Calibration of delay formulas for saturated and unsaturated signalized intersections in Istanbul

Abstract

In order to minimize delay incurred at signalized intersections and provide the intended improvements, the calculation of delay should have good correlation with the actual delay times. This study is aimed to provide calibration for various delay formulas in order to adapt them to the local conditions in Istanbul. Delay formulas of Highway Capacity Manual (HCM) (1997 and 2000) and Percentile Delay Method (PDM) were tested for validity at unsaturated signalized intersections. These formulas and Akcelik’s overflow delay formula were studied for an oversaturated intersection. The delay estimates were compared with field measurements and calibration models were generated for those formulas. Signal 97, Signal 2000 and Synchro 6 softwares were used for analysis of intersections and MS Excel was used for statistical analysis. As a result of the analysis of unsaturated intersections, HCM 1997 delay formula underestimated the delay compared to the actual delay measurements. HCM 2000 delay formula overestimated the delay through analysis by Signal 2000 and Synchro softwares. Both HCM 1997 and HCM 2000 delay estimates resulted with good relationship with the field measurements with coefficient of determination (R2) values around 0.85. The PDM has resulted with a poor relation with R2 value of 0.51. The analyses of the oversaturated intersection included testing the effects of queue length on HCM field measurement method for control delay, calculation of delay using HCM formulas (1997 and 2000) and PDM through computer softwares, and calculation of delay manually using HCM 2000 and Akcelik formulas and comparison of these delay estimates with the actual delay. The results of the analyses showed that the field delay measurement tends to be misleading for long queue lengths. Manual computation with HCM 2000 formulation resulted with delay estimates closest to the actual delay incurred among the other estimates.