A Hybrid Stochastic/Mechanistic Model for Determining Bridge Condition Rating

Bridges like other infrastructures such as roads and dams require proper inspection and maintenance programs to reserve bridge structural integrity and to ensure public safety. The structural integrity decreases through time due to several factors such as material degradation, environmental attacks, and an increase in usage over time. The consequence of improper inspection and maintenance programs can cause catastrophic bridge failure.

Dr. Rajwanlop Kumpoopong made a research which focused on developing a hybrid deterioration model that combines the quantitative analysis of damage mechanisms and the stochastic process based on the concept of Markov chain. The proposed model enhances the existing Markov deterioration model in a more reliable condition prediction of bridge elements at the network-level since relevant variables which affect the deterioration of bridge elements are properly addressed. Only the hybrid deterioration model for bridge deck is developed in this research.
The main objective of this research was to develop the methodology for developing the hybrid deterioration model for bridge deck that combines the quantitative analysis of crack mechanism due to truck traffic loads and the stochastic process based on the concept of Markov chain for the network-level bridge management. The methodology is then applied to develop the hybrid deterioration model for bridge deck under truck traffic condition in Thailand as a case study. The secondary objectives of this research were to: (1) study the effect of each truck category on the deterioration of bridge deck; (2) study the effect of upgrading regulatory truck weights on the deterioration of bridge deck; and (3) study the effect of overloaded truck for the transportation of sugarcane on the deterioration of bridge deck.

Conclusions

Bridge deck is the major bridge component that requires significant maintenance effort. The factor that contributes most to the deterioration process of bridge decks is concrete cracking which is directly related to the traffic volume and can be accelerated by increasing truck traffic. Bridge deck deterioration models in most recent bridge management systems are expressed in the form of probabilistic deterioration models such as the Markovian process. The transition probability matrix is developed to capture the deterioration mechanism of bridge decks.

The strength of the Markov deterioration model is its practicality when it is applied to forecast at the network-level bridge management. The condition of the bridge deck over time can simply be predicted through the multiplications of the initial state vector and the transition probability matrix. However, the Markov deterioration model has some weaknesses since the model may not reflect the actual condition of bridge decks. In the Markov chain models, relevant variables which affect the deterioration process of bridge decks, such as changes in traffic volume and environmental conditions are assumed to be constant throughout the analysis period. This may not be realistic and can result in the overestimation or underestimation of the forecasted bridge condition.

This research work aims to develop the methodology for developing the hybrid deterioration model for reinforced concrete (RC) bridge deck that combines the quantitative analysis of crack mechanism due to truck traffic loads and the stochastic process based on the concept of Markov chain for the network-level bridge management.

In this research, the finite element program that can simulate cracks in RC bridge deck due to truck traffic is developed to capture the crack mechanism. The results are presented by S-N relations of RC bridge deck at different levels of damage. The damage is defined in this study as the extent of crack on the bottom surface of the deck slab or the crack level. Distribution curves of the number of trucks required for the crack to reach each crack level are then developed from the simulated S-N relations and the truck weight information. From these distribution curves and the specified truck traffic volume and composition, transition probabilities based on the concept of Markov chain can be calculated. Based on the proposed technique, transition probabilities can be calculated for truck traffic data that is non-homogeneous with respect to both composition and volume which provides the flexibility of the model to be applied to highway bridges of different truck traffic conditions. The effect of specific overloaded trucks on the deterioration of DOH bridge deck is discussed through the case study of the transportation of sugarcane. For practical assessment of the deterioration of DOH bridge deck, a series of distribution curves of the number of trucks required for the crack to reach different crack levels for different truck categories and gross weights is provided.

His dissertation abstract is copied and posted below.

ABSTRACT

The concept of Markov chain has been widely used to predict the condition of bridge deck. The transition probability matrices, which are mostly developed by inspection data and experts’ opinion, are used to capture the deterioration of bridge deck. The Markov deterioration model developed by this approach although states deterioration in probabilistic sense may not be reliable because relevant variables which affect the deterioration of bridge deck such as changes in traffic volume and environmental conditions are not explicitly emphasized in the model. Concrete cracking which can be accelerated by increased truck traffic is the factor that contributes most to the deterioration of bridge deck and should be addressed in the deterioration model. This research work aims to develop the methodology for developing the hybrid deterioration model for reinforced concrete (RC) bridge deck that combines the quantitative analysis of crack mechanism due to truck traffic loads and the stochastic process based on the concept of Markov chain for the network-level bridge management. In this research, the finite element program that can simulate cracks in RC bridge deck due to truck traffic is developed to capture the crack mechanism. The results are presented by S-N relations of RC bridge deck at different levels of damage. The damage is defined in this study as the extent of crack on the bottom surface of the deck slab or the crack level. Distribution curves of the number of trucks required for the crack to reach each crack level are then developed from the simulated S-N relations and the truck weight information. From these distribution curves and the specified truck traffic volume and composition, transition probabilities based on the concept of Markov chain can be calculated. Based on the proposed technique, transition probabilities can be calculated for truck traffic data that is non-homogeneous with respect to both composition and volume which provides the flexibility of the model to be applied to highway bridges of different truck traffic conditions. The methodology is presented through the development of hybrid deterioration model for Thailand Department of Highways (DOH) bridge deck. In the presentation, the effect of each truck category and the effect of upgrading regulatory truck weights on the deterioration of DOH bridge deck are discussed. The effect of specific overloaded trucks on the deterioration of DOH bridge deck is discussed through the case study of the transportation of sugarcane. For practical assessment of the deterioration of DOH bridge deck, a series of distribution curves of the number of trucks required for the crack to reach different crack levels for different truck categories and gross weights is provided.