ANALYSIS OF FLOOD AFFECTED ROAD PRESERVATION PROGRAM USING THE MARKOV CHAIN MODEL

• Main Author: Bimo Aulia, Yoga
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/72040
• Institution: Institut Teknologi Bandung

Roads are one of the most important supports in the development of a region. In order for roads to function properly, it is necessary to manage roads properly and correctly as part of road asset management. In recent years, floods have disrupted the road management system by significantly increasing rehabilitation costs. The intensity and frequency of the flooding that occurs is uncertain, for example Bengkulu which was hit by floods in 2019 was also hit by another major flood after high intensity rains in 2021 and 2022. Pavement performance modeling is an important element of a pavement management system. Nearly 95% of the service level of road pavement is related to surface roughness, in other words surface roughness is the main variable that greatly affects the serviceability of road pavement. In this study, an analysis of the flooding’s effect, traffic, and maintenance history on the deterioration of pavement conditions in the future was carried out. The analysis was carried out on all national road sections in Bengkulu province. Road sections are categorized as segments affected by flooding and segments not affected by flooding according to the traffic category and maintenance history. The pavement condition parameters used in this study are the IRI (International Roughness Index) and PCI (Pavement Condition Index) values. Analysis of changes in pavement conditions in the future using the Markov Chain model. Later a comparison will be made between determining pavement conditions using random numbers from the LCG (Linear Congruential Generator) and those with the highest probability. In addition, the results of future pavement conditions with Markov Chain will be compared to the results of pavement conditions according to IRMS V.3. The determination of the treatment that will be carried out refer to the IRMS V.3 flexible pavement decision tree. Based on statistical analysis, it can be concluded that if handling is carried out, the flooded segment and the segment not affected by floods have no difference in pavement condition values. Meanwhile, if no action is taken, the flooded segment and the segment not affected by floods will have different pavement conditions. Based on the resulting transition probability matrix (MPT), traffic category conditions affect the level of decrease or increase in pavement conditions where high traffic has a higher rate of decline in pavement conditions than medium traffic and low traffic. The conditions between the flooded and non-flooded segments are also different in terms of changes in pavement conditions where the flooded segment has a faster rate of decline than the non-flooded segment. In the analysis of changes in pavement conditions in the future, the comparison of the number of handling decisions given to Markov Chain using random numbers and Markov Chain with high probability has little difference. Likewise, if the Markov Chain has the highest probability compared to IRMS V.3, it also shows different results, where the deterioration of pavement conditions, both IRI and PCI, is faster so that more frequent handling is carried out than using IRMS V.3, which tends to decrease road conditions more slowly.