Prediction and quantification models of bridge condition in Korea

In Bridge Management Systems (BMS), the role of condition profiles is important to predict the condition in the future subject to diverse circumstances. The accuracy of the profiles is directly related to the accuracy of BMS and the reliability of the profiles is important to produce the optimal solution to distribute a maintenance budget reasonably. To predict the future condition of bridges, condition rating models based on experts' opinions have been used widely because of the lack of quantitative inspection and maintenance records. A weakness of these conventional systems is that experts' opinions are very subjective, moreover, the quality of condition profile is not good enough to be used for more advanced BMS. To overcome this limitation, an advanced regression technique is suggested to make reasonable condition profiles. To apply the suggested technique, the database of the Highway Bridge Management System (HBMS) in Korea is used and condition profiles are produced.