Delay Analysis at Signalized Intersections Based on Malaysian Road Conditions

Transcription

1 Proceedings o the Eastern Asia Society or Transportation Studies, Vol.7, 2009 Delay Analysis at Signalized Intersections Based on Malaysian Road Conditions Nurikhwani Idayu ZAINAL ABIDIN PhD student School o Civil Engineering, Universiti Sains Malaysia, Engineering Campus, Nibong Tebal, Pulau Pinang, Malaysia Tel: ayuikhwani@yahoo.com Wan Hashim WAN IBRAHIM Proessor Dean, Faculty o Engineering, Universiti Malaysia Sarawak, Kota Samarahan, Sarawak, Malaysia Tel: wiwhashim@eng.unimas.my Ahmad Farhan MOHD SADULLAH Proessor Deputy General, Malaysian Institute o Road Saety Research (MIROS) Lot , Jln TKS 1, Taman Kajang Sentral, Kajang, Selangor Darul Ehsan, Malaysia Tel: Lee Vien LEONG Lecturer School o Civil Engineering, Universiti Sains Malaysia, Engineering Campus, Nibong Tebal, Pulau Pinang, Malaysia Tel: celeong@eng.usm.my Abstract: Capacity analysis or local roads in Malaysia has always been carried out using manuals rom other countries. These methods include the Arahan Teknik (Jalan) 13/87, the U.S. Highway Capacity Manual (HCM 2000) and aasidra. However, in 2006, the Ministry o Works Malaysia introduced Malaysia s very own Malaysian Highway Capacity Manual (MHCM 2006) with hopes that a more accurate outcome will be the result rom the usage o this manual. This paper discusses the comparison done to compare the results rom the our manuals above with the actual situation on-site. The outcome o the study shows that rom all our manuals, the MHCM 2006 gives the result closest to what was obtained on-site. Key Words: capacity analysis, MHCM 2006, control delay 1. Introduction The HCM 2000 deines capacity analysis as a set o procedures used to estimate the traic carrying capacity o transportation acilities over a range o deined operational conditions. The end result o the analysis is the level o service (LOS) o the particular intersection. This value o LOS is what will be used in order to plan and design uture acilities (Arnold and McGhee, 1995) In 2006, the Ministry o Works, Malaysia had introduced the Malaysia Highway Capacity Manual (MHCM 2006). This manual was to be used to help in conducting highway capacity analysis or our local roads. Prior to this, engineers and practitioners in Malaysia have been using other methods rom other countries or this purpose. The most common methods used were: Arahan Teknik (Jalan) 13/87, the U.S. Highway Capacity Manual (HCM 2000) and aasidra.

2 With the development o the MHCM 2006, it is hoped that this manual will be able to assist local engineers and practitioners in carrying out highway capacity analysis and designs or local roads. This is because this manual was based on local traic conditions and travel behavior, thus would better suit the local roads instead o using manuals orm other countries that might or might not suit our roads (Nurikhwani Idayu, 2007). The Arahan Teknik (Jalan) 13/87 originated rom the method used in the United Kingdom during the 50s and 60s. This method was developed by Webster and Cobbe and is clearly not based on local traic conditions as well as the travel behavior o road users here in Malaysia. The HCM 2000 on the other hand is a capacity analysis manual developed in the United States. This manual only has three vehicle categories which are the light vehicles, recreational vehicles and heavy vehicles (which includes trucks, lorries and buses). This can become a problem or Malaysia as Malaysia has a high number o motorcycles as can be seen in Table 1 and motorcycles clearly could not be classiied in either classes. Table 1: New Registered Vehicles by Type, Malaysia, (Source: Road Transport Department, 2007) Hire & Goods Year Motorcycle Motorcar Bus Taxi Drive Others Total Vehicle Car , ,765 2,620 4,358 2,545 69,234 30, , , ,343 2,947 5,257 1,860 65,160 28, , , , , ,786 6, , , , ,925 1,724 19,987 8, , , , ,635 2,883 24,316 11, , , , ,169 1,348 25,612 13, , , , ,446 1,242 25,415 16, , , ,753 1,014 5,542 1,231 29,975 17, , , ,116 1,290 7,746 1,797 33,169 18, ,363 1,020, , ,900 1,568 5,002 3,411 33,532 16,440 8 On the other hand, the aasidra has only two categories namely light vehicles and heavy vehicles. From these vehicle classes, the same problem is also aced where motorcycles are not taken into consideration or the capacity analysis. This will pose as a problem as motorcycles are too small to be considered as cars and totally eliminating them would greatly aect the outcome o the analysis as Malaysia has a large number o motorcycles on our local roads. The reason on why it is important to include the motorcycles in the analysis is because there is a high number o motorcycles on Malaysian roads. As given by the Arahan Teknik (Jalan) 13/87, the motorcycle has a passenger car equivalent (Pce) as 0.33 which literally means that it takes about one-third o the place o a passenger car. When this is translated into volume per passenger car unit (Pcu) or the year 1996 based on Table 1, the number o motorcycles will translate into pcu which is a very high number. When this high number is not included into the analysis, it will greatly aect the outcome o it as a large portion o the volume has been improperly disregarded. However, to compare these manuals, a certain parameter was to be determined irst. This

3 parameter must be something that not only can be calculated, but could also be determined onsite using manual observation. As been said earlier, the outcome or capacity analysis is the value o LOS. To determine the LOS, the value o delay is used as the reerence, where the lower the value o delay, the better grading o LOS was given or the intersection. However, only stopped delay can be determined on-site while by calculation, the delay obtained is the control delay. To overcome this problem, the relationship between stopped delay and control delay had to be determined. Akcelik (1988) claimed there was an assumption the the stopped delay o always 77% o the overall delay. With this assumption in mind, in order to convert the overall delay into stopped delay, a actor o 1/1.3 was applied to the calculated overall delay to obtain the value o stopped delay, thus, all manuals can now be compared. 2. Methodology For this study, a ew methods were used in order to come up with the appropriate results. First, traic counts were conducted in order to obtain data or the capacity analysis. Data rom a total o 12 sites were obtained or this study. During this traic count, the value o stopped delay was also observed manually, parallel to the traic count conducted. The TDC-8 (JAMAR Technologies) was used to determine the value o stopped delay on-site. Next, the data rom the traic count was analyzed using all our manuals which are the MHCM 2006, HCM 2000, Arahan Teknik (Jalan) 13/87 and aasidra. The calculations or Arahan Teknik (Jalan) 13/87 and MHCM 2006 were carried out manually with guidance rom the respective manual. For aasidra, the sotware itsel was used to carry out the calculation while HiCAP v.2.0 was used to calculate or HCM As commonly known by traic engineers and practitioners, in order to calculate the delay, there are a ew steps o calculation that must be done irst. The initial step or this set o procedures is to calculate the saturation low. From the estimated saturation low obtained, the v/c ratio could be calculated. All these values are important in order to calculate the delay that will be the end result in determining the LOS o the intersection. Dierent manuals have dierent ways o conducting the capacity analysis. For HCM 2000, two methods were used to estimate the saturation low while the rest o the procedures were the same. Equation 1 shows the calculation or estimated saturation low rom the HCM 2000 and this equation was used to dierentiate both methods. s= s N [1] o w HV g p bb a LU LT RT Lpb Rpb s = estimated saturation low rate (veh/hr) s o = base saturation low rate (pcu/hr/lane) N = number o lanes in lane group w = adjustment actor or lane width HV = adjustment actor or heavy vehicles g = adjustment actor or approach grade = adjustment actor or parking activity p

4 bb a LU LT RT Lpb Rpb = adjustment actor or bus blocking = adjustment actor or area type = adjustment actor or lane utilization = adjustment actor or let turning = adjustment actor or right turning = pedestrian adjustment actor or let turn movements = pedestrian adjustment actor or right turn movements The irst method is by using the number o vehicles and the heavy vehicle adjustment actor. In Malaysia, there is a wide variety o transportation on local roads. However, when using this method. it does not consider the dierent vehicle composition ound on our roads as the heavy vehicle adjustment actors or the HCM 2000 only takes into consideration recreational vehicles, trucks (which are deined as vehicles with more than our tires touching the pavement) and buses whereas in Malaysia, the vehicles have been categorized into ive classes (i.e. cars, motorcycles, lorries, trailers and buses). This method will be known as Method 1 in this paper. As or the second method used or this study, it represents the current practice o converting the vehicles into passenger car units (pcu) using the pce values obtained rom the Arahan Teknik (Jalan) 13/87. When using this method, the heavy vehicle adjustment actor is not used, thus the value or this adjustment actor is then given as This method will be known as Method 2 in this paper. Ater the estimated saturation low was determined, the capacity and v/c ratio is calculated. For the HCM 2000, capacity or each approach in an intersection is calculated using Equation 2. c i s i ci = si ( gi / C) [2] = capacity o lane group i (veh/hr) = saturation low rate or lane group lane i (veh/hr) (g i /C) = ratio o eective green time or lane group i Ater that, the v/c ratio is calculated by dividing the volume o the lane to its capacity as seen in Equation 3. v v X = v C i i i = = [3] c i gi si g i si C X i v i s i g i C = v/c ratio or lane group i = actual or projected demand low rate or lane group i (veh/hr) = saturation low rate or lane group i (veh/hr) = eective green time or lane group i (s) = cycle length (s)

5 When both the capacity and v/c ratio o the lanes have been determined, the delay is calculated. The delay used in this manual is the control delay. Control delay is a portion o the total delay attributed to traic signals operation or a signalized intersection which includes initial deceleration delay, queue move-up time, stopped delay and inal acceleration delay (TRB, 2000). To estimate control delay that is used by the HCM 2000, Equation 4 is used. d = d ( + d + d d [4] 1 PF) 3 d = control delay per vehicle (sec/veh) d 1 = uniorm control delay (sec/veh) PF = uniorm delay progression adjustment actor d 2 = incremental delay (sec/veh) d 3 = initial queue delay (sec/veh) To complete the calculation or the control delay using HCM 2000, other delays must also be calculated. These delays include the uniorm delay (Equation 5), the incremental delay (Equation 6) and initial queue delay. 2 g 0.5C 1 C d 1 = [5] g 1 min(1, X ) C d 1 C g X = uniorm delay (sec/veh) = cycle length (sec) = eective green time (sec) = v/c ratio or degree o saturation d 2 T k l c X 2 8kIX d 2 = 900T ( X 1) + ( X 1) + [6] ct = incremental delay (sec/veh) = duration o analysis period (hr) = incremental delay actor = upstream iltering/metering adjustment actor = lane group capacity = v/c ratio or degree o saturation For initial queue delay (d 3 ), this delay is estimated or additional delay caused by the initial queue as the residual queue rom the previous cycle. This happens when there is a remaining

6 queue rom the previous queue that did not cross the stop line during the previous green phase. When this happens, additional delay is experienced by vehicles arriving in this period as the vehicles rom the previous queue must irst clear the intersection. For MHCM 2006, the calculation o the estimated saturation low was carried out using Equation 7. so N w g p bb a LT RT s= [7] c s s o N w g p bb a LT RT c = estimated saturation low rate (veh/hr) = base saturation low rate (pcu/hr/lane) = no o lanes or lane group = adjustment actor or lane width = adjustment actor or approach grade = adjustment actors or parking activities = adjustment actor or bus blockage = adjustment actor or area type = adjustment actor or let turning vehicles = adjustment actor or right turning vehicles = adjustment actor or vehicle composition In regards to the special characteristic o motorcycles in the Malaysian vehicle volume on the road, the vehicle composition adjustment actor has been developed. This adjustment actor includes three actors; the adjustment actor or cars composition, heavy vehicle composition and motorcycle composition. This can be seen in Equation 8. = + + [8] c car HV motor car HV motor = adjustment actor or vehicle composition (cars) = adjustment actor or vehicle composition (heavy vehicles) = adjustment actor or vehicle composition (motorcycles) To calculate the respective vehicle composition adjustment actors as mentioned above, the ollowing equations are used. For car, Equation 9 is used. = e q Q car car car [9] car e car q car /Q = adjustment actor or vehicle composition (cars) = passenger car equivalent or passenger cars = proportion o cars in total volume

7 For motor, Equation 10 is used. = e q Q motor motor motor [10] motor = adjustment actor or vehicle composition (motorcycles) e motor = passenger car equivalent or motorcycles q motor /Q = proportion o motorcycles in total volume Three vehicle categories are included in the heavy vehicles adjustment actor. These three categories are lorries, trailers and buses. For this adjustment actor, the actors or each category are calculated respectively beore added together to determine the HV as can be seen in Equation 11. = + + [11] HV lorry trailer bus HV lorry trailer bus = adjustment actor or vehicle composition (heavy vehicles) = adjustment actor or vehicle composition (lorry) = adjustment actor or vehicle composition (trailer) = adjustment actor or vehicle composition (bus) The adjustment actors or the three categories are as in Equation 12, Equation 13 and Equation 14. qlorry = lorry elorry [12] Q lorry e lorry q lorry /Q = adjustment actor or vehicle composition (lorries) = passenger car equivalent or lorries = proportion o lorries in total volume = e q Q trailer trailer trailer [13] trailer = adjustment actor or vehicle composition (trailers) e trailer = passenger car equivalent or trailers q trailer /Q = proportion o trailers in total volume qbus bus = ebus [14] Q

8 bus e bus q bus /Q = adjustment actor or vehicle composition (buses) = passenger car equivalent or buses = proportion o buses in total volume To calculate the delay using this manual, the MHCM 2006 has adopted the equations used by the HCM 2000 (please reer Equations 4, 5 and 6). On the other hand, or both manuals, it is also desirable to aggregate the delay values to provide delay or an intersection approach and or the intersection as a whole (Ministry o Works Malaysia, 2006). This is done by using Equation 15 and Equation 16. divi d A = [15] v i d A d i v i = delay or approach A (sec/veh) = delay or lane group i (on approach A) (sec/veh) = adjusted low rate or lane group i (veh/hr) d Av A d i = [16] v A d i d A v A = average delay per vehicle or the intersection = delay or approach A (sec/veh) = adjusted low or approach A (veh/hr) The values o the control delay obtained rom calculation using MHCM 2006 and HCM 2000 was converted into stopped delay using the assumption discussed above in the previous section. For the Arahan Teknik (Jalan) 13/87, a more complicate approach is taken. First, the saturation low is determined according to the lane width where Equation 17 is used to estimate saturation low or lane widths greater than 5.5 meters. S W = saturation low (pcu/hr) = lane width (m) S = 525 W [17] However, Table 2 is reerred to or lane widths less than 5.5 meters.

9 Table 2: Relationship between eective lane width and saturation low (Source: Ministry o Works Malaysia, 1987) W (m) S (pcu/hr) Ater the saturation low has been determined using either Equation 10 or Table 2 above, a series o correction actors are applied to the obtained value. These correction actors include the correction actor or gradient, the correction actor or the turning radius and the correction actor is turning traic. Ater the saturation low has been determined by multiplying with all relevant adjustment actors, the determination o the y value is carried out. Y value is the ratio o low to saturation low and is determined using Equation 18. The y value used in calculation would be the highest y value calculated rom all approaches within that phase. y = q/s [18] y q S = ratio o low to saturation low = actual low on traic-signal approach converted to pcu/hr = saturation low or the approach in pcu/hr Next is the determination o lost time using the equation by Webster and Cobbe (1966) as seen in Equation 19 and the inter-green time as calculated using Equation 20. n n I a) + i 1 i= 1 L= ( l [19] I A l = the inter-green time between the phases (s) = the amber time, usually taken as 3 seconds = drivers reaction time at beginning o green per phase. In practice, this time is set to 2 seconds but 0-7 seconds can also be used I = R+ a [20] I R a = inter-green time (s) = all red interval (s) = amber time (s) Ater the lost time has been determined, the optimum cycle time is then calculated. The equation used by the Arahan Teknik (Jalan) 13/87 is given in the Road Research Technical Paper No. 56 as in Equation 21. For practical purposes, the cycle time should be between 45 seconds to 120 seconds

10 C o L Y C o 1.5L+ 5 = 1 Y = optimum cycle time (s) = lost time (s) = ratio o low to saturation low [21] Then the signal settings are determined where green time or each phase is calculated. For each individual phase, the Y value is irst calculated or the respective phases. Ater that, the eective green time or each signal phase is calculated using Equation 22. g n Yn = ( Co L) [22] Y g n Y n C o L = eective green time o the n th signal phase = calculated Y-value o the same signal phase = optimum cycle time (s) = lost time (s) The Arahan Teknik (Jalan) 13/87 uses the ollowing equation in estimating stopped delay or each approach. d c λ q x C(1 λ ) x d = (1 λx) 2q(1 x) [23] = average delay per vehicle = cycle time (s) = proportion o the cycle that is eectively green or the phase under consideration (i.e. g/c) = low (veh/hr) = degree o saturation, which is the ratio o actual low to the maximum low that can pass through the approach (i.e. q/λs) As or aasidra, the sotware itsel is used or this study and in this research, aasidra version 2.0 was used. This sotware is widely used by the Australian government and authorities. This sotware is also used by Australian academic and research organizations (Troutbeck and Akcelik, 1994). Ater all values o the delay were obtained, the analysis was carried out to compare all values with the actual value on-site. 3. Results and Discussion The values o delay, both calculated and obtained on-site can be summarized in Table 2. As can be seen in the table, all values o the control delay have been listed rom the respective manuals as well as the value o stopped delay on site that has been converted to control delay.

11 Site Table 2: Comparison between the delays rom the chosen manuals Delay by calculation Observed Control delay onsite (s) MHCM HCM 2000 Control Delay Control Delay Method 1 Method 2 Arahan Teknik (Jalan) 13/87 Stopped Delay Control Delay aasidra Control Delay Nasmir Nasmir Jln Bakar Arang Jln Pahlawan Jln Badlishah Jln Kg Perak Jln Kg Perak Jln Stadium Jln Raja Dihilir Jln Raja Dihilir Jln Tambun Jln Leong Boon Siew The dierence between these values can also be seen graphically as in Figure 1 below. Comparison o Delay between manuals and actual value Control Delay (sec/veh) On-Site MHCM HCM 2000 (M1) HCM 2000 (M2) Arahan Teknik (Jalan) 13/87 aasidra Nasmir 1 Nasmir 3 Jln Bakar Arang Jln Pahlawan Jln Badlishah Jln Kg Perak 1 Jln Kg Perak 2 Jln Stadium Sites Jln Raja Dihilir 1 Jln Raja Dihilir 2 Jln Tambun Jln Leong Boon Siew Figure 1: Comparison o Control Delay between Manuals and Actual Value From the diagram above, the MHCM gives a calculated value that is almost identical to the value obtained on-site. However, it can be seen clearly that some o the methods used have values that are way too much when compared to the actual value obtained on-site. This is especially clear in the HCM method or some o the sites even when the ormula used to calculate the control delay or both MHCM and HCM 2000 are similar. To urther compare the precision o the estimated control delay rom all manuals towards the

12 actual value obtained on-site, the results have been summarized graphically in another type o graph as can be seen below. This type o graph resembles a goodness-o-it type o graph where the y-axis and x-axis o the graph represents the estimated control delay o and actual control delay respectively. The linear line on the graph represents the situation when the actual value o the control delay equals to the estimated value calculated by the respective manuals. The scattered dots on the graph represent the control delay obtained rom the study. As seen rom Figure 1, the MHCM has the closest value with the actual value on-site. This is urther proved by the Figure 2 below. It can be seen that the dots that represent the delay value obtained rom the study are situated close to the linear line that represents the optimum situation when the estimated and calculated value o control delay are the same. From this observation, it can be summarized that the MHCM is the best manual to be used when conducting a capacity analysis or Malaysian roads as it well represents the actual local traic conditions and traic behavior. Comparison o MHCM and value on-site E s tim a te d v a lu e o c o n tr o l d e la y (s e c ) Actual value o control delay (sec) Figure 2: Comparison o control delay or MHCM and control delay on-site Figure 3 and Figure 4 represents the comparison done or HCM 2000 using both methods. A similar pattern was obtained where although some sites have results almost similar to the ones on-site, there were a ew results that were very ar rom the optimum situation, thus showing that the HCM 2000 may be good or some sites, but not good or other sites. From observation, the results that were o the chart were analysis or shared lanes or exclusive turning lane. The conclusion that can be made rom these igures is that HCM 2000 can be used to analyze through lanes in Malaysia, but or shared lanes or exclusive lanes, the results obtained were too ar rom the actual value.

13 Comparison o HCM 2000 (Method 1) and value on-site Estim ated value o control delay (sec) Actual value o control delay (sec) Figure 3: Comparison o control delay or HCM 2000 (Method 1) and control delay on-site Comparison o HCM 2000 (Method 2) and value on-site E s tim a te d v a lu e o c o n tro l d e la y (s e c ) Actual value o control delay (sec) Figure 4: Comparison o control delay or HCM 2000 (Method 2) and control delay on-site Figure 5 represents the situation or Arahan Teknik (Jalan) 13/87. As can be seen, the results are not even remotely close to the optimum situation as compared to MHCM and HCM Whilst the MHCM have all values close to the optimum situation, and HCM 2000 having a ew sites that are close, the Arahan Teknik does not have such values as all values are scattered ar away rom the optimum line. This actually should not be a surprise as the Arahan Teknik, though using a Malaysian name, was adopted rom a study in the 1960s.

14 Comparison o Arahan Teknik and value on-site 350 E s tim a te d v a lu e o c o n tr o l d e la y (s e c ) Actual value o control delay (sec) Figure 5: Comparison o control delay or Arahan Teknik (Jalan) 13/87 and control delay on-site Lastly, Figure 6 represents the situation aced when using aasidra. A similar pattern can be observed by aasidra as with the HCM 2000 where some sites have values that are close to the optimum values. However, there are still some values that are ar rom the optimum situation. Comparison o aasidra and value on-site E s tim a te d v a lu e o c o n tr o l d e la y (s e c ) Actual value o control delay (sec) Figure 6: Comparison o control delay or aasidra and control delay on-site 4. Conclusion From this study, it can be concluded that only the MHCM shows a consistent similarity to the actual value that was obtained on site. Other methods, particularly the HCM 2000 and

15 aasidra have some results that are similar to the ones on-site but also some results that are too ar o. This should not be surprising as the studies conducted to produce this manual were done throughout Malaysia, on our own local roads. This might be due to the act that among all manuals compared, only the MHCM takes motorcycles into consideration. As discussed earlier, the volume o motorcycles on Malaysian roads are almost hal o the total vehicle volume, being second only to passenger cars. When the motorcycle volume is disregarded, a large portion o the total volume is being compromised when it should not. The importance o these results is to avoid over-designing or under-designing Malaysian roads. This is because, capacity analysis is usually carried out to evaluate existing traic conditions and to orecast whether some changes must be done. The orecasted changes would then also be subjected to the capacity analysis to determine whether the planned changes are suitable or not in order to improve the intersection. When the capacity analysis is carried out using other manuals that are not suitable to local conditions, over-designing or under-designing might happen. This might lead to unnecessary upgrading o roads that costs a lot o money but the worst that can happen is that it might compromise the saety o the intersection. This study also proves the reliability o the manual when used to carry out capacity analysis and designs or local roads as proven by the comparison o all manuals with the actual result obtained on-site. Thereore it can be said that or capacity analysis on Malaysian roads, the MHCM is the most suitable method to be used. 5. Acknowledgement The authors would like to acknowledge the Highway Planning Unit (HPU) rom the Ministry o Works, Malaysia or inancing this study. 6. Reerences Akcelik, R., (1988). The Highway Capacity Manual Formula or Signalized Intersections. ITE Journal (March, 1988), pp Arnold Jr, E. D. and McGhee, C. C., (1995). A Survey o Capacity Analysis Practices in State Departments o Tansportation. ITE Journal (July, 1995), pp Ministry o Works Malaysia (1987). A Guide to the Design o Traic Signals. Arahan Teknik (Jalan) 13/87, Chapter 6. Ministry o Works Malaysia (2006). Malaysian Highway Capacity Manual. Chapter 3. Ministry o Transport Malaysia, Statistic unit. Nurikhwani Idayu, Z.A., (2007), Evaluating Signalized Intersection Capacity Based On

16 Malaysian Road Conditions, MSc thesis, Universiti Sains Malaysia. Road Transport Department o Malaysia. Transportation Research Board (2000), Highway Capacity Manual, Washington D.C Troutbeck, R. J. and Akcelik, R., (1994). Capacity Research and Applications in Australia. Second International Symposium in Highway Capacity, Country Reports (August, 1994), pp. 1-18