Double-deckers not right fit for S'pore roads? (Part 2)

sgbuses SG

This is a follow-up to the newspaper report discussed in this thread:

http://sgforums.com/forums/1279/topics/485771

The academic paper has been published. You will need a University or research subscription to access the full copy:

http://www.sciencedirect.com/science/article/pii/S0965856414002183

Methodology

Eight bus models were used in the research, and a week's worth of dataset from April 2011 (possible criticism: is the dataset too small?). The models were:

- Leyland Olympian 3-Axle
- Volvo B9TL WEG2
- Scania K230UB Euro IV
- Scania K230UB EEV
- Volvo B10M Mark IV (sample size indicates a single sub-variant was used)
- Mercedes-Benz O405 (capacity suggests DM, but large sample size indicates a broad fleet was used)
- Mercedes-Benz OC500LE
- Mercedes-Benz O405G (sample size suggests a broad fleet was used)

and the specifications based on sgwiki. And yes, I can forsee where the debate is going to start. A lot of technical information on sgwiki isn't accurate.

Boarding and alighting activities were timed separately.

The findings of the research are summarised as follow:

- Average boarding times ranged from 1.68s (MB OC500LE) to 2.02s (Volvo B10M) per passenger. Total average is 1.78s per passenger. O405G and LO3X fared worse than total average (1.78s and 1.93s respectively), while B9TL WEG2 was 0.02s better.

- The difference is apparent in the average alighting times per passenger. Total average is 1.48s per passenger. O405G fared best at 1.26s per passenger (obviously, two doors). OC500LE was second best at 1.48 per passenger. LO3X and Volvo B10M tied for the worst timing at 1.78s per passenger. B9TL WEG2 was second worst at 1.75s per passenger. Scania K230UB Euro IV, EEV and O405 were 1.54s, 1.50s, 1.52s per passenger respectively.

Paper's analysis

- High occupancy of bus actually made alighting faster, not slower.

- The more steps there are on the entrance of the bus, the longer it takes for passengers to board and alight.

- Passengers on DD buses tend to wait upstairs until the bus comes to a halt, delaying alighting time. Boarding time is not affected. Recommends that double-deck buses are more useful when there is high demand and many passengers alight at one particular bus stops.

- A large variation of boarding and alighting times between different bus models can lead to irregular frequency and service unrreliability.

My own analysis

- The paper seems to suggest double-deck buses should be deployed on long-haul routes, and routes where most people are transferring to MRT station rather than for point-to-point travel.

- It can be implied that older buses with steps should be retired in favour of new buses with step-free access to reduce dwell times. This is already the case as older models are quickly being replaced these days.

- It can also be implied that operators should use a single model, or bus models with similar dwelling times in a single route to increase reliability of that particular route. This importance has increased since the introduction of BSRF.

SBS2601D SG

No more access to such papers for me.

That aside, the purpose of this thread is to encoruage off-tangent disparaging remarks that reflect general immaturity and lack of inclusivity?

JurongWestresident SG

No more access to such papers for me.

That aside, the purpose of this thread is to encoruage off-tangent disparaging remarks that reflect general immaturity and lack of inclusivity?

Before things get any worse, can I just recommend that everyone in this forum stop doing the following:

sgbuses SG

Cheeky SBS2601D!

I thought the discussion could improve once the blanks are filled in. Anyone researching on Singapore bus transport will find this data a treasure trove.

Some of you may quickly notice that there is no data for Citaro, MAN A22 and A24 due to the age of the dataset. For those who are already gathering data at the loading thread, gathering this missing data is easy. You will need a stopwatch. During your journey:

1. Time boarding and alighting separately. Do one at a time (unless two people can do each). Start from the moment the BC presses the door open button, and stop the moment the BC presses the door close button.
2. Record number of card-paying passengers. Do not include cash-paying passengers. One way is to listen for the number of beeps on the reader.
3. Additional time (e.g. card errors, searching for card) should be factored in, as it was for this research paper.
4. Record the numbers over a range of routes (feeder, trunk, express).
5. Divide the total time the door stays open over the number of card-paying passengers.

If the paper's hypothesis is correct, I would expect the MAN A24 to improve its boarding and alighting time per passenger over the O405G. Not sure if there would be any difference between low-entry and low-floor configuration though.

sgbuses SG

Before things get any worse, can I just recommend that everyone in this forum stop doing the following:

It depends on what you want to get out of this forum. It can be sharing ideas, having a question answered, or even a simple suggestion.

201911 SG

Some of my thoughts:

-Obviously, newer rigid buses with no steps at the doors have faster boarding/alighting than older buses with steps. However, I think we can look at buses with 3 doors to further improve boarding/alighting time. I'm not an expert in this area, but I think some LHD buses with 3 doors are available in Europe, although RHD buses with 3 doors are hard to come by. It would require specific requests from the operators or LTA (whoever is purchasing buses) and close co-operation with the bus manufacturers, but 3 door RHD buses for SG or Asian markets should be explored.

-As SMRT has mentioned when they announced their purchase of their E500s recently, DDs are more suitable for long haul trunk routes where pax boarding/alighting activity is more concentrated at certain specific stops, rather than short haul feeder routes with frequent boarding/alighting at many stops.While DDs provide more capacity w/o the extra space of bendies, the question is how we can speed up boarding/alighting and reduce dwell time. The Borismaster has a 3rd door and extra staircase at the back, but this comes at the cost of reducing capacity since the back staircase takes up significant space.

While it's not perfect, I think the Borismaster's design is a step in the right direction, and further experimentation with redesigning DDs (or other bus types in general) should be done to improve passenger flow while minimising the trade offs. For example, the door width may be increased, or there can be a single staircase for the top deck leading directly to a dedicated entrance at the back.

-Bendies have no problem with bottlenecks at boarding/alighting with the presence of an extra door. Their main issue is the extra road space they occupy, and being more difficult to manoeuver around tight corners (leading to certain route restrictions). Years ago, SBST concluded that bendies were not suitable and sold off their 2 demonstrator buses. Thus, SBST uses DDs to provide extra capacity for both their long haul trunks as well as short haul feeders (which SMRT has rightly pointed out are more suited for bendy operations, due to shorter dwell time with easier boarding/alighting).

Under the government contracting model with all buses belonging to the govt/LTA in future, we could possibly see feeders currently under SBST eg 334, 222 that currently use DDs using bendies instead, whether operated by SBST (using buses leased from LTA) or by SMRT or a new operator. The issue is whether, and to what extent, LTA supports the use of bendies in Singapore. Another problem would be the possible lack of space with ever-shrinking bus interchanges.

carbikebus SG

If everyone think like what the prof think then the world wouldnt be making DDs lor..IMO DDs are suitable for trunk svc 13km above per direction..SDs with a few bendies suitable for short trunk or feeders..Dont like then fly kite

BusAnalayzer SG

The study only takes into account boarding/alighting to indicate DDs not fit. Big flaw in study:

1. does not take into account # more buses required if say all DDs on sv 88 were to be replaced by SDs.

2. does not take into account more qualitative factors like "comfort for passengers". Imagine standing all the way on 168 from Tampines to Woodlands

3. does not take into account how more crowded our interchanges would be if DDs were not there and only SDs were used. Lets for instance even take into account feeder sv like 334 that gets 100+ pax on every bus in PM peak. When the sole KUB comes, it can never take the load and still have 30-40 pax in quque (still wonder why its not become DD slot yet). Imagine then if 334 had all SDs like in previous days.. the number of buses you would had to add

On another note, sgbuses, I am not very sure why you opened this topic again for discussion. Was already discussed, no!?

carbikebus SG

Agree with Dato Busanalyzer...The one doing survey drive car himself?

I dun mind majority SDs for feeders but nt trunk svc pls..

sgbuses SG

The study only takes into account boarding/alighting to indicate DDs not fit. Big flaw in study:

1. does not take into account # more buses required if say all DDs on sv 88 were to be replaced by SDs.

2. does not take into account more qualitative factors like "comfort for passengers". Imagine standing all the way on 168 from Tampines to Woodlands

On another note, sgbuses, I am not very sure why you opened this topic again for discussion. Was already discussed, no!?

I think nobody would question the suitability of double-deck buses for long-haul routes. The research paper itself appear to confirm this as well.

I re-opened this topic not solely to start the double-deck v articulated bus debate. With the results we can properly critique on the research rather than dismissing it off-hand. Furthermore, these same set of numbers can also be used to recognize other possible issues.

Another way to look at this research is to think how can the interior configuration of the double-deck bus be modified to cater for feeder and short-haul routes. Would a straight and wider staircase design improve dwelling times because passengers could move up and down faster? Should we have two set of staircases? Three doors? More standing area? What aspects of any one particular bus model of the same type (i.e. OC500LE) made boarding and alighting so efficient that can be brought over to future procurements?

sgbuses SG

Some of my thoughts:

A custom design for Singapore might be just what we need.

Getting a three-door RHD product is challenging (and even then, the newly designed A95 is two-door configuration). Furthermore, manufacturers are more inclined to sell off-the-shelf modular designs, and options are getting less by the year.

The closest we have for a three-door RHD product is the Optare Tempo:

http://www.busnews.com.au/news/bus-news/1409/video-review-optare-tempo/

^tamago^ SG

To make things easier, consider a third, rearmost door with a smaller door width and a single door leaf.

carbikebus SG

Or a wider centre staircase and the exit door is just a window before the rear 1st axle..

phillipC SG

Professors very good at writting thesis or watever sis you call that.

Nothing practical

SBS7557R SG

I feel the title they used for the article was misleading in the first place. It should be smth like "Do double deckers lengthen boarding/alighting time for bus services?"

iveco SG

SBST has been a heavy user of DDs since the 1970s. Back then, they used bi-axle non-aircon ones which were only mid-capacity (90 pax or so), being only 10.5m long. It was only in 1993 that the LO3X ADD was introduced. Routes 240 and 334 had been dominated by bi-axle DDs since Day #1 of operation and it was inevitable that tri-axle ones would eventually take over.

Bendies, on the other hand, have had a short history, being only introduced by Tibs in 1996. Tibs's successor SMRT Buses may be right in pointing out that bendies are better suited for feeders. However, most of SBST's operational areas are not well-configured for bendies, neither is Singapore's infrastructure on the whole.

The capacity for a Mercedes O405G is 132, which is 7 less than that of the tri-axle Volvo Olympian. This difference is actually minimal.

Bi-axle DDs can be likened to maxi-buses like the Volvo B10M-70, which is 14m long. According to Singapore Buses Wikia, SBS997A could carry up to 110 pax.

carbikebus SG

Pls no tri axles 14 m buses for Sg,Unlike bendy theyre pain in ass for Sg road environment..Those buses are suitable for long distance routes..Sg roads are suitable for 12m buses while bendy best for feeders or short trunk svc.Like i emphasize before if only manufacturers like MAN/ADL/VDL/Volvo or Scania came out a custom tri axle 12.3 m with three door with single leaf type door for the rear most and dual staircase then we can judge the effiency and practicality of the DDs..If thats happening im sure LTA will get rid off the bendies for good.With 2 staircase(1 at most rear) and 3 doors commuters aint kiasi to go upper level even for 3-4 stops

SexyMichael SG

the legal passenger capacity on all buses are very likely to be computed on a formula that involves weight rather than actual standing space occupied. such a formula would allow PTOs to meet the <90% loading criterion with ease, although create real-world loading issues frequently as the actual loading of buses are always <<90% of their legal capacity.

'different buses for different situtations', if that is how the paper can be summed up. for the same number of alighting pax, bendies give generally quicker alighting times than double deckers due to the added door. SMRT has clearly recognised the strengths of the bendy and DD, although it still remains to be seen whether they would fully follow through with additional bendy/DD purchases.

as sgbuses has pointed out, redesigning the interior of the buses may be the way to go such that passenger flow within the bus is optimised and fewer instances of opposing passenger flows give rise to quicker boarding and alighting times.

to start the ball rolling...

1) stickers on the front door leafs stating the recommended mode of payment - this reduces congestion should a cash-paying passenger enter from the right door leaf then approach the validator after the stanchion

2) as tamago has mentioned, a third door similar to that on the NBfL - the drivetrain has to be able to accomodate a third door after the drive axle

carbikebus SG

Agreed,A fine example is A39 but its a bit too long at 13.7m,Something like 12.5m might be feasible..If any manufacturers can customise no harm PTOs/LTA give it a trial..Who knows MB/MAN/ADL might accept the ideas?Hong Kong also can order..

ButIAmAToilet SG

Or a tri-axle, 2-stairs DD, but very few seats downstairs? Confirm people will move up, and can hit around 110+pax? Standard 2-stairs DD like this MAN ND202 (I read up about it for random reasons, it is the only 12m 2-stairs DD I remembered so far) has a capacity of about 94 pax.