L,Tromba G, Castelli E, Fantini A, Tartari A, GAmbaccini M, Pani R, Bollini D, DelGuerra A (2000) Tomographic Imaging with Synchrotron Radiation. PhysicaMedica 16(3):155–159.

[96] Paul Scherrer Institute http://www.psi.ch/niag/what-is-neutron-imaging.Accessed 13.01.14.

[97] Pejryd L, Beno T, Carmignato S (2014) Computed Tomography as a Tool for
Kramer P, Weckenmann A (2010) Multi-energy Image Stack Fusion inComputed Tomography. Measurement Science and Technology 21(4):1–7.045105.
Kramer P, Weckenmann A (2011) Modellbasierte simulationsgestutzteMessunsicherheitsbestimmung am Beispiel Roentgen-CT. VDI-Berichte2149: 5. VDI-Fachtagung Messunsicherheit 2011 – Messunsicherheit praxisger-echt bestimmen, 08–09 November 2011, Erfurt, 13–22.

Kruth J-P, Bartscher M, Carmignato S, Schmitt R, De Chiffre L, Weckenmann A(2011) Computed Tomography for Dimensional Metrology. CIRP Annals60(2):821–842.

Kurfiss M, Streckenbach G (2012) 3-Dimensional X-ray Inspection of VeryLarge Objects – 600 kV Digital Laminography Offers a Solution. 18th WorldConference on Nondestructive Testing, 16–20 April 2012, Durban, South Africa.

Examining Surface Integrity in Drilled Holes in CFRP Composites. 2nd CIRPConference on Surface Integrity.

[98] Pfeiffer F, Bech M, et al (2008) Hard X-ray Dark-field Imaging Using a GratingInterferometer. Nature Materials 7:134–137. http://dx.doi.org/10.1038/nmat2096.

[99] Pfeiffer F, Weitkamp T, Bunk O, David C (2006) Phase Retrieval and Differen-tial Phase-contrast Imaging with Low-brilliance X-ray Sources. Nature Physics2:258–261. http://dx.doi.org/10.1038/nphys265.

[100] Picouet PA, Teran F, Gispert M, Font M, Furnols M (2010) Lean ContentPrediction in Pig Carcasses, Loin and Ham by Computed Tomography (CT)Using a Density Model. Meat Science 86:616–622.

[101] Pittino G, Fritz L, Hadwiger M, Geier G, Rosc J, Habe D, Pabel T (2010)Stahlfaserbewehrter Spritzbeton – CT basierende Visualisierung und




http://www.goodwinsteelcastings.com/en/facilities/foundry

http://www.goodwinsteelcastings.com/en/facilities/foundry

















































































































http://www.marketresearch.com/search/results.asp%3Fcategoryid=0%26qtype=2%26title=%26publisher=Global+Industry+Analysts%26query=computed+tomography%26lprice=%26uprice=%26datepub=%26regionid=%26submit2=Search




http://dx.doi.org/10.1016/j.compositesb.2014.01.027

http://dx.doi.org/10.1016/j.compositesb.2014.01.027








http://www.fi.dk/tilskud/bevillingsoversigter/2011/202-8-mio-til-strategisk-forskning-i-sundhed-foedevarer-og-velfaerd/projektresumeer



http://www.nanoeva.de/fileadmin/images/events/phone.pdf

http://www.nanoeva.de/fileadmin/images/events/phone.pdf








http://www.ndt.net/exhibit/list.php3%3FglobalSearch=1%26searchmode=AND%26Find=computed+tomography+market

http://www.ndt.net/exhibit/list.php3%3FglobalSearch=1%26searchmode=AND%26Find=computed+tomography+market





http://www.nikonmetrology.com/

http://www.nist.gov/pml/data/xraycoef/index.cfm

http://4nsi.com/industrial-x-ray-news/taking-3d-x-ray-inspection-to-a-new-dimension-4d-computed-tomography



http://xviewct.com/





http://dx.doi.org/10.1088/0957-0233/23/12/125401





http://www.psi.ch/niag/what-is-neutron-imaging

http://www.psi.ch/niag/what-is-neutron-imaging




http://dx.doi.org/10.1038/nmat2096

http://dx.doi.org/10.1038/nmat2096

http://dx.doi.org/10.1038/nphys265






999)f the–11.ssed

dentp. on22.tiga-11th

ssed

011)ility

andssing

es – ISO

dorf.Mea-nda-dorf.ellennung

g forlogy

LB,Meat

thersity

ctionional

ting

intyphy.06–ww.

3D-pu-

stria,

Fully Cast

ty in


Berechnung der Orientierung von Stahlfasern. Proc. of iCT 2010, Wels, Austria,27–29 September 2010.

[102] Pough FH, Rogers TH (1947) Experiments in X-ray Irradiation of Gem Stones.American Mineralogist 32:31L 43.

[103] PRWeb press release http://www.prweb.com/releases/industrial_x_ray/inspection_systems/prweb8849962.htm. Accessed 04.12.13.

[104] Randall MG, Bose A (1997) Injection Molding of Metals and Ceramics. MetalPowder Industry. ISBN 13: 978-1878954619.

[105] Rebuffel V, Dinten J-M (2006) Dual-energy X-ray Imaging: Benefitsand Limits. Proc. 9th Eur. Conf. on Non Destructive Testing (ECNDT), Berlin,September 25–29, 2006.

[106] Requena G, Cloetens P, Altendorfer W, Poletti C, Tolnai D, Warchomicka F,et al (2009) Submicrometer Synchrotron Tomography of MultiphaseMetals Using Kirkpatrick–Baez Optics. Scripta Materialia 61:760–763.

[107] Romvari R, Dobrowolski A, Horn P, Allen P (2003) CT-Examination of PigCarcasses. EUPIGCLASS Workshop Roskilde, 6–7th October 2003.

[108] Roth H (2006) Aussagekraftige Resultate, QZ Jahrgang 51 (9), pp. 54–56, CarlHanser Verlag, Munchen.

[109] Salvo L, Di Michiel M, Scheel M, Lhuissier P, Mireux B, Suery M (2012) UltraFast In Situ X-Ray Micro-Tomography: Application to Solidification ofAluminium Alloys. Materials Science Forum 1713:706–709.

[110] Sauerwein C (2010) Rekonstruktionsalgorithmen und ihr Potenzial. Fraun-hofer IPA Workshop F 207: Hochauflosende Rontgen-Computertomographie-Messtechnik fur mikro-mechatronische Systeme, 79–90.

[111] Sehested E, Vangen O (1989) Computer Tomography, a Non destructiveMethod of Carcass Evaluation. 41st Annual Meeting of the EAAP, Helsinki1st July 1988.

[112] Standfest G, Kranzer S, Plank B, Petutschnigg A (2010) Dunky Anwendung derSub-mm CT zur Charakterisierung von Holzwerkstoffen. Proc. of iCT 2010,Wels, Austria, September 2010.

[113] Schardt P, Deuringer J, Freudenberger J, et al (2011) New X-ray TubePerformance in Computed Tomography by Introducing the Rotating Enve-lope Tube Technology. Medical Physics 31(9):2699–2706.

[114] Schonherr T, Altmann S (2009) Hitech: Hochauflosende Computertomografiein der Papiertechnik. PTS News 01/2009, 34–35.

[115] Schuster KC, Schadel U, Salaberger D, Kastner J (2008) Computertomographiezur Visualisierung von Wasser auf Textilien – Computed Tomography forVisualisation of Water Distribution on Textiles. Proc. of iCT 2008, Wels,Austria, 26–28 February 2008.

[116] Sørensen T (2012) CT Scanning in the Medical Device Industry. ConferenceIndustrial Applications of CT Scanning – Possibilities & Challenges in theManufacturing Industry, IPU and Technical University of Denmark.

[117] Spinelli M, Carmignato S, Affatato S, Viceconti M (2009) CMM-based Proce-dure for Polyethylene Non-congruous Unicompartmental Knee ProsthesisWear Assessment. Wear 267:753–756.

[118] SPS Inspection Systems http://www.shawinspectionsystems.com/products/comet/x-ray%20tubes.htm. Accessed 04.12.13.

[119] Stampanoni M, Wang Z, Thuring T, David C, Roessl E, Trippel M, Kubik-HuchRA, Singer G, et al (2011) The First Analysis and Clinical Evaluation of NativeBreast Tissue Using Differential Phase-Contrast Mammography. InvestigativeRadiology 46(12):801–806. http://dx.doi.org/10.1097/RLI.0b013e31822a585f.PMID 21788904.

[120] Sterzing A, Neugebauer R, Drossel W-G (2013) Metal Forming – Challengesfrom a Green Perspective. Proc. Intern. Conf. on Competitive Manufacturing(COMA 2013), Stellenbosch, South Africa, 19–24.

[121] Sun W, Brown SB, Leach RK (2012) An Overview of Industrial X-ray ComputedTomography. NPL Report ENG 32. ISSN 1754-2987.

[122] Szabo C, Babinsky L, Verstegen MWA, Vangen O, Jansman AJM, Kanis E (1The Application of Digital Imaging Techniques in the In Vivo Estimation oBody Composition of Pigs: a Review. Livestock Production Science 60(1):1

[123] Technical University of Denmark http://www.imaging.dtu.dk. Acce20.01.14.

[124] Tan Y, Kiekens K, Kruth J-P, Voet A, Dewulf W (2011) Material DepenThresholding for Dimensional X-ray Computed Tomography. Intern. SymDigital Industrial Radiology and Computed Tomography, Berlin, June 20–

[125] Tan Y, Kiekens K, Welkenhuyzen F, et al (2013) Simulation-aided Investion of Beam Hardening Induced Errors in CT Dimensional Metrology.

ISMTII Symposium, Aachen, July 1–5, 2013.[126] TU Munchen https://www.tum.de/. Accessed 27.01.13.[127] University of Manchester http://www.mxif.manchester.ac.uk. Acce

20.01.14.[128] Van Bael S, Kerckhofs G, Moesen M, Pyka G, Schrooten J, Kruth J-P (2

Micro-CT-based Improvement of Geometrical and Mechanical Controllabof Selective Laser Melted Ti6Al4V Porous Structures. Materials ScienceEngineering A: Structural Materials: Properties Microstructure and Proce528(24):7423–7431.

[129] VDI/VDE 2617 Part 13 (2011) Accuracy of Coordinate Measuring MachinCharacteristics and their Testing – Guideline for the Application of DIN EN10360 for Coordinate Measuring Machines with CT-sensors, VDI, Duessel

[130] VDI/VDE 2630 Part 1.2 (2010) Computed Tomography in Dimensional

surement – Influencing Variables on Measurement Results and Recommetions for Computed Tomography Dimensional Measurements, VDI, Duessel

[131] VDI/VDE 2630 Part 2.1 (2013) Computertomografie in der dimensionMesstechnik – Bestimmung der Messunsicherheit und der Prufprozesseigvon Koordinatenmessgeraten mit CT-Sensoren, VDI, Duesseldorf.

[132] Verburg JM, Seco J (2012) CT Metal Artifact Reduction Method CorrectinBeam Hardening and Missing Projections. Physics in Medicine and Bio57:2803.

[133] Vester-Christensen M, Erbou SGH, Hansen MF, Olsen EV, ChristensenHviid M, Ersbøll BK, Larsen R (2009) Virtual Dissection of Pig Carcasses.

Science 81(4):699–704.[134] Vester-Christensen M (2009) Image Registration and Optimization in

Virtual Slaughterhouse, (PhD thesis) DTU Informatics, Technical Univeof Denmark, Lyngby, Denmark.

[135] Yague-Fabra JA, Ontiveros S, Jimenez R, et al (2013) A 3D Edge DeteTechnique for Surface Extraction in Computed Tomography for DimensMetrology Applications. CIRP Annals 62(1):531–534.

[136] Wang X, Meier D (2011) MicroCT with Energy-resolved Photon-counDetectors. Physics in Medicine and Biology 56(May (9)):2791–2816.

[137] Weckenmann A, Kraemer P (2009) Assessment of Measurement UncertaCaused in the Preparation of Measurements Using Computed TomograIMEKO XIX World Congress Fundamental and Applied Metrology,

11.09.2009, Lisbon, Portugal, 978-963-88410-0-11888–1892, http://wimeko2009.it.pt/Papers/FP_145.pdf. as of 05.10.2009.

[138] Weckenmann A, Kramer P, Potschies B, Welz A (2008) Prufkorper fur dieComputertomographie von elektronischen Baugruppen. Industrielle Comtertomographie Fachtagung – Proceedings, 26–28.02.2008, Wels, AuAachen: Shaker, 2008, 978-3-8322-6949-459–64.

[139] Wenzel T, Stocker T, Hanke R (2009) Searching for the Invisible using

Automatic X-ray Inspection. Foundry Trade Journal of the Institute ofMetals Engineers 183(3666).

[140] Werth Messtechnik GmbH http://www.werth.de/. Accessed 31.05.13.[141] Wilhelm RG, Hocken R, Schwenke H (2001) Task Specific Uncertain

Coordinate Measurement. CIRP Annals 50–2:553–563.





http://www.prweb.com/releases/industrial_x_ray/inspection_systems/prweb8849962.htm

http://www.prweb.com/releases/industrial_x_ray/inspection_systems/prweb8849962.htm










































http://www.shawinspectionsystems.com/products/comet/x-ray%20tubes.htm

http://www.shawinspectionsystems.com/products/comet/x-ray%20tubes.htm

http://dx.doi.org/10.1097/RLI.0b013e31822a585f









http://www.imaging.dtu.dk/

http://www.imaging.dtu.dk/







https://www.tum.de/

http://www.mxif.manchester.ac.uk/

http://www.mxif.manchester.ac.uk/































http://www.imeko2009.it.pt/Papers/FP_145.pdf

http://www.imeko2009.it.pt/Papers/FP_145.pdf








http://www.werth.de/



Documents

Industrial applications of computed tomography · tomography, sometimes abbreviated to XCT or iCT (i for industrial), isthemethodofusingX-rayradiationtotakeanumber of two dimensional