Download in pdf format - Computer Vision Laboratory

221
Chapter 6
Bonding Defect
Imaging in Glulam with
Novel Air-Coupled
Ultrasound Testing
Sergio J. Sanabria
Laboratory for Wood Physics, Institute
for Building Materials, ETH Zurich,
Switzerland
Jürg Neuenschwander
Swiss Federal Laboratories for
Materials Science and Technology,
Switzerland
Roman Furrer
Swiss Federal Laboratories for
Materials Science and Technology,
Switzerland
Peter Niemz
Laboratory for Wood Physics, Institute
for Building Materials, ETH Zurich,
Switzerland
Urs Sennhauser
Swiss Federal Laboratories for
Materials Science and Technology,
Switzerland
ABSTRACT
The objective of this chapter is to provide an overview of novel non-destructive
testing methodologies for bonding quality assessment in glued laminated timber
developed within a recently completed Swiss National Science Foundation research
project (Sanabria, 2012). The focus is set on air-coupled ultrasound testing, which
has previously been applied to wood-based panels typically up to 50 mm thick. A
novel prototype capable of transmitting ultrasound signals through up to 500 mm
thick glulam was developed. A computerized-scanning system allowed imaging of
DOI: 10.4018/978-1-4666-4554-7.ch006
Copyright ©2014, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
the position and geometry of defects within the bonding planes. A normal transmission setup allows a global assessment of defective bonding planes stacks. Latest
results are as well shown for a recently patented slanted lateral transmission setup,
which allows separate assessment of individual bonding planes with unlimited beam
height and length. The investigations allowed an improved understanding of the
wave propagation phenomena in thick laminated timber components through both
analytical calculations and finite-difference numerical simulations. An overview
of the main findings is as well provided. Future research is planned to combine the
developed theoretical and experimental tools in a tomographic inspection method.
INTRODUCTION
The use of wood as a construction material currently experiences a renaissance
not only due to its undisputable renewable, environmental-friendly and aesthetic
nature, but as well owing to its high strength to weight ratio, durability and resistance against chemical attack, easy machining and predictable fire performance,
which are competitive with respect to established construction materials like steel
or reinforced concrete (Forest Products Laboratory, 2010).
During the last century, developments in fabrication technology and a declining volume of large, old-growth timber have led to the progressive replacement of
traditional solid wood structural members by highly engineered adhesively-bonded
composites. Glued laminated timber (glulam) is a layered composite manufactured
by gluing and stacking timber lamellas (for softwoods typically 35 to 40 mm thick
and 100 to 300 mm wide), which are sawn and finger-jointed parallel to the wood
grain direction. This arrangement homogenizes mechanical properties, and allows
the fabrication of large structural beam or column products of straight or curved
form, from 2 laminations up to typically 2.5 m high and 50 m long stacks (Bodig
& Jayne, 1982; Dunky & Niemz, 2002; EN 386, 2001).
With the proliferation of glued timber constructions, there is an increasing concern
about safety problems related to the delamination of timber glue lines. Delaminations are caused by manufacturing errors and in-service unfavorable load combinations (Hansson & Larsen, 2005). A moisture-resistant adhesive and a homogeneous
stress distribution in the glue lines are required for a durable bond. Polyurethane
adhesives require a minimal ambient humidity for curing whereas urea resins can
be hydrolytically removed due to temperature and humidity gradients (Kägi et
al., 2006; Schrödter & Niemz, 2006). Residual stresses are induced by improper
adhesive curing, long-sustained loads and swelling and shrinkage due to moisture
gradients in large glued cross-sections (Gustafsson et al., 1998). Bond delamination
is influenced by both the adhesive and wood phases (Frihart, 2009), the mechanisms
222
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
leading to debonding initiation are still not well understood (Bucur, 2011). Several
recent building collapses were related to bonding failure, which should be prevented
in the future with a timely defect detection (Blass & Frese, 2010).
It is therefore a demanding necessity to develop measurement technologies which
assess bonding defects in glued timber structural members during their full life cycle,
detecting both defective glue lines during manufacturing, tooling or processing and
in-service developed delaminations. Non-destructive testing technologies, which
probe the full structure without damage or modification and in a cost-efficient way,
are desirable. Apart from avoiding predictable hazards for lives and property, the life
span of existing glued timber constructions can be optimized if accurate information is available about when, where and which restoration of glue lines is required.
BACKGROUND
Current standardized delamination testing methods for timber laminates rely on
destructive tests for random specimens extracted from a specific production process or from an in-service structural member (increment core) (ANSI/AITC 19011-2007, 2007; EN 386, 2001). Mechanical tests and accelerated aging tests with
changing climatic conditions provide direct measurements of the ultimate strength
and fracture behavior, microscopic methods probe adhesive-wood interaction and
fracture at microstructural and chemical level. These methods provide information
on significant drifts from established production parameters and can identify onsite deterioration processes which affect the full structural member. However, they
can neither provide quality assurance for each single manufactured specimen nor
identify localized damage for in-service structures.
The established non-destructive testing methodologies for adhesive bond testing
in glued timber structures are still at a primary stage. Condition assessment practice
generally relies on techniques that have proven to be adequate for decay inspection,
such as visual inspection, tapping, resistance drilling, or stress wave timing in suspected damage regions, but which are not necessarily able to detect inner lamination
faults in structural laminated beams. Feeler gauges can be used to manually estimate
crack depth, the process is however slow and only detects edge open cracks with
straight paths. Proof loading provides information about the remaining structure
strength provided that live load can be applied to the structure in a cost-efficient
and safe way, which is often only possible for timber bridges (Aicher, 2008; Kasal,
2010; Ross et al., 2004).
An alternative to these methods is to non-destructively probe the structures with
propagating radiation or vibration fields, from which delamination signatures can be
extracted. The investigation of such methods for wood has significantly advanced
223
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
in the last two decades, yet some important limitations still exist. Thermography
allows accurate subsurface delamination imaging, but is limited to millimeterthick laminates, such as veneer overlays, due to the inherent thermal damping of
the porous wood microstructure (Meinlschmidt, 2005; Tsuchikawa et al., 2000).
The same restrictions apply for optic methods, which are based on the detection
of surface deformations (Castellini et al., 2008). X-rays are able to penetrate thick
structural members. Radiation safety needs to be strictly considered for a secure
equipment operation. Off-the-shelf portable systems are available which have been
applied in glulam constructions (Hasenstab et al., 2004). However, radiographic
images obtained with these systems provide a reduced sensitivity to debonding.
Applications for finger joint production control have been reported (Hu & Gagnon,
2007). A significantly higher contrast is obtained with three-dimensional computed
tomography reconstruction (Hu & Gagnon, 2007; Sirr & Waddle, 1999). However,
in this case a large number of radiographic projections is required in a full range of
orientations between sample and radiation path, which in structural applications is
often not feasible due to geometric and measurement time constraints.
Mechanical vibration methods have a long standing tradition in wood science;
however, they have only been applied to a minor extent to lamination assessment in
glued timber. Low-frequency (<1 kHz) modal analysis performs well for dynamic
measurements of global strength properties of timber laminates and is an interesting
alternative for quality assurance during manufacture (Gsell et al., 2007), difficulties in defect localization and a high sensitivity to boundary conditions however
constrain on-site application. Acoustic emission is based on passive sensor technologies which are potentially capable to detect and localize fracture initiation in
glulam (Dill-Langer et al., 1999), yet a practical application requires a large number
of sensors to be permanently embedded in each structural member. Ultrasound
testing is a well-established tool for non-destructive testing of adhesively bonded
composite materials due to its high sensitivity to cracking and delamination, which
behave like mechanical discontinuities to ultrasonic wave propagation (Maeva et
al., 2004). Moreover, the technology is safe, low-cost and easily portable on-site.
Typical testing frequencies for wooden structures range between 20 and 200 kHz.
Sufficient ultrasound energy coupling into structural timbers has traditionally required
the pressing of the transducers onto the sample surfaces (contact techniques). These
methods have shown potential for the detection and characterization of cracks and
delaminated areas in timber structural members. Stress wave timing allow edge open
crack depth estimates (Garab et al., 2010), the measurement of diffuse ultrasonic
energy (acousto-ultrasonics) is applicable to bond curing monitoring and finger-joint
production control (Anthony & Philipps, 1993; Beall, 1989). Large planar cracks and
lamination defects in structural glulam have been detected with longitudinal wave
through-transmission and shear wave pulse-echo measurements with perpendicular
224
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
insonification with respect to the bonding planes (Aicher et al., 2002; Dill-Langer et
al., 2005a, 2005b; Dimanche et al., 1994). A shear wave spring-loaded pulse-echo
array transducer for reinforced-concrete industry has been applied to single-sided in
situ inspection of glulam; cracks could be detected with both perpendicular and parallel insonification with respect to the lamination planes (Hasenstab, 2007; Hasenstab
& Osterloh, 2009). Overall, a minimum acceptable measurement reproducibility
(15% amplitude error) requires a delicate and generally manually supported adjustment of the coupling pressure at each measurement point (Dill-Langer et al., 2005b;
Hasenstab, 2006), which in practice restricts the inspection to a reduced number of
measurement points. Therefore, only rough defect position and geometry estimates
are practical. Moreover, ultrasonic waves strongly interact with the heterogeneous
structure of wood, leading to complex wave propagation phenomena and a strong
variability in the ultrasound signal detected in defect-free regions, which complicate
the discrimination between defects and material heterogeneity, especially if only
few measurement points are available. A discussion of the main ultrasound wave
propagation phenomena in glued timber structures is provided in the next sections.
Air-Coupled Ultrasound (ACU) is a relatively new transducer technology which
overcomes the contact limitation between ultrasound transducers and solid samples.
The transducers are held separated by an air gap from the sample surface, which
allows a simple acquisition of a large grid of measurement points, for example, by
means of a mechanical scanner, and a flexible positioning and orientation of the
transducers with respect to the inspected samples. This shows a high potential for
the discrimination of defects from background variability. Moreover, the measurement reproducibility significantly outperforms the one of contact techniques, with
<1% reproducibility error after one year measurement (Sanabria et al., 2010). These
highlights provide an unexplored potential for structural health monitoring based
on difference imaging with respect to a defined reference state. The classical constraint of ACU is a low coupling efficiency of ultrasound signals into solid materials
(<0.5% of input pressure into wood, Figure 1), which has traditionally limited the
inspection to thin plate materials. In the case of wood industry, <50 mm thick woodbased composites (particleboards, fiberboards, plywoods and laminated veneered
lumbers) are typically tested with this methodology (Benedetti, 2003; Blum, 1997;
Fagus GreCon Greten GmbH & Co. KG, 1994). A careful adjustment of the air
gap between transducers and sample by means of metering rollers in contact with
the sample allows a higher coupling of ultrasound energy (power sonic resonance)
and has achieved inspection of up to 200 mm thick laminated veneered lumbers
(Fuchs, 2011). Recent developments in off-the-shelf transducer technologies have
recently allowed tomographic inspection of 300 mm thick concrete without any
contact between sample and measurement system (Hall, 2011).
225
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
Figure 1. ACU delamination assessment principle. A quasi-specular reflection of
ultrasound waves is observed at wood-air interfaces, these lead on one hand to
substantial coupling loss (Lcoupling) and on the other hand to an increased signal
attenuation (Ljoint) across a bonding defect.
In this context, the investigation of the application of air-coupled ultrasound to
the bonding quality assessment of glued laminated timber structures is timely. A
four-year Swiss National Science Foundation research project has addressed this
research gap in the frame of a recently completed doctoral thesis (Sanabria, 2012).
The investigations covered three fundamental topics:
•
•
•
The development of a theoretical model of ultrasound wave propagation in
glulam based on state-of-the-art mechanical models for solid wood.
The design and implementation of an ACU prototype system capable of
transmitting and detecting ultrasound signals through structural multilayer
glulam in flexible configurations.
The theoretical and experimental investigation of the applicability of specific ACU inspection configurations for the detection and characterization of
bonding defects in glulam.
Next, a comprehensive overview of the results of these investigations is provided
including latest non-published results.
226
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
ACU Wave Propagation in Timber Laminates
The bonding assessment is based on the measurement of the attenuation of an
ultrasound beam transmitted between a transmitter/receiver transducer pair (Tx/
Rx) through a delaminated timber glue line (Figure 1). Ultrasonic waves are quasispecularly reflected at air/solid interfaces, which leads to a significantly increased
attenuation if a small air gap (crack or debond) is present at the glue line. The voltage level recorded at the receiver transducer is directly proportional to the pressure
level integrated on its surface. The ratio between the voltage level recorded through
a reference (defect-free) and a delaminated glue line gives the attenuation or contrast
Ljoint at the joint, which is conveniently expressed in logarithmic units with Ljoint =
6, 12, 20, 40, 60 dB corresponding to an attenuation factor of x2, x4, x10, x100,
x1000 in a linear scale. For insonification perpendicular to the bonding planes, the
maximum available contrast is given by Ljoint = 20 log10 (4ZaZw/(Za+Zw) 2), where
Za, Zw are the acoustic impedances of air and wood, defined as Z = (ρ C)0.5, where
ρ and C are the density and stiffness for each material.
In the case of Norway spruce (Picea abies Karst.), a softwood, one of the most
representative wood species used in glulam production, the maximum contrast is Ljoint
= 50 dB. The variations are not large between wood species, with for example Ljoint
= 56 dB for European beech (Fagus sylvatica L.), a representative hardwood. The
ultrasound beam is transparent to well-glued joints; the adhesive product therefore
does not play a significant role. The contrast is reduced with decreasing air gap
separation between timber lamellas and decreasing testing frequency, due to the
constructive interference of ultrasound pulse echoes at the glue joint. The maximum
contrast of Ljoint = 50 dB is reduced down to 20 dB for 10 µm thin air gaps between
delaminated interfaces and typical 100 kHz pulses. This value is of the order of the
size of the cellular micro voids in wood, which defines a minimum air separation
between timber lamellas, thus, in practice, an efficient blocking of the transmitted
ultrasound beam is expected by asymptotically thin discontinuities between timber
lamellas and with a lateral surface larger than the extension of the pressure field
incident into the glue line (Sanabria et al., 2010, 2010d).
The assessment of bonding planes requires the transmission of an ultrasound
beam through at least its adjacent timber lamellas. The wave propagation in timbers
is consequently strongly influenced by the heterogeneous and anisotropic wood
material structure. The density distribution, the annual ring structure and the grain
angle are the main material parameters influencing the ultrasound wave propagation. In a first approximation, a homogenized density distribution is assumed and
the long dimension of the lamellas is considered to be well-oriented to the wood
grain direction (L), the annual ring structure being the main influence in the wave
propagation. In particular, as a consequence of wood’s anisotropy, a normally inci227
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
dent ultrasound beam experiences significant trajectory shifts by an angle χ within a
timber lamella (Figure 2). The angle χ is a function of the ring angle ϕ between the
insonification axis and the direction tangential to the annual rings (T). In general,
the ring angle ϕ locally changes within each timber lamella, so that the angle χ is
position-dependent.
A two-dimensional Finite-Difference Time-Domain (FDTD) numerical simulation model was developed to simulate ACU wave propagation through arbitrary
heterogeneous and anisotropic material distributions. The model allows the direct
definition of material properties at each pixel with high accuracy and numerical
stability. By combining established orthotropic stiffness models of wood, which are
defined with respect to the material axes (L, R, T), and by performing local transformation of the stiffness tensor as a function of ϕ, the curvature of the annual rings
in individual timber lamellas is fully characterized. Air-coupling, bonding planes
and delaminations are directly implemented by defining the mechanical properties
of air and adhesive at the corresponding pixels. The two-dimensional model implements wave propagation through arbitrary sections of a glued laminated timber
beam.
The FDTD model was validated by comparing the simulated χ for lamellas with
constant ϕ against analytical calculations based on plane wave solutions of the wave
propagation equations. Figure 2a shows simulation time snapshots of the calculated
Figure 2. Finite-difference time-domain (FDTD) simulation of ACU wave propagation
through Norway spruce timber lamella as a function of annual ring orientation ϕ.
Due to the material anisotropy, the ultrasound beam deviates from the insonification
direction by a ϕ-dependent angle χ. a) Simulation snapshots of the stress fields σyy
through air and wood. b) Comparison of χ calculated from analytical plane wave
solutions with geometrically calculated values from FDTD snapshots, a total of 90
simulations were run for each of the represented ϕ - χ pairs.
228
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
stress fields σyy illustrating the ACU transmission of an ultrasound pulse through
a 40 mm thick timber lamella. At the interfaces (air-wood, or wood-wood) three
modes are coupled (quasi-longitudinal QP, quasi-shear in-plane QSV and shear outof-plane SH), out of which the QP mode is dominant. The shifts of the QP mode
range between -30° and 10° and tend to align the ultrasound beam with the principal
material axes (R and T), if the insonification direction coincides with these axes
no shifts (χ = 0°) are theoretically expected. The shift χ obtained from the FDTD
simulations is in excellent agreement with the analytical calculation (Figure 2-right),
the uncertainties (<3°) being rather due to small physical differences between the
limited-width pulsed ultrasound beam and the analytical plane wave front than due
to inaccuracies of the model. In fact, by doubling the width of the ultrasound beam
the uncertainties are reduced to <0.3°.
A similar agreement is observed between FDTD and analytical prediction for
other relevant wave propagation parameters, for example, the ultrasound transmission
through a delaminated glue line as a function of the air gap separation. In addition
to the described wave propagation phenomena, the strong gradients of mechanical properties across the periodical earlywood/latewood transitions in the annual
rings and the grain angle distortion around knots introduce strong scattering in the
transmitted pressure fields and contribute to the overall variability of the ultrasonic
signal in defect-free regions. The porous microstructure of wood introduces at 100
kHz a material attenuation coefficient of 2.4 dB cm-1 in the RT plane. Furthermore,
as a consequence of the strong anisotropy between grain (L) and cross-grain (RT)
directions, an incident circular-section ultrasound beam deforms into an elliptical
shape elongated along L when transmitted through wood, leading to a lower lateral
resolution along L. Since the described phenomena are accumulative, the wave
propagation becomes increasingly complex with longer propagation paths in wood.
Development of ACU Prototype for
Structural Timber Inspection
An air-coupled ultrasound system prototype was developed to inspect structural
timbers up to a thickness of 500mm (Figure 3). State-of-the-art high-efficiency aircoupled ultrasound transducers (50 mm diameter 100 kHz Gas Matrix Piezoelectric
Composites, The Ultran Group Inc., State College, PA, USA) were combined with
high-power pulsed electronics (>1000 Vpp), low-noise receiver electronics (input
referred noise of 0.9 nVrms Hz-0.5) and dedicated signal processing. This system
allowed bridging the total 115 dB signal loss through a defect-free 280 mm thick
glulam beam (including coupling loss and material attenuation) with a signal-to-noise
ratio (SNR) of 40 dB, which provided enough dynamic range to resolve bonding
defects from the variability in ultrasound signals introduced by the timber material.
229
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
Transmitter and receiver are positioned on two opposing surfaces of the inspected
glued laminated timber beam. An arbitrary waveform generator generates an ultrasonic
pulse which is amplified by the pulser and excited into the transmitter transducer.
The propagated signals are captured by the receiver transducer and amplified by the
low-noise receiver chain, the resulting time waveforms are digitized by an analog
to digital converter for further signal processing. The transmitter-receiver pair is
moved with respect to the sample surfaces with a computerized mechanical scanning
system; at each scanned pixel ultrasound signals are recorded. From the recorded
datasets, ultrasound images of the defect positions and geometries are calculated.
The excitation and recording of ultrasound signals are synchronized with the scanner movement by means of a control trigger. Due to the pulsed operation principle,
there is usually no need of physically blocking spurious signals transmitted in air
around the sample, which is particularly attractive for on-site inspection. Moreover,
the adjustment of the air gaps between transducers and sample is not critical, allowing for fully non-contact operation.
Air-Coupled Ultrasound Normal Transmission Testing
A global bonding quality assessment was investigated with a normal transmission
setup (NT). Figure 3 illustrates the geometrical configuration and data acquisition
sequence. The transmission is performed perpendicular to the bonding plane stack;
the transmitter and receiver transducers are scanned as a fix unit in a raster fashion
along the beam width (Y) and length (X). For each scanned position a pulsed ultrasound time waveform is recorded. The recorded datasets are processed to generate
images of the internal bonding defects of the sample; the peak value of a selected
Figure 3. Air-coupled ultrasound system referred to normal transmission setup
230
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
time portion of the recorded ultrasound waveforms is represented as a function of
the scanned position. The calculated amplitude values are represented normalized
in logarithmic scale with respect to its mean value in reference well-glued regions
(Amplitude(dB) = -Attenuation(dB)). Low signal levels are consequently associated
with defect positions.
Figure 4 demonstrates ACU NT imaging of delaminations induced by strong
climate variations. The test sample consists of two 20 mm thick lamellas, one
made from Norway spruce and the other one from European beech, which were
glued together with a one-component polyurethane adhesive. The samples were
then exposed to exterior climatic conditions for several months until delaminations
occurred as a consequence of internal stresses developed in the bond line due to
climate variations. The delamination process was accelerated by the gradient of
mechanical properties between both wood adherends. The delamination depth was
measured at the sample edges with a 100 µm thick feeler gauge and is marked in
the ACU image. Drops of ACU signal amplitude were consistently identified at
delamination positions. The ACU image shows additional delamination indications
(DI) at positions that were not accessible by the feeler gauge. Overall, a broad range
of bonding defects (non-glued areas, saw cuts, cracks, non-adherent glue, debonding
of vibration-welded joints) can be detected as long as there exists an asymptotically
thin material discontinuity between timber lamellas. For laminates of comparable
thickness, the lateral resolution in the images is limited by the beam width of the
ultrasound beam excited by the ACU transmitter (around 35 mm) With additional
deconvolution image processing, >12 mm diameter adhesive droplets were detected
in 10 mm thick test samples (Sanabria et al., 2009, 2010, 2010b).
Figure 4. ACU NT imaging of delaminations induced by strong climate variations.
The ACU assessment successfully detects the delamination areas identified with the
feeler gauge, moreover providing additional defect indications (DI).
231
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
As the thickness of the timber laminate increases, so does the variability of the
ultrasound signals recorded in defect-free regions and the spread of the ultrasound
beam within the sample, leading to an overall reduction in the lateral resolution of
the defect images. Figure 5 shows an amplitude profile along the length of a glulam
beam in which non-glued regions of specific lengths were defined. The same glulam
beam was assessed after gluing together, two, four and six timber lamellas, the gluing defects always located at the middle bonding plane. As expected, the lateral
resolution is anisotropic and lowest along the grain direction, with a 20 dB amplitude
drop at 22, 55 and 110 mm from the defect in the wood grain direction, and
at 9, 16 and 24 mm in cross-grain direction for 78, 150 and 230 mm thick glulam,
respectively. The described method is thus well-suited for the detection of lengthwise
elongated cracking and delamination, which frequently occur in glulam constructions (Sanabria et al., 2010c, 2010d, 2011a, 2011b).
Air-Coupled Ultrasound Slanted Lateral Transmission Testing
The setup outlined in Figures 3 to 5 performs a global assessment of the full laminated stack, allowing the identification of defective areas along their length and
width. For structural glulam with a large number of lamellas (beam height >300
mm), and especially for on-site inspection of glulam constructions, it is important
to determine which specific bonding planes are defective. This information allows
to quantify the structural relevance of the defect and to decide whether either no
action, reinforcement or a full replacement of the faulty member is required. A
novel ACU slanted lateral transmission setup (SLT) has been developed to address
these requirements (Sanabria, et al., 2011c, 2013) (Figure 6). An ultrasound beam
Figure 5. Detection of gluing defects along the length X of glulam beam as a function of the number of laminations. The method provides a global assessment of
bonding quality along the full lamination stack. The lateral resolution decreases
with increasing number of laminations.
232
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
Figure 6. Air-coupled ultrasound slanted lateral insonification of glulam: Left)
setup; Right) ultrasound waveforms in well-glued and defective bonding planes.
Individual bonding planes can be tested separately; the assessment is not limited
by the height and length of the glulam beam.
is transmitted and received through opposite lateral faces of the beam. The inclination of the ultrasound beam (typically <10°) is adjusted with respect to the wave
propagation models to achieve a controlled refraction path across single bonding
planes. A user-defined wave propagation path is obtained by adjusting the relative
orientation and inclination of the ACU transducers. The transducers are then scanned
as a single unit in beam height (Z) and length (X) directions. The assessment is not
limited by the height and length of the beam. Defects in individual bonding planes
are successfully separated. For example, in Figure 6 the third glue line contains a
full width non-glued area whereas the fourth glue line is defect-free. At the glued
position, the ultrasound beam is transmitted through the pre-calculated defect-free
wave propagation path (T1-R1) and a signal is received through wood at the Rx
(Glued). In the case of the defective glue line, the ultrasound beam refracted into
the sample is scattered by the gluing defect off the pre-calculated path (T2-R2), and
a reduced signal level for waves propagating through wood is therefore observed at
the receiver position R2 (Defect). Spurious paths diffracted in air around the sample
(Air) arrive delayed in time and can be filtered out in time with the pulsed electronics.
Defect maps indicating the position and extension of the delaminated areas are
obtained with this method, following a similar data evaluation principle of the recorded datasets as in the case of the ACU NT setup. Amplitude drops in the images are associated to bonding defect positions.
Figure 7 illustrates ACU SLT imaging of delaminations in a commercial multilayered glued laminated timber sample (200 x 395 x 750 mm3). The sample was
repeatedly tested, first as provided by the manufacturer, then after introducing one
saw cut D2 in the bonding plane B5, and finally after introducing additional saw
cuts D1, D3 and D4 in the bonding planes B3, B7 and B8, respectively. The mean
233
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
Figure 7. ACU SLT imaging of glulam sample with multiple bonding defect regions
D1, D2, D3, and D4 (saw cuts at bonding plane positions): a) setup; b) and c) ACU
SLT images. Experiments were performed for two specific measurement configurations leading to different refraction paths, so that the ultrasound beam interacted
with two timber lamellas (∆h = 2d, Figure 7b) and with three timber lamellas (∆h
= 3d, Figure 7c). For each case, ACU SLT images were obtained before introducing
the defects (left), after introducing the defect D2 (middle), and after introducing the
remaining defects D1, D3, and D4 (right). The assessment allows identification of
the defective bonding planes (B3, B5, B7, and B8), providing information about the
length and width extension of the defect areas.
234
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
timber lamella thickness is d = 40 mm. Experiments were performed for two specific
measurement configurations: in the first one the ultrasound beam refracted within
the sample interfaces a reduced portion of the beam height ∆h corresponding to
two timber lamellas (∆h = 2d, Figure 7b), in the second one the refracted beam
interfaces three timber lamellas (∆h = 3d, Figure 7c). The bonding planes B1, B2
… B9 are marked in the ACU SLT images with horizontal lines with respect to the
Z positions insonified by the transmitter transducer Tx. Full-width defects lead to
an amplitude reduction in the ACU SLT images in a region below the tested bonding plane with a Z extension equal to ∆h. This region corresponds to positions for
which the propagation of the transmitted ultrasound beam is blocked by the bonding defect (Figure 7a). Similarly, defect areas of a fraction f = 0…1 of the beam
width are identified in the ACU SLT images as a reduced amplitude region with a
Z extension of f·∆h.
For example, for ∆h = 2d = 80 mm (Figure 7b), full-width defects (D1, D2, and
D3) lead to a reduced ACU signal in a 80 mm Z-region below the respective defective planes B3, B5 and B7. Similarly, the half-width bonding defect D3 can be
identified as a reduced amplitude region with a Z extension of 0.5∆h = 40 mm
below the defective plane B8. In the case of the SLT assessment with ∆h = 3d
(Figure 7c), the full-width defects introduce attenuation regions of ∆h = 3d = 120
mm, whereas the half-width defect introduces an attenuation region of 0.5∆h = 60
mm. The defective bonding planes are consistently identified in both measurement
configurations. The length of the defective areas is directly determined from the X
extension of the reduced amplitude regions in the ACU SLT images.
The choice of the insonified beam height portion ∆h is a trade-off between
ultrasound signal variability in defect-free regions, contrast at defect areas and
signal-to-noise ratio at the receiver transducer. The higher ∆h the longer the wave
propagation paths in wood and consequently the higher the signal variability and the
higher the attenuation – consequently the lower the signal-to-noise ratio. Conversely,
the higher ∆h the smaller the sample width fraction of the tested bonding planes
interacting with the ultrasound beam and thus the better lateral resolution along
the sample width. The higher ∆h as well the lower the spurious ultrasound energy
diffracting around the defect being captured by the receiver and thus the higher
contrast at defective areas. Overall, values of ∆h = K·d with K = 1…3 were shown
to provide a good performance. The values of K were generally chosen as integer
values 1, 2 and 3 for simplicity in the interpretation of the ACU images, although
other values are as well possible. For K > 1, an overlap of the information extracted
from adjacent bonding planes is present in the ACU images. This is for example
clearly visible in Figure 7c, where the bonding defects D1 and D2 are present in the
same length region X and in two bonding planes B3 and B5 separated by two timber
lamellas. Since in this case ∆h = 3d, there is an overlap between the defect regions
235
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
D2 and D1 in the ACU image, which reduces the discrimination capacity along Z.
However, since delaminations are assumed to be scarcely located in both X and Z,
this overlaps are in general not problematic, and otherwise can be further resolved
by repeating the assessment with lower K in the conflictive regions.
The described SLT setup assumes a constant wave propagation path between
transmitter and receiver transducers. However, as discussed in the theoretical section, the wave propagation paths in timber laminates are highly influenced by the
annual ring structure in the timber lamellas. In order to gain further understanding
about these mechanisms, finite-difference time-domain simulations were run for
specific defect inspection scenarios as a function of the annual ring curvature, which
was characterized by the position of the stem pith for each sawn timber lamella. In
order to keep the number of case studies tractable, a series of practical simplifications were introduced. The annual rings were assumed to grow cylindrically with
respect to the tree stem. All lamellas were of a same representative size (40 x 200
mm2) and showed the same year ring curvature. The horizontal pith position was
equal to center width of the timber lamella. The vertical position of the pith y was
chosen as a multiple of the lamella width d, with y/d = -3, -2, -1, 0, 1, 2, 3. Negative and positive y/d values correspond to convex and concave year ring curvatures,
respectively. y/d = 0 corresponds to a scenario in with the pith P is located at middle
lamella width. The highest absolute y/d values are given by the usual diameter of the
tree stems from which the lamellas are cut from (Norway spruce). Gluing defects
of 25%, 50%, 75% and 100% width section were assessed with ∆h = 2d, the defects
were centered at middle width except for the 50% width defects, which were as well
simulated at both edge positions. The insonification position was adjusted to achieve
the maximum possible interaction with the defect. The theoretical wave propagation
paths were calculated by assuming an isotropic material with sound speed equals to
the longitudinal wave velocity parallel to the timber lamellas. The orientation and
position of the ACU transmitter Tx was adjusted accordingly.
Figure 8 represent the waveform peak stress distribution (σyy) simulated within
the glued laminated timber samples as a function of y/d and the defect size, which
gives an idea of the main wave propagation paths. Overall, the annual ring structure
has a strong influence on the ultrasound wave propagation, leading to an energy
distribution which is not concentrated within the calculated propagation paths, but
that to a large extent “follows” the annual ring curvature. This leads to significant
position drifts in the wave propagation paths, and to spurious signal coupling in the
timber lamellas adjacent to the inspected ones.
Some general trends can be identified. Differentiated stress distributions are
obtained for concave and convex year ring orientations. The simulated stress distribution fits closer to the theoretical wave propagation paths the higher the absolute
value of y/d, namely the lower the annual ring curvature. Above y/d = 2 and below
236
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
Figure 8. Finite-difference time-domain simulation of ACU SLT inspection of glued
laminated timber beams for specific defect scenarios and year ring curvatures. The
peak waveform stress σyy at each pixel of the computation domain is represented,
normalized with respect to the mean value along Z. The same logarithmic scale is
applied to all snapshots. The pixel size is 200 µm and the time step 100 ns. The ring
curvature is the same for all lamellas of each sample and defined with respect to
the vertical position y of the pith with respect to the lamella thickness d (40 mm).
The size of the simulated glued laminated timber cross-sections is 200 x 320 mm2.
237
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
y/d = -2 the observed variations were negligible. The highest beam distortion was
observed when the pith was present within the lamellas (y/d = 0). This effect is as
well experimentally observed in Figure 7. A reduced signal amplitude is consistently observed along X around Z positions 250 to 350 mm (Figures 7b, 7c), corresponding to the interactions of the ultrasound beam with the timber lamella between
B6 and B7, which contains the pith. In practice, such timber lamellas are generally
avoided during glulam manufacture due to their lower mechanical properties.
An analysis of the attenuation of ultrasound signals simulated at the receiver
transducer with specific defect sizes with respect to well-glued material revealed for
y/d = 3 attenuation factors of -45 dB, -21 dB, -6 dB and -3 dB, for defect sections of
100%, 75%, 50% and 25%. Considering the signal variability in defect-free regions,
the detectability of defects sections <50% is problematic. An in-depth discussion
of the SLT method including several techniques to improve the robustness of the
assessment and defect detectability can be found in Sanabria et al. (2013).
COMPLEMENTARY INVESTIGATIONS
In this chapter the discussion has been restricted to air-coupled ultrasound methodologies for bonding quality assessment of timber laminates. Within the framework of
the project, additional investigations were carried out in the field of non-destructive
bonding quality assessment of glued timber.
An automatized ultrasonic point contact method was developed to optimize the
data acquisition and reproducibility of the ultrasonic contact technique. With this
purpose, the off-the-shelf shear wave spring-loaded pulse-echo array transducer,
which was previously discussed in the Background section, was combined with
the computerized mechanical system used in the ACU investigations. The method
is single-sided and shows an amplitude reproducibility error <10%. Experiments
were performed for commercial glued laminated timber samples with artificial defects and for a 90 year old roofing glulam. With appropriate data evaluation, edge
delaminations deeper than 20 mm can be successfully detected in the signature of
the surface wave and large scale delaminations (>80% of beam width) in the backwall echo. The data interpretation was supported by finite-difference time-domain
simulations (Neuenschwander et al., 2013; Sanabria et al., 2011b).
An X-ray limited-angle computed tomography (LCT) method has been investigated for the detection of gluing defects in timber laminates. In order to overcome
the limitations associated to conventional computer tomography methods (previously
discussed in the Background section), only a reduced number of radiographies in
a small angular range (0.6 to 1.8°) parallel to the inspected timber glue lines were
used in the three-dimensional reconstruction. The method was implemented in a
238
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
laboratory environment with a microfocus system. A dramatic reduction in the
measurement and reconstruction times was achieved. Moreover, density gradients
in the wood timber lamellas (earlywood/latewood transitions, wood knots) were
filtered out while highlighting the information from the glue lines. The assessment
is theoretically independent of the height and length of the inspected sample. Air
gaps down to 150 µm could be detected in 10 and 65 mm thick samples with a lateral
resolution >5 mm (Sanabria et al., 2011d).
Finally the ACU NT method described has been applied to the characterization of
material properties in particleboard composites. ACU images of ultrasound velocity
and amplitude were correlated with X-ray radiographies of the density distribution of
the samples. Correlations between acoustic parameters and the density and particle
geometry were found, which were as well influenced by the testing frequency. FDTD
simulations allowed interpretation of the observed trends in terms of multi-scale
porosity and grain scattering (Sanabria et al., 2013b).
CONCLUSION
Air-Coupled Ultrasound (ACU) is a well-suited non-destructive testing technology for the structural health monitoring of glued laminated timber, both during
manufacture and at the construction site. The inspection does not require contact
between transducers and samples, which leads to highly reproducible measurements (<1% amplitude error). Moreover, a large grid of measurement points can
be efficiently acquired by attaching the transducers to a computerized scanning
system, the ultrasound beam excitation can be flexibly controlled by adjusting the
transducer position and orientation. This allows higher resolution in the imaging
of defect positions and geometries, together with a better differentiation of defects
from background material variability. The classical constraint of this technology is
the inefficient coupling of ultrasound waves into solid materials, which has classically limited the assessment to <50 mm thick wood-based composites. However,
with a careful system design based on state-of-the-art transducers and electronics,
the feasibility of ACU transmission through up to 500 mm thick glued laminated
timber has been demonstrated.
Ultrasonic waves strongly interact with air-solid discontinuities and therefore
show a high sensitivity to cracking and debonding. An ultrasound beam transmitted with normal incidence through a delaminated timber glue line is attenuated by
up to 50 dB, gaps as thin as 10 µm provide a contrast of 20 dB. Ultrasonic waves
do not only interact with the inspected glue lines, but also with the adjacent timber
lamellas. Wood is an anisotropic, heterogeneous and porous material, all of which
have a strong influence in the wave propagation. An ultrasound beam propagating
239
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
within timber is gradually shifted from the insonification direction along trajectories, which are a complex function of the annual ring and grain structure. Moreover,
mechanical gradients at annual ring and knot heterogeneities scatter the transmitted
fields. The porous structure additionally introduces strong damping in the signals
transmitted through wood. A finite-difference time-domain simulation tool has been
developed to efficiently simulate pulsed ultrasound wave propagation in timber
laminates, arbitrarily heterogeneous and anisotropic materials can be simulated by
locally defining the mechanical properties at each pixel.
A normal transmission setup has been first investigated, which transmits an
ultrasound beam perpendicular to the bonding stack, providing a global assessment for each length and width position along the glued laminated timber sample.
Delaminations and cracks induced by exterior climatic conditions were successfully
detected. The ACU assessment is in excellent agreement with delamination indications obtained by sizing edge open cracks with a feeler gauge, providing moreover
information about inner defects. For thin laminates (<40 mm) the lateral resolution in the images is limited by the beamwidth coupled by the ACU transducers,
defects <20 mm size are typically assessed. As the thickness of the timber laminate
increases, so does the variability of the signals recorded at defect-free regions and
the spread of the ultrasound beam, leading to an overall lateral resolution loss. The
reduction in lateral resolution is more significant in grain than in cross-grain direction, making the method well-suited for the assessment of lengthwise elongated
cracking and delamination.
For structural glulam with a large number of lamellas (beam height >300
mm) and constrained access to the beam surfaces parallel to the bonding planes, a
slanted lateral transmission setup has been proposed. Here an ultrasound beam is
coupled with a small inclination with respect to the inspected glue lines. This setup
allows determining which specific bonding planes of the glued stack are defective.
Moreover, it is not limited by the height and length of the inspected sample. The
applicability of this method has been successfully validated in multi-layered glued
laminated timber beams; the position and geometry of multiple defects were successfully identified. The defect detectability limits were as well discussed with the
help of finite-difference time-domain simulations. The timber lamellas strongly
scatter the incident ultrasound fields, leading to an energy distribution which is not
concentrated within the pre-calculated propagation paths. The energy distribution
is strongly coupled with the annual ring structure, leading to significant position
drifts in the wave propagation paths, and spurious signal coupling in the timber
lamellas adjacent to the inspected ones. Defects down to 50% width section were
successfully assessed in commercial glued laminated timbers with general annual
ring orientations in each timber lamella. The described method an installation led
to a Swiss patent application (Sanabria et al., 2011c).
240
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
FUTURE RESEARCH DIRECTIONS
An in-depth discussion of the SLT method including several techniques to improve
the robustness of the assessment and defect detectability is under preparation.
So far, the developed Finite-Difference Time-Domain (FDTD) numerical simulation model has been applied to predict and interpret the ACU wave propagation
phenomena occurring in timber laminates for a priori defined transducer excitation, material properties and defect configurations. The model has been verified
with analytical results and shows a good quantitative agreement with parametric
experimental studies (not shown). The question naturally rises, whether the inverse
problem is as well feasible, that is, whether experimentally acquired ultrasound
fields can be used as input to the simulation model. The objective is to partially
compensate for known material heterogeneity and anisotropy in wood material, in
order to improve the detectability and lateral resolution of the bonding assessment.
Further research is planned in this direction.
The developed NT and SLT setups have been tested with real glued laminated
timber samples in a laboratory environment. With minor modifications they are
portable to both glulam production lines and the construction site. A dedicated
study for a larger volume of laminates is still necessary to establish the technology
in both scenarios.
REFERENCES
Aicher, S. (2008). Verfahren und Aussagemöglichkeiten bei der Begutachtung von
Holzkonstruktionen. Paper presented at the Fachtagung Bauwerkdiagnose. Berlin,
Germany.
Aicher, S., Dill-Langer, G., & Ringger, T. (2002). Non-destructive detection of
longitudinal cracks in glulam beams. Otto-Graf-Journal, 13, 165–181.
ANSI/AITC 1901-1-2007. (2007). Standard for wood products: Structural gluedlaminated timber.
Anthony, R. W., & Philipps, G. (1993). An update on acousto-ultrasonic applied to
finger joints. Paper presented at the 9th International Symposium on Non-destructive
Testing of Wood. Madison, WI.
Beall, F. C. (1989). Monitoring of in-situ curing of various wood-bonding adhesives using acousto-ultrasonic transmission. International Journal of Adhesion and
Adhesives, 9, 21–25. doi:10.1016/0143-7496(89)90142-5.
241
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
Benedetti, P. (2003). Method and device for testing the presence of defects in the
production of plane boards. EP Patent 1324032A1.
Blass, H. J., & Frese, M. (2010). Schadensanalyse von Hallentragwerken aus Holz.
Tech. rep. Karlsruhe, Germany: Karlsruher Institut für Technologie (KIT).
Blum, R. (1997). Verfahren zur Erkennung von Spaltern in Span- und MDF-Platten
und Vorrichtung zur Durchführung des Verfahrens. DE Patent 19519669C1.
Bodig, J., & Jayne, B. A. (1982). Mechanics of wood and wood composites. Scarborough, Canada: Van Nostrand Reinhold Publishing.
Bucur, V. (2011). Delamination in wood, wood products and wood-based composites.
Berlin: Springer. doi:10.1007/978-90-481-9550-3.
Castellini, P., Abaskin, V., & Achimova, E. (2008). Portable electronic speckle
interferometry device for the damages measurements in veneered wood artworks.
Journal of Cultural Heritage, 9(3), 225–233. doi:10.1016/j.culher.2008.05.002.
Dill-Langer, G., Aicher, S., & Bernauer, W. (2005). Reflection measurements at timber
glue-lines by means of ultrasound shear waves. Otto-Graf-Journal, 16, 273–283.
Dill-Langer, G., Bernauer, W., & Aicher, S. (2005b). Inspection of glue-lines of
glued-laminated timber by means of ultrasonic testing. Paper presented at the 14th
International Symposium on Nondestructive Testing of Wood. Eberswalde, Germany.
Dill-Langer, G., Höfflin, L., & Aicher, S. (1999). Fracture of solid wood and glulam
in tension perpendicular to the grain monitored by acoustic emission. Paper presented at the 1st RILEM Symposium on Timber Engineering. Stockolm, Sweden.
Dimanche, M., Capretti, S., Del Senno, M., & Facaoaru, I. (1994). Validation of
theoretical approach for the detection of delamination in glued laminated beams.
Paper presented at the First European Symposium on Nondestructive Evaluation
of Wood. Sopron, Hungary.
Dunky, M., & Niemz, P. (2002). Holzwerkstoffe und Leime: Technologie und Einflussfaktoren. Berlin, Germany: Springer. doi:10.1007/978-3-642-55938-9.
EN 386. (2001). Glued laminated timber - Performance requirements and minimum
production requirements.
Fagus GreCon Greten GmbH & Co. KG. (1994). Device for monitoring the quality
of laminar materials, preferably based on wood, such as chipboard or plywood, by
means of an ultrasonic transmitter. DE Patent 3742844C2.
242
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
Forest Products Laboratory. (2010). Wood handbook - Wood as an engineering
material. Tech. Rep. FPL-GTR-190. Madison, WI: Department of Agriculture,
Forest Service.
Frihart, C. R. (2009). Adhesive groups and how they relate to the durability of bonded wood. Journal of Adhesion Science and Technology, 23, 601–617.
doi:10.1163/156856108X379137.
Fuchs, M. (2011). New technologies for glow and blister detection widens on-line
evaluation of board characteristics by implementing new calibration methods. Holztechnologie, 52(5), 17–21.
Garab, J., Toth, A., Szalai, J., Bejo, L., & Divos, F. (2010). Evaluating glued laminated beams using a non-destructive testing technique. Transactions of Famena,
34(4), 33–46.
Gsell, D., Feltrin, G., Schubert, S., Steiger, R., & Motavalli, M. (2007). Cross-laminated timber plates: Evaluation and verification of homogenized elastic properties.
Journal of Structural Engineering, 133(1), 132–138. doi:10.1061/(ASCE)07339445(2007)133:1(132).
Gustafsson, P. J., Hoffmeyer, P., & Valentin, G. (1998). DOL behaviour of endnotched beams. European Journal of Wood and Wood Products, 56, 307–317.
doi:10.1007/s001070050325.
Hall, K. S. (2011). Air-coupled ultrasonic tomographic imaging of concrete elements.
(PhD Thesis). University of Illinois at Urbana-Champaign, Urbana, IL.
Hansson, M., & Larsen, H. (2005). Recent failures in glulam structures and
their causes. Engineering Failure Analysis, 12(5), 808–818. doi:10.1016/j.engfailanal.2004.12.020.
Hasenstab, A. (2006). Integritaetspruefung von Holz mit dem Zerstoerungsfreien
Ultraschallechoverfahren. (PhD Thesis). Technische Universitaet Berlin, Berlin,
Germany.
Hasenstab, A. (2007). Ultraschall-Echo zur Ortung von Rissen in Brettschichtholz
(BSH). Paper presented at the DGZfP-Jahrestagung. Fürth, Germany.
Hasenstab, A., & Osterloh, K. (2009). Defects in wood non-destructive locating
with low frequency ultrasonic echo technique. Paper presented at the NCDTE’09,
Non-Destructive Testing in Civil Engineering. Nantes, France.
243
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
Hasenstab, A., Osterloh, K., Robbel, J., Krause, M., Ewert, U., & Hillemeier, B.
(2004). Mobile Röntgenblitzröhre zum Auffinden von Holzschäden. Paper presented
at the DGZfP-Jahrestagung. Salzburg, Germany.
Hu, L. J., & Gagnon, S. (2007). X-ray-based scanning technique for non-destructive
evaluation of finger-joint strength. Paper presented at the 15th International Symposium on NDT of Wood. Deluth, MN.
Kägi, A., Niemz, P., & Mandallaz, D. (2006). Influence of moisture content and
selected technological parameters on the adhesion of one-part polyurethane adhesives under extreme climatic conditions. European Journal of Wood and Wood
Products, 64, 261–268.
Kasal, B. (2010). In-situ assessment of structural timber: state-of-the-art, challenges
and future directions. Advanced Materials Research, 133-134, 43–52. doi:10.4028/
www.scientific.net/AMR.133-134.43.
Maeva, E., Severina, I., Bondarenko, S., Chapman, G., O’Neill, B., & Severin, F. et al.
(2004). Acoustical methods for the investigation of adhesively bonded structures:
A review. Canadian Journal of Physics, 82(12), 981–1025. doi:10.1139/p04-056.
Meinlschmidt, P. (2005). Thermographic detection of defects in wood and woodbased materials. Paper presented at the International Symposium of Non-destructive
Testing of Wood. Hannover, Germany.
Neuenschwander, J., Sanabria, S. J., Schütz, P., Widmann, R., & Vogel, M. (2013).
Delamination detection in a 90 year old glulam block with scanning dry point-contact
ultrasound. Holzforschung. doi:10.1515/hf-2012-0202.
Ross, R., Brashaw, B. K., Wang, X., White, R. H., & Pellerin, R. F. (2004). Wood and
timber - Condition assessment manual. Madison, WI: Forest Products Laboratory.
Sanabria, S. J. (2012). Air-coupled ultrasound propagation and novel non-destructive
bonding quality assessment of timber composites. (PhD Thesis). ETH Zurich, Zurich, Switzerland.
Sanabria, S. J., Furrer, R., Neuenschwander, J., Niemz, P., & Sennhauser, U. (2010c).
Air-coupled ultrasound wave propagation in glued laminated timber structures applied to bonding quality assessment. Paper presented at the 2010 IEEE International
Ultrasonic Symposium. San Diego, CA.
Sanabria, S. J., Furrer, R., Neuenschwander, J., Niemz, P., & Sennhauser, U. (2011).
Air-coupled ultrasound inspection of glued laminated timber. Holzforschung, 65,
377–387. doi:10.1515/hf.2011.050.
244
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
Sanabria, S. J., Furrer, R., Neuenschwander, J., Niemz, P., & Sennhauser, U. (2013).
Novel slanted incidence air-coupled ultrasound method for delamination assessment
in individual bonding planes of structural multi-layered glued timber laminates.
Ultrasonics, 53(7), 1309-1324. doi: 10.1016/j.ultras.2013.03.017.
Sanabria, S. J., Hilbers, U., Neuenschwander, J., Niemz, P., Sennhauser, U., & Thömen, H. et al. (2013b). Modeling and prediction of density distribution and microstructure in particleboards from acoustic properties by correlation of non-contact
high-resolution pulsed air-coupled ultrasound and X-ray images. Ultrasonics, 53(1),
157–170. doi:10.1016/j.ultras.2012.05.004 PMID:22677469.
Sanabria, S. J., Mueller, C., Neuenschwander, J., Niemz, P., & Sennhauser, U. (2009).
Air-coupled ultrasound inspection of glued solid wood boards. Paper presented at
the 16th International Symposium on Nondestructive Testing and Evaluation of
Wood. Beijing, China.
Sanabria, S. J., Mueller, C., Neuenschwander, J., Niemz, P., & Sennhauser, U. (2010a).
Air-coupled ultrasound as an accurate and reproducible method for bonding assessment of glued timber. Wood Science and Technology, 45, 645–659. doi:10.1007/
s00226-010-0357-z.
Sanabria, S. J., Mueller, C., Neuenschwander, J., Niemz, P., & Sennhauser, U. (2010b).
Nondestructive glue line assessment in timber laminates with an air-coupled ultrasonic technique. In V. Bucur (Ed.), Delamination in wood, wood products and wood
based materials. Berlin, Germany: Springer. doi:10.1007/978-90-481-9550-3_19.
Sanabria, S. J., Neuenschwander, J., Niemz, P., & Sennhauser, U. (2010d). Structural
health monitoring of glued laminated timber with a novel air-coupled ultrasound
method. Paper presented at the 11th World Conference on Timber Engineering,
Riva del Garda. Trentino, Italy.
Sanabria, S. J., Neuenschwander, J., Niemz, P., & Sennhauser, U. (2011b). Monitored
assessment of structural integrity of multilayered glued laminated timber beams
with air-coupled ultrasound and contact ultrasound imaging. Paper presented at the
17th International Non-destructive Testing and Evaluation of Wood Symposium.
Sopron, Hungary.
Sanabria, S. J., Neuenschwander, J., & Sennhauser, U. (2011c). Air-coupled ultrasonic contactless method for non-destructive determination of defects in laminated
structures. CH Patent Application 01354/11, PCT/EP2012/065593.
245
Bonding Defect Imaging in Glulam with Novel Air-Coupled Ultrasound Testing
Sanabria, S. J., Wyss, P., Neuenschwander, J., Niemz, P., & Sennhauser, U. (2011d).
Assessment of glued timber integrity by limited-angle microfocus X-ray computed tomography. European Journal of Wood and Wood Products, 69, 605–617.
doi:10.1007/s00107-010-0509-8.
Schrödter, A., & Niemz, P. (2006). Studies on the influence of temperature and
timber moisture on the failure behaviour of selected adhesives under tensile load.
Holztechnologie, 47, 24–32.
Sirr, S. A., & Waddle, J. R. (1999). Use of CT in detection of internal damage and
repair and determination of authenticity in high-quality bowed stringed instruments.
Radiographics, 19(3), 639–646. PMID:10336193.
Tsuchikawa, S., Takahashi, T., & Tsutsumi, S. (2000). Non-destructive measurement of wood properties by using near-infrared laser radiation. Forest Products
Journal, 50(1), 81–86.
246

Download Report