Keywords

1 Introduction

Recently, interest in preventive maintenance of bridges and infrastructure has increased, and research has been conducted on the performance of structures and maintenance costs [1,2,3,−4]. After collecting the safety inspection and in-depth inspection data of the national road bridges, the condition prediction model was established through polynomial regression analysis considering the deterioration rates [5]. In order to make these studies practical, optimization algorithms were proposed to take account of the various constraints and correlations between members.

2 The Optimal Maintenance Algorithm for Time and Costs

To optimize time point and cost for maintenance are required to satisfy correlation in each member and various constraints. And then, a genetic algorithm has some effective advantages to estimate various variables and constraints. So, in this study, there is suggested the model to predict optimal time point and cost for bridge maintenance, using a genetic algorithm(G,A).

2.1 The Prediction Model of Bridge Member Condition and the Maintenance Cost

The prediction model of bridge member condition is made to forecast each member condition rate up to the target life of the bridge. And this model is based on safety inspection and in-depth inspection data of the national road bridges in Korea since 1995. In Fig. 1, the condition prediction model was developed by polynomial regression results which were based on bridge condition ratings, and the difference in the degree of deterioration expressed the deterioration ratio in the point of maintenance before and before after.

Fig. 1
figure 1

Concept about prediction model of bridge member condition and maintenance cost

Full size image

The rating maintenance cost function was proposed by specialists’ opinions and maintenance rating cost data, and it was shown that the curve was newly updated by rating resulting from maintenance.

2.2 Optimization Model Using the Genetic Algorithm

The genetic algorithm (G.A.) is suitable as maintenance optimization methods because they do not have to be constrained by differential and continuous conditions in the purpose function. So, in this study, it was applied to. Figure 2 is described in the genetic algorithm for the absence-specific correlation and constraints.

Fig. 2
figure 2

Optimal algorithm for bridge maintenance about time point and cost using G.A.

Full size image

Each Member of the bridge is needed to consider the correlation with the increased deterioration ratio. When it was reached condition-index Grade D(0.3), it has to replace as the maintenance and resets as the initial deterioration ratio(a). At this time, it should be considered some relationship with each member and decided to main or sub-member. Then, it is calculated the cost of bridge maintenance using relationship about bridge members.

3 Case Study of Optimal Algorithm About Maintenance Time Point and Cost

In this study, to confirm the applicated performance ability about the optimal algorithm, it conducts a case study using by deck and pavement members of the common bridge in Korea. As shown from the Table 1, it is represented factors - the deterioration ratio, deterioration increase ratio, C.I.- Cost function, and constraints. Each factor is set into the algorithm and calculates the optimal time point and cost of maintenance. And then, it makes the maintenance scenarios with the member-specific correlation. The deck and pavement maintenance scenario of Fig. 3 is represented the optimal maintenance scenario considering the member-specific correlation and scale for the target life of 100 years.

Table 1 The factors of optimal algorithm about maintenance using G.A.
Full size table
Fig. 3
figure 3

The scenario study about bridge maintenance

Full size image

4 Conclusions

In this study, to optimize time point and cost for maintenance, there is suggested the model to predict optimal time point and cost for bridge maintenance. Then, it is represented the optimal algorithm about maintenance of bridge using and genetic algorithm.

This algorithm can establish an optimal maintenance scenario considering correlation about each member of bridge to maintenance efficiently and reasonably. In addition, it is expected to be able to further apply a variety of member-specific correlations.