교량 모니터링 빅데이터를 이용한 광안대교의 교통량 의존 변위 추정 모델

Abstract

In this study, we developed a vertical displacement prediction model that relies on the real-time traffic volume of the Gwangan Bridge. The influence of large vehicles and small vehicles on the structural displacement of bridges is confirmed and the actual traffic volume in the bridge is negatively affected by fatigue and aging of bridges. And to clarify the effect. Based on the developed model, we developed a model that can predict the strain and expansion / contraction, and tried to find ways to utilize the billing system in order to minimize the impact on the structural displacement of the bridge by distributing traffic volume. In order to use all of the variables in the modeling process of displacement prediction depending on the traffic volume, structuring was carried out. In this process, a correlation regression based regression mode and a principal component regression model Respectively. There was no significant difference in the RMSE of each of the two models. The developed model was completed with training data for 2006. However, it was tried to verify whether the developed model can be applied even at the present time 10 years later. An independent sample t-test was conducted using the traffic volume by vehicle type in 2006 and 2017 as variables. As a result of verification, small - sized vehicles increased and large - sized vehicles did not increase. For the developed model, we compare the predicted value with the actual value by applying the validation data set that has passed 10 years. The validation result shows that the structured regression model has a wider deviation than the principal component regression model. The independent sample t-test was conducted using the mean difference of the residuals of the development model, and the structured regression model required an update. Three updated models were developed by combining various data of 2006 and 2017. The developed updater model was tested with RMSE. As a result, the structural regression model shows a large deviation in RMSE and MAPE. In the use of developed models, a model was developed to predict strain using structured independent variables in the same way. And by applying the development model that revealed that the vertical displacement was displacement due to traffic load, the vertical displacement value acting on the new axis displacement of the bridge was applied to the live load value to develop the expansion displacement prediction model. Finally, since this traffic affects the aging of the bridge, the distribution of traffic can be made. These proposals presuppose an increase in the maintenance budget for a preemptive response to longevity, such as repair and improvement of performance that reflects the level of needs of citizens, in a situation where bridges are aging over time. However, considering the fact that the proportion of domestic SOC budget including maintenance budget is decreasing every year and that Gwangan Bridge is scheduled to be free in 2028, the government should supplement the insufficient budget by improving the operation method. Improved operating methods will contribute to longevity as well as reduced bridge fatigue.