75
International RILEM Conference on Materials, Systems and Structures in Civil Engineering
Conference segment on Service Life of Cement-Based Materials and Structures
22-24 August 2016, Technical University of Denmark, Lyngby, Denmark
coefficients were measured on concrete specimens under standard curing conditions (20°C
and RH > 95%) for given ages, 28d and 56d, by RCM method.
For tunnel segments, the target mean values for D
nssm
are respectively 6.5×10
-12
m
2
/s and
4.5×10
-12
m
2
/s for 28d and 56d. The mean values of measured D
nssm
are 4.68×10
-12
m
2
/s (28d)
and 2.95×10
-12
m
2
/s (56d), much lower than the required values. The variation coefficient of
this D
nssm
is 9.9% (28d) and 8.7% (56d), smaller than the expected level of 20% in design
phase, especially for 56d age. For concretes in bridge structures, much lower mean values,
compared to the required values, are obtained for both prefabricated and cast-in-place
elements. The D
nssm
dispersion of prefabricated bridge elements is very near 20% for both
28d and 56d ages, but the variance ranges from 4% to 44% for cast-in-place elements. Most
diffusivity values conform to the complaince critera (6) and (7).
3.3 Concrete cover thickness
The concrete cover was measured through the common midpoint (CMP) method and the
principle of CMP method detects the depth of reinforcement bars through transmitting an
electromagnetic wave pulse and receiving the reflected waves from the steel bars by two
adjacent antennas. The measurement error of the device used in HZM project is evaluated as
around 3mm through the comparison with the measurement on extracted cores from test
segments of tunnel. For information, the compliance criteria for the construction errors of
concrete cover are within -5mm/+18mm.
For the prefabricated tunnel segments, the required values for thickness are respectively
50mm (intrados) and 70mm (extrados). The measurements show that the mean values are
53.5mm (intrados) and 73.4(extrados) and the construction error, corresponding to the
distance between 5% percentile value and mean value, is 5.8mm (intrados) and 6.4mm
(extrados), slightly larger than the expected construction tolerance, 5.0mm, for prefabricated
elements. For the cast-in-place concrete elements in bridge structures, the design value for
concrete cover thickness for bridge decks is 45mm, and the measured mean value attains
50.5mm with only 1.11mm as standard deviation (SD) value; the design thickness for piers, in
splashing zones, is 70mm and the measured values have 80.0mm as mean value and 6.74mm
as SD value. The design values for bearing platforms, in immerged and splashing zones, are
respectively 60mm and 80mm, and their measured values have mean values of 71.0mm and
95.5mm and SD values of 6.85mm and 4.22mm.
4. Durability
assessment
4.1 Assessment model
The assessment model for chloride-induced corrosion is also adapted from the analytical
model of Fick’s second law, similar to the design model in (1). With the corrosion initiation
specified as DLS, the assessment equation writes,
cr
0
0
0
Cl
0
1 erf
0
2
,
d
s
x
G
C
C
C
C
D
t t t
(8)
76
International RILEM Conference on Materials, Systems and Structures in Civil Engineering
Conference segment on Service Life of Cement-Based Materials and Structures
22-24 August 2016, Technical University of Denmark, Lyngby, Denmark
with C
0
standing for the initial chloride concentration in concrete (%binder). Hereafter, this
model is to be used through full probabilistic approach to evaluate the failure probability of
durability,
f
SL
SL
prob
0
p t
G t
(9)
The service life is judged to be satisfied or not by comparing the failure probability at t
SL
with
the target failure probability
p
target
. Thus, the assessment model contains six parameters: C
cr
,
C
s
, C
0
, x
d
, D
Cl
0
and n(or ). The statistical properties of the parameters C
cr,s
and ageing
exponent n have been investigated in depth on the basis of the long-term exposure tests in
design phase [7], and the same statistical properties are retained in durability assessment. For
the initial concentration C
0
, a rectangle, or uniform, distribution is calibrated for the structural
concretes from the chemical analysis of raw materials. For the chloride diffusivity
D
Cl
0
, the
collected data are used to update the statistical properties, and the collected data are also used
to update the statistical properties of concrete cover thickness.
In the construction phase, different protection measures are adopted, including the silane
impregnation on concrete surface, epoxy-coating on steel bars,
stainless steel bars, coating on
steel surface and cathode protection of steel bars in concrete elements. These measures are
adopted to increase the durability safety margin and/or to compensate the possible adverse
effect from construction defects. In durability assessment, these measures are taken into
account after the following assumptions: (1) the silane impregnation is assumed to delay the
accumulation of chloride on concrete surface to its stabilized value by 10 years in the
assessment; (2) the epoxy-coatings have no impact on the corrosion initiation of
reinforcement steel bars in the durability assessment; (3) the C
cr
of stainless bars is
conservatively taken as 5 times the value of the conventional carbon steel bars; (4) the
cathodic protection passivates completely the reinforcement steel once it is put into function,
keeping the steel in passivation during its active duration.
4.2 Assessment through full probabilistic approach
The fully probabilistic analysis for the durability assessment of RC/PC elements is realized
through Monte-Carlo simulations. A computer-based program is developed especially to
perform the probabilistic assessment. In the simulation, six parameters are considered as joint
occurrence random variables. For a given exposure age, the Monte-Carlo simulations are
performed to calculate the failure probability of (9), and 1,000,000 samplings are used to
ensure the solution of “real” probability. Accordingly, the failure probability is solved with
time from t=0 to t=120 years. Some assessment results are presented in Figure 2 for bridge
elements in splashing and tidal zones.
Globally the concrete elements all achieve reliability index above 2.0 at 120 years. The bridge
elements in atmospheric zones have very low failure probability, i.e.
p
f
<10
-3
at 120 years and
the corresponding reliability index >3.0. The silane impregnation is generally adopted as
protection measure for elements, the stainless bars are adopted for pylons and the epoxy-
coated bars are used respectively for decks and box girders. The bridge elements in
splashing/tidal zone, piers and bearing platforms, have satisfactory durability safety margin: