Development of Lifetime Bridge Management System For Expressway Bridges in Japan
Development of Lifetime Bridge Management System For Expressway Bridges in Japan
Development of Lifetime Bridge Management System For Expressway Bridges in Japan
2
Nippon Expressway Research Institute, Co., Ltd, Japan
3
Nippon Expressway Research Institute, Co., Ltd, Japan
ABSTRACT
These days, proper maintenance and management of deteriorating expressway bridges are becoming
serious issues in many countries worldwide. It is necessary to apply proper maintenance actions in
order to extend the service life of existing bridges and to minimize their lifetime maintenance &
rehabilitation costs. In order to make a proper long-term maintenance strategy, it is necessary to
evaluate current condition and predict future deterioration for existing bridge, choose the best
maintenance actions at appropriate application times, and determine their expected durations of effect.
Under these circumstances, recently privatized Japanese expressway companies have been trying to
develop a lifetime bridge management system (BMS) for intercity expressway bridges in Japan. The
main objective of this study is to develop a framework for an appropriate lifetime maintenance
strategies for deteriorating bridges with emphasis on concrete superstructures.
Recent study revealed that the most relevant mechanisms for deterioration of concrete structures are
corrosion of rebar due to chloride induced deterioration or carbonation. Condition evaluation method for
bridge elements under several deterioration factors such as chloride induced deterioration and
carbonation is introduced using quasi-quantitative grading method. The grading method is based on
the results of visual inspection.
One of the most important tools included in the bridge management system is deterioration prediction
model. The deterioration prediction curve was calibrated based on the further inspection results and
engineers’ judgment. Several maintenance scenarios are taken into account in order to compare the
life-cycle maintenance cost. An optimum lifetime maintenance strategy for expressway bridges is
proposed based on the comparison of present values of expected cumulative maintenance costs.
The proposed bridge management system can be used as a useful tool for decision makers to
determine the best maintenance strategy.
KEY WORDS
BMS, expressway bridges, deterioration prediction, lifetime maintenance strategy
1. INTRODUCTION:
These days, proper maintenance and management of deteriorating highway structures are becoming a
serious issue in many countries. It is necessary to apply proper maintenance actions to highway
bridges in order to extend the lifespan of existing bridges, and minimize the repair and rehabilitation
cost. Leading works on BMS in Japan include a study by Miyamoto et al [1]. The research uses the
results of regression analysis for statistically processed inspection data and neural network models to
evaluate the soundness of concrete beams and slabs using the average soundness for load bearing
capacity and durability obtained from the output of an expert system.
Managing infrastructure and rational decision-making is a complex activity requiring both
well-experienced engineering knowledge and economical considerations. It is bridge agency’s
responsibility to ensure bridge structures can cope with the increase in demand without loss of service
to the public under the constraints of shrinking budget and lack of experienced maintenance engineers.
Right now, intercity expressway systems in Japan are maintained and operated by recently privatized
expressway companies from former Japan Highway Public Corporation. One of the three expressway
companies, West Nippon Expressway Company Limited (NEXCO-West), is responsible for
maintenance and operation of more than 3,000 expressway bridges. These bridges have been in
service for 20 years on average. Many of them have been deteriorated due to heavy traffic and severe
environmental conditions, and some of them are in need of repair and/or reinforcement. However, so
far maintenance actions have primarily been conducted as corrective maintenance, with decisions
based mainly on the professional engineering judgment and budgetary restrictions within a short-term
basis.
Under these circumstances, Nippon Expressway Research Institute (NEXCO Research Institute) has
developed a total bridge management system (NEXCO-BMS) in 2003 in order to support finding the
best solutions for lifetime maintenance scenarios for deteriorating expressway bridges.
For most of the decision makers, the most difficult uncertainties associated with actions and decisions
were produced by the very absence of data where statistics could be of little help. In order to make
better decisions based on life-cycle concepts, better data is required, and the data must be turned into
information and put to effective use through advanced decision support. Therefore, the collected data
should be shared by everybody involved in the maintenance business in the highway agency so that
the decision-making process could be more effective. NEXCO has developed integrated maintenance
database such as inventory data, inspection records and repair/reinforcement history for over 3,000
bridges. That database could be of great help to predict future conditions of bridges by calibrating
deterioration prediction curve created by mathematical equations.
In general, deterioration of highway structures progresses at an increasing rate with time. Therefore,
total life-cycle cost could be reduced by applying preventive maintenance in early stage of deterioration
than postponing it until the deficiencies become evident. This is because deterioration of bridges
accelerates year by year [2].
Currently, the importance of performing preventive maintenance actions in order to reduce the life-cycle
cost and achieve a longer service life has been widely recognized. However, in reality, it is difficult to
allocate the limited budget to preventive maintenance actions due to the shortage of budget and
accumulating backlogs of essential maintenances. In addition to it, a strategy of preventive
maintenance may be more difficult to justify because the public’s expectation is that the worst roads
demand immediate attention. Furthermore, the public often interprets activities related to preventive
maintenance actions as “fixing something that isn’t broken” [3]. Life-cycle cost analysis (LCCA) is an
effective tool in order to prove the cost-effectiveness and return on investment of such a preventive
strategy relative to traditional strategies. LCCA is an evaluation technique applicable to the
consideration of certain transportation investment decisions [3]. Specifically, when it has been decided
that a project will be implemented, LCCA will assist in determining the best (the lowest-cost) way to
accomplish the project. The LCCA approach enables the total cost comparison of the competing
maintenance alternatives, each of which appropriate for implementation.
In order to perform LCCA, it is important to develop a system that can express the present condition of
structures, the effects of various maintenance actions, and the prediction of future conditions.
The main objective of this study is to develop a framework for determining the best maintenance
scenario for expressway bridges with emphasis on concrete superstructures. Based on the prediction
of future condition states for expressway bridges, various kinds of maintenance actions are applied
depending on the condition states. The distribution of the first application time of each maintenance
action is determined based on the condition transition time calculated by the deterioration prediction
model. To find the lifetime maintenance cost for each maintenance scenario, Monte Carlo simulation
was performed with the sample numbers of 10,000. Both zero and non-zero discount rates are used to
help understanding the effect of time value of money on results.
The optimum lifetime maintenance strategy for highway bridges should be determined based on the
comparison of present values of expected cumulative maintenance costs under different maintenance
scenarios. It is concluded that the preventive maintenance option is the most cost effective
maintenance scenario. In addition, iterative application of preventive maintenance can guarantee the
better condition states compared to the corrective maintenance options in which maintenance actions
are not applied until the onset of corrosion in rebar. The proposed framework can be useful for decision
makers to determine the best maintenance strategy.
2. CONDITION EVALUATION
2.1 Definition of condition state
There are several methods to evaluate the condition state of a concrete structure. It is desirable that the
assessment of various performances of a structure be conducted using quantitative indicators such as
chloride ion content in the concrete, carbonation depth, or corrosion of steel rebar. However, in practice,
obtaining all quantitative data for all structures is not economically efficient. Therefore, in most BMSs
the condition states are verbally described and the quasi-quantitative grading method is widely applied
to the evaluation of concrete structures.
This quasi-quantitative grading method is used also in NEXCO-BMS, since the visual inspection results
are recorded based upon the discrete condition states. General definition of each condition state in
NEXCO-BMS is defined as shown in Table 1 [4]. Each grading is related to a possible maintenance
strategy such as preventive maintenance, corrective maintenance with repair or rehabilitation. State V
represents the critical condition that means failure of an element. Fig. 1 graphically shows the visual
appearance of RC deck slab in State I to V in the inspection manual [5].
∂C ( x, t ) ⎛ ∂ 2 C ( x, t ) ⎞
= D⎜⎜ ⎟⎟ (1)
∂t ⎝ ∂x
2
⎠
⎛ x ⎞
C ( x, t ) = C0 ⎜1 − erf ⎟ + C ( x ,0 )
(2) ⎝ 2 D ⋅ t ⎠
where
C (x, t) = Chloride ion content at the depth of x (cm) and at time t (years) (kg/m3)
C0 = Chloride ion content on the concrete surface (kg/m3)
D = Apparent diffusion coefficient of chloride ions (cm2/year)
erf = error function
C (x, 0) = Chloride ion content at the depth of x (cm) and at time 0 (year) (kg/m3)
As the deterioration of a structural member progresses into state III (see Table 3), corrosion of rebar
would start and it becomes the dominant indicator of deterioration instead of chloride ion content.
Corrosion of rebar can be predicted by eq. (3) [9]:
3.8 × 10 −3
Y= × e 0.15t (3)
φ
where
Y = Corrosion of rebar (%) at time t (years)
t = Elapsed time after the chloride ion content reached 1.2kg/m3 at position of steel rebar
φ = diameter of the steel rebar (cm).
In general, most of the decisions during maintenance and management of bridge structures shall be
made under conditions of uncertainty. In this sense, bridge performance, its remaining life, and
expected lifetime maintenance cost can be predicted only in terms of a probabilistic way. Uncertainty is
naturally associated with random phenomena because the exact realization of a phenomenon cannot
be determined with certainty [10]. The conceivable or possible realizations may be described in terms
of a range of possibilities, with their respective relative likelihoods of occurrence (e.g., with a probability
density function). In other words, if the state of nature is basically random, it cannot be described with a
deterministic model; its description must include a measure of its inherent variability and thus
uncertainty. For practical purposes, the required description may have to be limited to the main
descriptors of interest, which are the central value (e.g., the mean or median) and its measure of
dispersion (e.g., standard deviation or coefficient of variation). Available observational data are
normally used to estimate the central value and the degree of dispersion of the possible realizations.
The probabilistic approach to predict the deterioration of bridge elements is most commonly carried out
by Markov chain method used in PONTIS BMS [11]. Current condition states of bridge elements are
investigated and Markov transition probabilities between different condition states are used to evaluate
and predict future performance of deteriorating structures. However, currently predominant Markov
based management systems cannot address the environmental and material properties of each
structure because they are not based on the physical behavior of the deteriorating structures.
In the deterioration prediction model described in this paper, eqs. (2) and (3) are used to perform a
probabilistic prediction of future deterioration for concrete bridges under chloride induced deterioration.
As mentioned previously, the corrosion of steel begins when the chloride concentration reaches the
critical value. Many experimental studies on existing bridges showed that the concentration level of
chloride ion at the position of rebar varies in a very broad range. For instance, the cracking due to
shrinkage and/or tensile loading condition affects the overall permeability of the concrete member and
the distribution of the chloride concentration with respect to time. In addition, perfect control of
water-cement ratio is very difficult in practice, mainly due to the fact that course aggregate and fine
aggregate contain certain amount of water and the moisture contents of aggregates vary with time,
depending on the temperature and the relative humidity of the environment [12]. Therefore, most of the
influencing variables exhibit significant randomness. The effect of uncertainties associated with the
influencing variables on chloride penetration increase as the life span of a structural system increases
[13].
Therefore, in this paper, the randomness of most variables is taken into account, and a probabilistic
prediction of deterioration for concrete bridges is performed based on eqs. (2) and (3). For problems
involving random variables with known (or assumed) probability distributions, Monte Carlo simulation is
required to predict the future events [1]. This involves repeating a simulation process, using in each
simulation a particular set of values of the random variables generated in accordance with the
corresponding probability distributions. By repeating the process, a sample of solutions, each
corresponding to a different set of values of the random variables, is obtained. A sample from a Monte
Carlo simulation is similar to a sample of experimental observations.
In this study, the Monte Carlo simulation method is used efficiently to evaluate the effect of propagation
of uncertainties on the distributions of these parameters. Based on the simulation results on
distributions of X and Y, the probability of transition of condition states can be predicted.
The input parameters used in the prediction of future condition states are tabulated in Table 3. All
random variables are assumed to have lognormal distribution. Each random variable is characterized
by its mean value μ, and standard deviation σ. Numbers in parentheses in Table 3 represent mean
values and standard deviations associated with lognormal distribution.
Calculating the time required for the chloride ion content to up-cross a given threshold level in Table 2
gives the time required for the condition state of the structure to transit to the next condition state. The
probability of the event P [X (t) ≥ XCRITICAL] can be evaluated by Monte Carlo simulation as follows:
m
P[ X (t ) ≥ X CRITICAL ] ≅ (4)
M
where: m and M are the number of samples satisfying the condition X(t) ≥ XCRITICAL and sample size of
the simulation, respectively. The probability of (4) is computed at specific points in time. The period of
staying in state I and II (i.e., before the onset of corrosion) can last for many years depending on the
corrosion resistance of the steel, the thickness and quality of concrete cover, and other corrosion
protection measures applied to the structure [12].
Similarly, probability of corrosion of rebar to up-cross a given threshold level can be computed in the
same manner.
Fig. 2 shows the result of Monte Carlo simulation for chloride ion content at the position of rebar with
respect to time. Based on these values, corrosion of rebar will occur around t = 20 years in average. Fig.
3 shows the result of Monte Carlo simulation for corrosion of rebar with respect to time. Relatively large
standard deviation in corrosion of rebar Y is attributed to the large scatter in corrosion initiation time.
Table 4 shows the probability of being in each condition state with respect to time. The bridge agency
that owns a group of bridges wants to know when the first bridges are eligible for repair, in order to
allocate sufficient funds. Therefore, finding the distribution of the transition time of condition states can
be useful information. If a bridge owner is responsible for a group of structures whose material
performance and environmental conditions are similar to one another, the information shown in Table 4
can be used to predict the number of bridges that need to be repaired or rehabilitated at a given point in
time.
POSITION OF REINFORCEMENT, X (KG/M )
1.6
3
CHLORIDE ION CONCENTRATION AT THE
10,000 SAMPLES
1.4 STATE III
1.2
1.0 STATE II
MEAN
0.8 W/C = LN(50%, 5%)
3
C0 = LN(2.0, 0.2), KG/M
0.6
HCOVER = LN(3.5, 0.35), CM
0.4
STD. DEVIATION
STATE I
0.2
0.0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
TIME, T (YEARS)
Time, t (Yrs) State I (%) State II (%) State III (%) State IV (%) State V (%)
10,000 SAMPLES
700000
500000
MAJOR CONC. REP.
400000
300000
DESALINATION LINING A
200000 LINING B
100000
0
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
TIME, T (YEARS)
75000
10,000 SAMPLES
REBUILD
MEAN CUMULATIVE COST (U)
v = 6%
50000
25000
LINING B
LINING A
0
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
TIME, T (YEARS)
9.0
8.0
約3.0
7.0
co (kg/m3)
6.0
5.0
4.0
約2.0
3.0
約1.0
2.0
約1.0
1.0
約1.0
0.0
0m
汀 線 100m 250m 500m 1000m
Fig. 6 Relationship between C0 and distance from coast [16]
1 1
2 1
1
1
0 10 20 30 40 50
Year
Fig. 7 An example of regression analysis for deterioration prediction [15]
2. A probabilistic approach to predict the future condition states of the structure was discussed. Monte
Carlo simulation method is used efficiently to evaluate the effect of propagation of uncertainties on the
future condition states of the structures.
3. A methodology to determine the optimum maintenance strategies were discussed based on the
comparison of cumulative lifetime maintenance cost under different maintenance scenarios.
4. From the present value of the expected cumulative maintenance cost under different maintenance
scenario, it is concluded that the preventive maintenance strategy applying lining A is the most cost
effective maintenance scenario both in the case study for chloride induced deterioration and
carbonation. In addition, iterative application of preventive maintenance can also guarantee the better
condition states compared to the corrective maintenance options in which maintenance actions are not
applied until the onset of the corrosion in rebar.
5. The results obtained in this study illustrates that the most attractive maintenance scenario could be
the one which frequently applying the preventive maintenance action with lowest unit cost before it is
too late. This kind of analysis can be of great help to prove that the preventive maintenance scenario
actions could be a better option than corrective maintenance actions to minimize the expected
cumulative life-cycle cost.
This conclusion is in good agreements with those demonstrated by Frangopol et al. [17] and Kong and
Frangopol [13].
REFFERENCES
[1] Miyamoto A., Kushida M., Adachi Y., and Matsumoto M., (1997), “Development of Bridge
Management System (BMS)”, Proceeding of the Japan Society of Civil Engineers (JSCE), No. 560,
VI-34, pp. 91-106
[2] Nakatani , S. et al. (2004), “Current Status of Japanese Road Bridges and Future Perspectives”,
October 2004, 2nd International Conference on Bridge Maintenance, Safety and Management, Kyoto,
Japan
[3] Department of Transportation, Office of Asset Management (2002), “Life-Cycle Cost Analysis
Primer”, Washington DC
[4] JH-BMS Committee (2003) “Report on Committee for deterioration prediction and evaluation of
bridges”, February 2003, Tokyo, Japan
[5] JH (2003), Japan Highway Public Corporation, “Inspection Manual for Highway Structures”, Tokyo,
Japan
[6] Yokoyama et al (2001), “Development of Bridge Management System taking account of the Life
Cycle Cost”, The 56th Annual Conference of JSCE, CS6-003, 260-261
[7] JSCE (2001), “Standard Specification for Concrete Structures -2001 [Maintenance], Working Group
on Standard Specification for concrete Structures, JSCE, Tokyo, Japan
[8] Matsushima, M. (2001), “Optimum Repair Level Considering LCC of RC Structures”, Maintaining the
Safety of Deteriorating Civil Infrastructures, 259-271, Miyamoto A. & Frangopol D. M. Eds., September
2001, Practical Maintenance Engineering Institute of Yamaguchi University, Yamaguchi, Japan
[9] JCI Rehabilitation Committee (1998), "Committee Report on Rehabilitation of Concrete Structures",
36-52, October 1998, Japan Concrete Institute, Tokyo, Japan
[10] Ang, A. H-S. and Tang, W. H. (1990), Probability Concepts in Engineering Planning and Design,
Volume II – Decision, Risk, and Reliability, John Wiley & Sons, New York
[11] Thompson, P. S., Small, E. P., Johnson M., and Marshall, A. R. (1998), “The PONTIS Bridge
Management System”, Structural Engineering International, IABSE, 8(4), 303-308
[12] Kong, J.S., Ababneh, A.N., Frangopol, D.M., and Xi, Y. (2002), “Reliability Analysis of Chloride
Penetration in Saturated Concrete, Probabilistic Engineering Mechanics 17, 305-315, Elsevier
[13] Kong, J.S. and Frangopol, D.M. (2003). “Life-cycle reliability-based maintenance cost optimization
of deteriorating structures with emphasis on bridges,” Journal of Structural Engineering, ASCE, 129(6),
818-828.
[14] Matsumoto M. (2005), “Development of Lifetime Maintenance Strategies For Highway Structures”,
Master’s Thesis for University of Colorado at Boulder, 2005, CO, USA
[15] Yokoyama et at (2007), “Development of Bridge Management System for Expressway Bridges in
Japan”, The Second International Workshop on Lifetime Engineering of Civil Infrastructure, Nov.28-30,
2007, Yamaguchi, Japan
[16] Wada Y., Sakai S., Ohshiro T., Homma A. and Ogata N., “The measure towards advanced of
Bridge Management System for Expressway Bridges in Japan” The Fourth International Conference on
Bridge Maintenance, Safety and Management (in Press), Korea, 2008
[17] Frangopol, D.M., Kong, J.S., & Gharaibeh, E.S. (2001). “Reliability-based life-cycle management
of highway bridges,” Journal of Computing in Civil Engineering, ASCE 15(1), 27-34