All the traffic models discussed at the Public Inquiry were based on the TRANSYT program developed by the Transport Research Laboratory (TRL). We are very grateful to TRL for allowing KVCA to use a low-cost 'educational' TRANSYT licence, so that Kirkstall residents could be properly represented alongside major companies like BHS. Engineering software is normally expensive because the user base is small and the big property companies can afford to pay. This software would otherwise have been completely unaffordable for a non-profit residents' group. TRL have addressed the problem of social exclusion and we thank them for their help.
TRANSYT is a 1970's Fortran program, which still shows signs of its punched card ancestry, although the latest version comes with a Windows front end. There is nothing old-fashioned about the mathematics inside the package, which is widely used in many countries for designing networks of linked traffic signals. The network is modelled as a series of traffic lanes or "links" connecting the signalised junctions, or "nodes". The user must specify which nodes are linked together, and the length and "saturation flow" for each link. The saturation flow is the maximum possible flow along each link, if the traffic signals were permanently on green. This depends on the road geometry, but is typically around 2000 vehicles (or "pcus") per hour for a single traffic lane with no turning movements.
Kirkstall Node Diagram All ten green nodes are signalised, but node 6 has no signals.
Traffic engineers must specify the signal sequence for each node: i.e. which link movements can take place simultaneously on a green light without the vehicles hitting each other. It is also possible to model flared traffic lanes, left and right filters, steady flows and 'give-way' situations.
Users must provide an estimated traffic flow for each link. This is usually based on current vehicle counts, plus an allowance for the extra traffic expected from the new development. There are often small discrepancies between these data, so TRANSYT adjusts these initial estimates to achieve a degree of congruence between the numbers of vehicles entering and leaving each link.
The software shares out the available 'green time' between the competing traffic flows, and also adjusts the signal timing at different nodes to reflect the distances between them, so that vehicles passing through a well-designed network should see the lights change to green as they approach each node.
Although TRANSYT uses a rugged minimisation algorithm to optimise the signal timings, there are physical limits to any network capacity, so the software reports the percentage saturation, the number of stops and the length of any internal queues which develop along each link.
Kirkstall Network Links (KVCA model, based on Leeds City Council estimates. The other models are similar.)
link | description | length | satn. | flow | comments |
---|---|---|---|---|---|
77 | Bridge Road BHS site entrance | no signals | 47 | give way to eastbound traffic | |
101 | Bridge Road westbound straight ahead | 145 | 1900 | 610 | |
102 | Savins Mill Way left turn | 100 | 1785 | 844 | |
103 | Bridge Road right turn | 85 | 1710 | 452 | |
104 | Bridge Road eastbound straight ahead | 85 | 1915 | 876 | limit queue to 12 cars |
105 | Savins Mill Way right turn | 100 | 1785 | 220 | |
202 | Savins Mill Way westbound through junction | 135 | 1940 | 752 | |
203 | Savins Mill Way left turn to Morrisons | 135 | 1740 | 424 | |
204 | Morrisons right rurn out | 200 | 1760 | 300 | |
205 | Morrisons left turn out | 200 | 1720 | 269 | |
206 | Savins Mill Way right turn to Morrisons | 80 | 1725 | 177 | |
207 | Savins Mill Way eastbound through junction | 80 | 2000 | 310 | |
301 | Abbey Rd straight ahead then right to Savins Mill Way | 200 | 2300 | 412 | |
302 | Abbey Rd straight ahead & left turn | 200 | 3300 | 862 | |
303 | Kirkstall Lane right turn | 230 | 1675 | 163 | |
304 | Kirkstall Lane straight ahead | 230 | 1790 | 562 | |
305 | Kirkstall Lane left turn | 230 | 1770 | 161 | |
306 | Commercial Road left turn | 55 | 1785 | 46 | |
307 | Commercial Road straight ahead | 55 | 3970 | 1308 | limit queue to 13 cars |
308 | Bridge Road left turn | 100 | 2000 | 596 | limit queue to 13 cars |
309 | Bridge Road straight ahead | 100 | 1785 | 420 | |
401 | Commercial Road right turn | 65 | 3300 | 471 | limit queue to 10 cars |
402 | Commercial Road southbound straight ahead | 65 | 3300 | 876 | limit queue to 10 cars |
403 | Commercial Road northbound straight ahead | 105 | 3000 | 1048 | |
404 | Commercial Road left turn | 105 | 1710 | 700 | |
405 | Savins Mill Way right turn | 140 | 1900 | 364 | |
406 | Savins Mill Way left turn | 140 | 2000 | 254 | |
501 | Commercial Road Southbound to Beecroft Street | 125 | 3600 | 1240 | requires substantial road widening |
502 | Beecroft Street lower westbound | 100 | 1730 | 283 | |
503 | Commercial Road northbound to Beecroft Street | 200 | 4000 | 1645 | requires substantial road widening |
601 | KDC main exit on Beecroft Street | no signals | 352 | ||
602 | Beecroft Street upper westbound | 130 | 1980 | 78 | |
603 | Beecroft Street lower eastbound | 120 | 1980 | 266 | |
701 | Morris Lane right turn | 200 | 1600 | 140 | |
702 | Morris Lane straight ahead & left turn | 200 | 1750 | 286 | |
703 | Kirkstall Lane westbound all movements | 200 | 1950 | 590 | |
704 | Kirkstall Hill all ways (BHS) left + straight (LCC) | 120 | 1900 | 694 | see also 707 |
705 | Kirkstall Lane eastbound right turn | 240 | 1800 | 125 | |
706 | Kirkstall Lane eastbound straight ahead & right turn | 240 | 1900 | 564 | |
707 | Kirkstall Hill proposed RIGHT turn lane (LCC only) | 120 | 1665 | 96 | see also 704 |
801 | Kirkstall Hill southbound | 145 | 1940 | 298 | |
802 | Burley Hill northbound | 200 | 1970 | 685 | |
803 | Beecroft Street upper eastbound main flow | 100 | 1685 | 123 | |
804 | Beecroft Street upper eastbound left flare | 145 | 1685 | 49 | |
901 | Bridge Road straight ahead | 210 | 1740 | 431 | |
902 | Bridge Road left turn | 210 | 1740 | 977 | |
903 | Wyther Lane all movements | 70 | 1900 | 762 | limit queue to 4 cars |
904 | Leeds & Bradford Road all movements | 200 | 1665 | 502 | |
1001 | Wyther Lane southbound all movements | 60 | 1845 | 999 | limit queue to 4 cars |
1002 | Wyther Lane northbound all movements | 200 | 1710 | 461 | |
1003 | Broad Lane all movements | 200 | 1910 | 335 | |
1401 | Bridge Road eastbound straight ahead | 140 | 1965 | 1155 | limit queue to 18 cars |
1402 | BHS Site Exit right turn | 50 | 1871 | 50 | |
1403 | BHS Site Exit left turn | 50 | 1791 | 134 | |
1404 | Bridge Road westbound straight ahead | 70 | 2000 | 1344 | limit queue to 11 cars |
Link 1404 is the pinch point for cyclists, where BHS squeeze two congested lanes down to 6 metres, with no cycle lane.
TRANSYT prints a small "+" sign in the results table to warn users that an internal queue has exceeded the physical capacity of the road network, however it continues with the modeling process, so that inexperienced users may not immediately realise that their "solution" could not be implemented on the ground. To avoid this problem, users can impose "penalties" when queues approach the physical length of the connecting links. This will cause the minimisation software to seek other timing solutions where the queueing problems are less serious, or do not occur.
Unfortunately, in the case of the Kirkstall network, there are no other solutions which completely avoid these problems. The highway network is saturated with cars during the morning and evening peaks, and has been so for almost ten years, so the effect of altering the signal timings is merely to move the queues from one location to another, without solving the problem as a whole. Although BHS claimed to have overcome the traffic problems, they have not actually done so. They systematically underestimated the traffic that would be generated by their development, they did not include the traffic from the re-developed Kwik Save site, they omitted Beecroft Street (which constrains the possible timing solutions) and they have five times too many vehicles on Wyther Lane than would fit into the available space. The BHS traffic model imposes no queue length penalties on the critical Wyther Lane links 903 amd 1001, which in reality can only 'hold' five cars in each direction. This causes the program to calculate unduly optimistic vehicle flows.
Extract from the BHS transyt output file showing unrealistic queues on links 704, 903 and 1001.
The problems at Wyther Lane are not new. They were an issue in the 1997 - 98 Public Inquiry into the Morrison's Supermarket development. The traffic flows estimated in 1997 are almost the same as those used today, confirming that the local highway network is fully saturated, and the capacity cannot be increased. Automatic traffic counts recorded over the last ten years are almost flat, as would be expected if the road network were operating at full capacity with no opportunity for improvement.
TRANSYT 12 users: Download the BHS input data. Download the BHS link diagram. Download the KVCA input data. Download the LCC link diagram. Run these tests for yourself!
Although there is no overall solution, we were able to show that cyclists could be more safely accommodated on link 1404 without causing excessive detriment to car drivers. (There is some detriment, but the effect is small. In reality, even without development, there are already massive queues on all the approach roads at peak times.) If bus lanes were provided we could have optimised the bus service as well. We say that developers should be paying for these improvements to favour alternative, less polluting modes of transport, rather than providing for an ever increasing number of private cars.
Please use the browser "Back" button to return to the main site.