Traffic flow characteristics and modeling for estimation fo AADT in selected rural highways

Abstract

This study is performed to analyze traffic flow characteristics and estimation of AADT from short counts on selected rural highway. The North Bengal Corridor has been selected for this purpose based on availability of 7 years of daily flow data and 13 weeks of hourly data on Jamuna Multipurpose Bridge and I year of daily flow data on Nalka- Hatikamrul-Bonpara road. From the analyses of traffic flow data, it has been found that the basic flow patterns on the selected route are repetitive in nature. Analyses such as hourly, daily, weekly, seasonal variations, directional distribution, traffic composition, traffic growth pattern have been performed. It was observed that the normal flow pattern is sometimes affected by certain external factors such as Eid festivals, national strikes, transport strikes, flood etc. Considering the critical flow value caused by Eid festivals, emphasis has been given in analyzing the impact of Eid festivals on traffic flow. The AADT of 1999 and 2005 has been found to be 2416 and 5288 respectively. Taking the average from 1999 to 2005, the AADT is 3749. Growth rate of total traffic is 14% per annum. From flow pattern analyses, it has been found that maximum and minimum hourly flow occurs at I :00-2:00 and 5:00-6:00 carrying 5.68% and 2.25% of total daily traffic respectively. Maximum daily flow occurs on Friday (East to West 15.16% and West to East 15.25%) and December carries maximum monthly volume (9.45% of yearly volume). Dry season (Nov. to Apr.) carries slightly higher percentage of traffic (51.48%) than rainy season (May to Oct.) Average directional distribution is almost equal, but it varies significantly in hourly pattern. In all the cases, individual vehicle class shows different pattern from total traffic. Therefore, class-wise separate analyses have been performed and factors have been established. From traffic composition analysis, it has been observed that, trucks and buses together comprise of 84.06% of traffic stream. For estimation of AADT from short counts, expansion factors as well as regression models have been established. Their accuracy has been found to be around 80%, by checking with external data. Calibration curves from regression analysis have also been established.