Traffic speed along a roadway segment is the function of a number of factors like geometry, traffic and weather conditions. In this paper, the effect of weather conditions both the speed-flow and flow-occupancy relationships was studied. The data used in the analysis were obtained from traffic management system installed in the Incheon International Airport Expressway. In order to quantify how seriously weather affects traffic flow conditions, the impact in term of traffic speed was described. Regression analyses were performed to select proper models representing the speed-flow and flow-occupancy relationship for uncongested operation. From the research, weather conditions reduced the slope of flow-occupancy function and caused a downward shift in the speed-flow function. That is, the free flow speeds are reduced. Increase of traffic volume by weather event at the YeongJong Grand Bridge (site2) is much larger than at the BwangHwa Grand Bridge (sitel), which makes travel speed down significantly. The ratio of free flow speed reduction was observed 7% and 2% by snowy and rain, respectively. Meanwhile, the ratio of speed reduction was shown that snowy and rainy night are each 5% and 6% drop. That is, the effect of snowy day and night is similar while the effect of rainy night is larger than rainy day in terms of speed redution In case of site 2, during day time, the ratio of speed reduction was turned out no difference according to weather events like snowy and rainy conditions. However, the ratio of speed reduction during night time was decreased 3% and 7% in each snowy and rainy conditions.

An automatic traffic recorder (ATR) makes it easy to detect a vehicle’s speed and the number of axles because it produces highly precise data; thus, it can be used to obtain major traffic characteristics for several purposes such as vehicle classification. However, the failure of a piezo sensor or inductive-loop sensor often results in missing data. Moreover, there is no method available for evaluating the life of a sensor installed in a pavement. Therefore, such sensors require maintenance after a failure occurs, resulting in a large amount of missing traffic volume data. If the quantified life expectancy of a sensor could be obtained, a maintenance and rehabilitation strategy could be established and the economic feasibility could be determined. In this study, the 5-year maintenance history of a sensor newly installed in a national highway was reviewed. Normal, gamma, lognormal, and Weibull distributions were fitted to the data, and the results were compared. The Weibull distribution was used to predict the life. When the maintenance histories of piezo and inductive-loop sensors were fitted to the Weibull distribution, the scale parameters were 79.23 and 89.72 and the shape parameters were 3.28 and 2.83 for the piezo sensor and loop sensor, respectively. The mean life and median life of the piezo sensor were calculated to be 71.05 and 70.85 months, respectively. Those of the loop sensor were 79.92 and 78.82 months, respectively.