Technical Memo

Preliminary Review of Non-Destructive Tests and Coring results

Project: KVMRT - Malaysia

Structure: Tunnels
Object: Concrete Platform

IMM SA Job N.: 21-0050

Authors: M. Di Tommaso (MDT) Ref: TM-0050A

Recipients: MMC Gamuda Date: 10/05/2021

Annexes/ --
Format:

Background
The undersigned, consultant in concrete technologoy and general manager of IMM SA (Swiss
accredited materials testing laboratory and consulting firm in the field of construction materials
technology), has been requested by Messrs. MMC Gamuda KVMRT SDN BHD, hereafter referred to as
the client, to preliminary review the results of the non-destructive testing and coring campaign used to
evaluate a section of about 355 m in length, located in TU4 and TU3 tunnels of the above mentioned
project. These further investigations into the concrete condition were prompted by the NCR reported in
figure 1, where contamination of the area of placement of concrete was observed during concreting
works.

TM0050A Page 1 of 13
 Figure 1.

TM0050A Page 2 of 13
Structural Detailing
It is understood by the undersigned that the original design for the undeground trackworks was as
shown in figure 2, but the actual construction solution adopted is shown in figure 3.
Figure 2.

Figure 3.

As it has been acknowledged by us, the serviceability of stage 2 concrete for its intended scope must
be granted for the first 250 mm of an ideal concrete strip, in a similar fashion to the layout for elevated
sections reported in figure 4 below. Hence, for safety, the most critical part can be taken as a strip of
300 mm wide concrete spanning from the outer side towards the tunnel walls. Concrete is, finally,
defined as massive because of the light reinforcement used (control of cracking only).

TM0050A Page 3 of 13
 Figure 4.

Non-Destructive Surveys
Messrs. IMM – SOIL Bhd, an ISO/CEI 17025 accredited laboratory was appointed by the client to carry
out non-destructive testing (NDT) surveys of the condition of the concrete of the trackwork for the stage
2 casting of concrete (total length of 355 m x 1 m width, from Ch. 23’960 to Ch 24’210).
The concept of the intervention was to cover the entire surface with a high frequency ground
penetrating radar scanner (GPR) to evaluate possible defects within the layer of concrete of stage 2,
as sketched in figure 5 below. This approach, which is fast and very reliable as our professional
experience and as the massive amount of literature data show, allows to pin-point doubtful areas where
a further, more detailed but even more time consuming, Ultrasound Pulse Echo (UPE) scanner can be
used to investigate potential defects and mechanical weaknesses in more detail.
Figure 5.

TM0050A Page 4 of 13
The chosen approach of combining an electromagnetic technique (GPR) with a sonic technique (UPE)
is considered adequate for the scope and it is, furthermore, recommended by state-of-the-art reports
in the literature on the subject, such as Rilem’s Non-Destructive Assessment of Concrete Structures:
Reliability and Limits of Single and Combined Techniques or IAEA’s (International Atomic Energy
Agency) Guidebook on non-destructive testing of concrete structures.

GPR Surveys
GPR surveys were conducted by Messrs IMM-SOIL lab from 24th April 2021 to 02nd May 2021, in the
presence of representatives from the client, for APPS to KBNS station and, in order to make comparison
between the concrete quality in these suspected locations with a reference zone where documented
quality of concrete works was good, another site at HKLX was investigated.
Typical patterns of scans are reported in figure 6 excerpted from Messrs IMM SOIL BHD preliminary report.
A series of nine (L1 to L9) longitudinal equally spaced scans covering 355 m in length and 1.0 in width
was carried out, and potential defective areas have been marked for a second stage investigation by
means of UPE (Ultrasound Pulse Velocity) and coring.
Note. L9 is the line closest to the tunnel walls.

Figure 6

When dielectric properties of the concrete change because of large voids filled with air or water, or
because of the presence of metallic reinforcement or foreign objects (plastic or wood for instance),
these features reflect electromagnetic waves, the GPR’s antenna detects these reflective surfaces and
the algorithms within the software used to elaborate data visualize these defects or anomalies with scale

TM0050A Page 5 of 13
of colours in an otherwise homogeneous background. The example of figure 7 below gathered from L9
(hence close to the tunnel walls) shows an approximately 2.5 m long (potential) interface between
stage 2 concrete and stage 1 concrete (turquoise reflections) resting at about 0.1 m below the rebars
cage (shown as bright red reflections). This zone has been marked for further investigations by means of
UPE as it can be read in the image itself. Since this reflection appears at the depth where concrete from
stage 2 and 1 join, it might be caused by a potential delamination which needs to be assessed both by
direct methods (coring) and by UPE tests.
Figure 7

Besides having located areas where at the interface between stage 2 and stage 1 concrete strong
reflections are present, other potential areas of defective concrete might be located within the depth
of the reinforcement as shown in figure 8 below. These may due to segregation or incorporation of
debris/foreign objects (to be verified by further UPE).

TM0050A Page 6 of 13
 Figure 8

When an image like the above is compared with an image gathered from a, so-called, well-supervised
section within the same project, such as the section represented in figure 9 below, the homogeneity of
concrete for a well conducted work is clear even to the untrained eye in the field of NDT.
In the image below it is in fact visible the only expected strong reflection given by the rebars at depth
of about 0.1 m only, over an otherwise homogenous background. Because no interface is visible
between stage 2 and stage 1 concrete at about 0.20-0.22 m it should be concluded that the two layers
are fully bonded and they are of similar properties.

TM0050A Page 7 of 13
 Figure 9

Interpretation of GPR data

For the sake of the interpretation and based on what it has been communicated to the undersigned
by the designer, the critical sections to be thoroughly investigated are those covering the first 300 mm
of width of the 1.0 m wide platform.
From GPR data it seems possible that four type of defects may be present in order of importance:
1. Local detachment at the interface between new and old concrete (below the rebars level);
2. Segregation caused by unsuitable consistency/mix type at placing (random);
3. Top-down cracking of high shrinkage mixes;

UPE Survey
The Ultrasound Pulse Echo technique uses sonic waves (shear waves precisely) travelling through the
thickness of the concrete, where features having different elastic modulus (i.e. different mechanical
properties) may cause reflections. This method is usually coupled with GPR to have a more
comprehensive view of the interior of the concrete by using two different “eyes”, one working with
electromagnetic waves and dielectric properties and the other working with elastic waves and
mechanical properties.

TM0050A Page 8 of 13
At the time of preparation of this preliminary report, UPE elaborations are not available yet. However,
raw data from an area where mortar like material was extracted (ref. to Core 5) show, for instance, a
shear wave velocity Vs = 1800 m/s.
From the laws of continuum mechanics (ref. for instance to the classic Timoshenko & Goodier’s Theory
of Elasticity), for a Poisson coefficient of concrete-like materials, of 0.15 a compressional velocity of Vp
= 2700 m/s can be calculated.
Using the following standard relationship between Vp and the dynamic modulus of elasticity E for a
concrete-like material of apparent density of about = 23.0 kN/m3

(1 − )
=
(1 + )(1 − 2 )

It is possible to estimate a value of dynamic elastic modulus of the area tested by UPE of about 16 GPa,
which is indicative of mortar-like materials, more than concrete.

Coring Survey
In areas that showed anomalies at the interface between stage 1 and stage 2 concretes, such as those
illustrated in figure 7, six 100 mm diameter cores were taken to include stage 2 and some of the stage 1
underlying concrete. Some pics of the coring campaign are reported in figure 10. The following
observations can be made which are, by the way, in remarkable agreement with the predictions made
by the GPR data.

 Core C1: Stage 1 concrete and stage 2 concrete are bonded, but the coring process has
broken an already weak interface;
 Core C2: Stage 1 concrete and stage 2 concrete are bonded, but the coring didn’t break the
interface which is therefore better bonded than for core C1;
 Core C3: Stage 1 concrete is cracked and it appears toward the top such as a mortar like
material. Stage 2 concrete is fully detached from casting since the pre-existing crack in stage 1
concrete did not propagate bottom-up into the upper layer as clearly visible;
 Core C4: Stage 1 concrete shows grout-like material towards the top. Stage 2 concrete is fully
detached in service (not damaged by coring).

TM0050A Page 9 of 13
 Figure 10

TM0050A Page 10 of 13
In addition, core C5 in figure 11, shows, again, mortar-like material at the top and it appears that
bonding pre-existed to coring, since the crack in stage 2 concrete has propagated most likely top-
down into stage 1 concrete.

Figure 11

Preliminary Commentary
The area investigated shows potential issues with local detachment between stage 1 and stage 2
concrete. Stage 1 seems to have some pre-existing cracks and stage 2 seems to have been subjected
to severe shrinkage. Also, some cracks appear to have been poorly repaired with grout like material
creating an interposed weak layer between stage 2 and stage 1.
The possible severe shrinkage caused by restrained contraction of stage 2 and the consequent top-
down cracking, may have caused stage 1 either to crack, when bonding between the two layers was
very good, or to detach when bonding was very poor. If bonding is good in fact, tensile strains induced
by shrinkage of stage 2 become tensile stresses at the interface of stage 1. Whereas, if bonding is bad
tensile strains induced by the shrinkage of stage 2 become pull-off stresses for the interface of stage 1.

TM0050A Page 11 of 13
Hence it seems that uncontrolled volumetric instability of stage 1 concrete caused either detachment
(when adhesion was low from the time of casting) or cracking (when adhesion was high from the time
of casting).
Concrete in areas were cores show mortar-like material has very low compressional velocities consistent
with a dynamic modulus of about 16 GPa, which is half of what it should be expected for a C40, cube
concrete mix, for instance. Reason for this might be found in segregating mixtures, sand-rich mixtures,
water-rich mixtures, cracking, etc.

Percentage of Defects in the Test Area

A preliminary counting of linear meters of defective or potentially defective concrete with the total
scanned liner meters shows the following (see annex 1):

 Defects seem to be more concentrated in the L1-L4 sections, which are also expected to
provide structural response in service;
 The most critical defects are, in our opinion, related to detachment caused by shrinkage
movements of stage 2 over stage 1;
 Also, in areas where concrete looks segregated/fine/mortar-like, the predicted mechanical
properties inferred by using the shear waves recorded with UPE are poor, indicative of low
strength/mortar like material;
 In some lines these defects reach also up to 40-50% of the entire scanned length.

Recommendations
The proposed GPR has shown with remarkable accuracy the locations of detachment, which were
clearly confirmed by coring. Also, thickness of stage 2 concrete has been predicted very accurately
with GPR. UPE data are still under elaboration, but the use of shear waves velocity shows that, when
concrete appears defective, also the mechanical properties, evaluated by means of the dynamic
modulus of elasticity, are poor.
In view of the findings of this preliminary campaign and considering that the GPR has proved to predict
very accurately the presence of defective concrete when it comes, specially, to
bonding/delamination, and, finally, considering that UPE (much slower technique) could be used,
randomly, to assess mechanical properties and refine further investigations when required, we propose
the following approach.

TM0050A Page 12 of 13
  Scan the entire 27 km length of the platform in one scan only of L2 type (located therefore
roughly midspan within the 300 mm wide strip of concrete of most interest for the mechanical
response of the element in service).
 Every 2.0 km test a 100 m2 area (100 m length and 1.0 m wide) with L1 to L9 scans and additionally
gather information by means of UPE of concrete mechanical properties by recording shear
waves.

This is expected to take about 35 worked days to completion, excluding reporting.

The L2 type of scanning will ensure that, if any detachment surface exist between stage 2 and stage 1
in proximity to the track, it will be crossed by the transect of the GPR.
For any evidence of detachment, the 100 m2 investigation will be prompted automatically. This can
obviously inflate the time for completion of works enormously but it cannot be predicted at this stage.

To sum-up:

 Best case scenario: scan 27 km with an L2 scan and, every 2.0 km, execute a 100m2 investigation
including up to L9 plus shear waves for mechanical properties.
 Worst case scenario: scan 27 km with an L2 scan and every time you have evidence of
detachment between stage 2 and stage 1 concrete expand the investigation to an area of
100m2 including up to L9, plus measure shear waves for mechanical properties.

TM0050A Page 13 of 13
ATTACHMENT 1
(Sketch of Testing Layout)
C001 C002
C004
C003