Abstract:
The concept of road maintenance management through the
life cycle of roads by analyzing deterioration behavior and
predicting the future service quality of roads is an essential
component of maintenance planning. From a study using the
HDM-4 deterioration model, it was found that the model was
overly complex, making it impractical for practical application.
This research aims to develop a suitable model for predicting the
deterioration behavior of asphalt concrete pavement roads,
aligning with the service provision characteristics and user
behavior in Thailand. The research employs support vector
machine (SVM) for regression analysis in developing the
deterioration model. The analysis includes the average annual
daily traffic (AADT) and International Roughness Index (IRI) data
for National Highway No. 1 (1,396 samples), National Highway No.
2 (898 samples), National Highway No. 3 (703 samples), and
National Highway No. 4 (566 samples) surveyed between 2019
and 2023. The evaluation of the developed model yielded MAPE
= 10.92%, RMSE = 0.4100, and R-Squared = 0.8747. When
applying the model to predict the service life of National Highway
No. 1, Sutheewittaya School Junction - Dong Jampa, Saraburi
province, when setting the target service level at an average IRI
of 3.50 m/km, as per government service performance
certification, it was observed that segments with service levels
not lower than the specified target are expected to remain in
service from 2024 to 2028. The analysis results can be utilized for
planning purposes, determining when each road segment should
be maintained to enhance the efficiency of road maintenance
management within the constraints of available budgets