6 Guide to Safety at Sports Grounds, Sixth Edition, Sports Grounds Safety Authority (SGSA), Department for Digital, Culture, Media and Sport (DCMS), 2018.
There is evidence that due to increased collision avoidance behaviours, the flow rate is reduced more than VSBs alone where
- Items such as bottles or cups are placed on the top of bollards or people are sat on low-rise bollards;
- The height of bollards exceeds 1.2m;
- Staff are positioned at VSBs.
It is recommended that these factors be considered within the operational plan.
Flow rates and throughput
Flow rates are used in event venues to:
- estimate how long it takes for spectators to pass through elements inside and outside the grounds(gates, doorways, stairs, etc.);
- be used as a basis for determining throughput - calculating entry capacity, exit capacity and emergency exit capacity;
- help determine the maximum capacity of the full venue (alongside risk assessment methodologies);
- plan crowd movement in the vicinity of the venue (i.e. “Zone Ex” and “Zone 5”); and
- support planning of security overlays, crowd management, transport requirements, etc.
Flow rates are typically considered as the number of persons per metre width per minute (p/m/min). The Guide to Safety at Sports Grounds (the “Green Guide”6) outlines maximum flow rates currently to be used for calculation purposes using this unit of measurement:
- On a stepped surface, 66 p/m/min
- On a level surface, 82 p/m/min
6 Guide to Safety at Sports Grounds, Sixth Edition, Sports Grounds Safety Authority (SGSA), Department for Digital, Culture, Media and Sport (DCMS),2018.
 |
 |
 |
 |
 |
 |
|
| Level of service for walking (Note: there are different scales for other activities such as using stairs or queuing) |
Free circulation, with ample space for conflicting movements and opposing flow. | For one directional flows, free circulation. For reverse and crossing flows, minor conflicts. |
Some restriction in selection of walking speed and ability to pass others. High probability of conflict, and increased likelihood of restricted movement in opposing flow. |
Restricted and reduced walking speed for most spectators, due to restrictions in passing slower moving pedestrians. | The majority of spectators would have their walking speeds reduced and restricted. | All spectator movements are reduced to shuffling, with unavoidable moments of contact with other pedestrians. The spectator experience is similar to moving queues. |
| Level of service (LoS) | A | B | C | D | E | F |
| Flow rate (p/m/min): | 21-23 | 23-33 | 33-49 | 49-66 | 66-82 | Restricted, lower than LoS E |
| Density (people/m2) | <0.31 | 0.31-0.43 | 0.43-0.72 | 0.72-1.08 | 1.08-2.15 | >2.15 |
LoS A is the highest, and most free-flowing level of service, providing the most circulation freedom for spectators; and LoS F is the most restrictive of conditions, providing a less comfortable experience for spectators. LoS F is not necessarily unsafe, but it is more likely that it poses risks to safety (arising from higher crowd densities).
Bollard arrays in passages
Where spectators use straight line passages (e.g. paths, roads or bridges), the movement is almost completely unidirectional, and there is the potential to position VSBs perpendicular to the direction of channel.
Placing a bollard line within a passage has been observed to reduce flow rates across the full width in the order of 10%, where the bollard dimensions are less than 1.2m in height and 0.3m in width.
While ideally the flow would be head-on to the bollard line, it is recognised that this is not an exact science. However, there is a theoretical basis to support the 10% reduction up to 20º deviation from the perpendicular10.
Testing has shown that there is some variation on the densities observed when a bollard array is positioned on the entrance or at the end of a passage2.
However, these outcomes are similar to those within a passage and it is recommended that the 10% reduction is applied in these circumstances.
If the configuration is different to this (i.e. not in a straight line), it can impact how much space people perceive.
Configuration and perception
Angle of approach
Where arrays are configured with even spacing but not in a continuous straight line, the diagonal bollards along the line of sight appear closer together to pedestrians approaching straight on.
If the VSB alignment is not perpendicular to the angle of approach (±20º), then where higher density crowd conditions (>0.4p/m2) are expected, the recommended reduction in throughput to be adopted is to treat the bollards as obstacles.
For example, if the array were constituted of 10 bollards with a width of 50mm each, then the capacity of this route should be calculated with 500mm removed from the total width of the passage.
As demonstrated on the previous page, to avoid over- or under-utilisation within the same bollard array, a mix of approach angles on the same array or succession of arrays should be avoided.
If a mix of approach angles must be used, it is important to consider the impact this might have on usability of each area of the array.
Desire lines on approach to VSBs
Observations show that downstream destinations along with the immediate built environment create desire lines, which can impact the distribution/volume of people flow across the width of the array.
As a result, and depending on the specific topography of each site, bollard arrays are not always used evenly across their entirety.
It is important to consider these desired paths in planning the configuration of the array (see the previous pages on Configuration and perception, and Angle of approach, p19-20).
Bollard arrays that precede obstacles
The presence of obstacles in the vicinity of VSBs has the potential to reduce the flow of pedestrians moving through a bollard array.
For example, see the configuration (below). Crowds move through a passage with a widening at the end.
The bollard array is positioned and configured well: in the widening, with an angled bollard line on the right-hand side (viewed from above) to maximise the angle of approach.
However, the crowd density at the bollard array is increased compared to the approach, despite the increase in space.
This indicates that other factors impact the density and lower walking speeds in this area, such as the food kiosk and its queue, signage and people meeting/dwelling.
These create several obstacles on both sides of the array and reduce usable space.
It is important to carefully consider the placements of VSBs in areas with high levels of dwelling and queuing.
Static features i.e. VSBs, may be inaccurately identified as a greater obstacle compared to transient queues and groups meeting and dwelling.
Reconfiguring waiting spaces and food and beverage positions may reduce the impact of the barriers on spectator movement.
This may not be immediately obvious unless the location is observed during busy periods.
Bollard arrays at junctions
Significantly lower flow rates have been observed at junctions (i.e. where pedestrians select a change in route/direction).
When there is a change of direction or a merging of flows beyond the VSBs, this has been observed to affect the speed at which people approach the decision point.
It was also found that, when there is a sharp turn following the bollard array, densities can be significantly increased on the corner of the turn.
This level of crowding means there is unavoidable movements of contact with other spectators, and crowd movement can be reduced to a shuffle.
Corner bollard arrays
Given the nature of dispersal routes away from venues, there are instances where crowds reach a location where two onward routes at right angles are available. Both of these routes would require protection, and it may be that the nature of the land ownership is such that the natural positioning of VSBs would result in a right-angle configuration.
Where the aggregate width of the two onward routes is greater than that which feeds it, then the bollard arrays themselves do not inhibit free flow in their own right.
The main consideration is that these bollards are positioned in a location where there are turning movements involved, and that they are placed in a location where there is decision-making within the crowd – and both of these features result in crossing movements within the crowd as the bollard arrays are approached.
For example, consider the adjacent configuration. The impact of the bollards on flow is small given that the overall capacity of the onward routes is greater than the approach.
However, the crowd density in the vicinity of the bollard array is increased compared to the approach, particularly as spectators making a turning movement will cross paths with those continuing straight on and also attempt to ‘hug’ the corner.
It is important to carefully consider the placements of VSBs in areas that are key way finding decision points, as the complexity of movements in these areas may be exacerbated by the introduction of additional infrastructure.
Affordance and VSB design
Observations show that at low to moderate densities, i.e. where pedestrians have more freedom of choice over their movements, there appears to be an avoidance of the darker paviours, which form part of the VSB structure.
This is more pronounced when there is a strong contrast between pavement and paviour (e.g. a black paviour on a lightly coloured pathway).
Observations showed that where the contrast is lower (e.g. in wet conditions where the
paviour colour more closely matched the pathway), pedestrians did not exhibit this behaviour as frequently.
Operational interventions
It is recommended that the relationship between existing operational interventions and VSBs is considered at the earliest (design) stage. The most common operational interventions in this context are those supporting either crowd management or security operations.
Examples are filter cordons used to slow down and filter pedestrians approaching for the purpose of monitoring them more closely; and barrier cordons used to segregate, stop or control the crowd’s movement.
Existing operational interventions may be placed in line with the VSBs where they can actively support the existing objectives and where not contributing to undue obstruction or confusion to oncoming pedestrians.
It is recommended that the placement and design of a VSB array should not result in additional operational interventions offsetting adverse impacts.
Should an operational intervention be necessary to ensure that VSBs do not adversely impact comfort, convenience or other pedestrian experiential factors, the selected design and product should be reconsidered as a poor design can seldom be offset effectively through operational management.
Flat-top features
Observations have shown spectators may place objects on flat-top VSBs (e.g. rubbish such as food packaging, plastic bags and bottles).
This can impact how people move through the array (e.g. people will move to avoid the rubbish on the barrier and alter their bodies to a greater degree than if there was just a bollard).
The presence of these objects creates additional collision avoidance behaviours as people are moving/turning their upper bodies away from a perceived obstacle.
In effect, the placement of objects makes it somewhat harder to avoid collision at around elbow height.
Observations showed the objects placed on top of VSBs become dislodged over time and thus increased the risk of slips, trips and falls.
Planners should consider the design of bollards where litter might accumulate and conduct a risk assessment for hazards including slips, trips, falls, broken glass etc.
Slanted tops are often available as a design choice and are less likely to have litter placed on them because it cannot be easily deposited.
If flat-top bollards must be used, consider stewarding, messaging and/or see-through bins, to reduce littering on and around the bollards.
Seat-height features
Where predominantly static activities take place (e.g. waiting to meet other people) VSBs and integrated parts of an array (e.g. trees and fencing) can increase pedestrian comfort by way of offering space to sit or lean.
Observations show that these activities were particularly prevalent in locations where bollards are up to 900mm in height.
However, at peak times (e.g. on egress) static activities like waiting or sitting can pose a significant obstacle to pedestrian flow.
Sitting and leaning activities can be deterred through good design, for example through convex or undulating shapes as well as more innovative ideas including landscaping where appropriate.
Alternative targeted operational interventions should ensure that these behaviours do not cause obstacles during peak flow, including through the use of stewards.
If predominantly static activities are likely to be an issue alternative locations should be provided (e.g. meeting points, pick-up/drop-off areas etc).
'Stop and Go' systems
“Stop and Go” or “Stop and Hold” crowd management techniques are often used in advance of a VSB array or at the array.
The data shows that where “Stop and Go” systems are put into operation in advance of the VSBs (i.e., upstream), flow rates at the bollard line can approach the maximum guidance values (i.e., approximately 80 p/m/min). This is a result of the initial surge in pedestrian movement following a release of the cordon.
These high flow rates are not sustained for significant periods of time. Soon after the initial peak, where pedestrians no longer have clear space on the approach to the bollards, the crowd movement becomes a more constant stream and flow rates settle to more typical values for Zone Ex (i.e., 50-60 p/m/min).
Where the array is also used as the management point for the “Stop and Go” cordon, this reduces the volume of the initial surge of demand, thereby ‘releasing’ those in being held less quickly.
If the intention of the system is to speedily fill the next element (e.g. loading a platform within a small train headway) then the positioning of the control line in the vicinity of the array may not support the operational objective.
Pedestrian and vehicle interaction
VSB site consideration will identify if the barrier design would adversely influence interactions between pedestrians and vehicles operating on site, and the additional risks caused to pedestrians due to this.
These risks are especially important for temporary barrier installations, and spaces where pedestrians are known to typically interact with vehicles during ingress or egress from an event.
As an example, care should be taken to ensure the operation of barriers that are manually opened and closed to allow vehicles on/off a site do not result in the unnecessary build up of pedestrian queues.
This may mean pedestrians choose to walk in areas beyond the protection given by the HVM, to avoid the queue, thus putting them at a higher risk from vehicles.
Previous studies
Previous work examined the impact of VSBs (specifically, permeable bollards) focussed on transport environments and exit movement1. Pedestrian flow rates were captured (i.e., the number of people who could pass through the bollards over time).
It was found that when dispersing from building exits, pedestrian flow was not significantly impacted by bollard arrays that were positioned at least 3m away from the exit, when there were gaps of 1.2m between the bollards.
Bollards placed closer to an exit (e.g., 2m away) may reduce the flow of people by up to 9%. Bollards placed within a corridor may reduce the flow of people by 8%, but this has not been systematically tested with realistic densities for large crowd flow.
Field observations at nine sites (including a sports stadium) found that these types of bollard arrays had only subtle impacts on individual pedestrian behaviour2. It was found that bollard lines running perpendicular to pedestrian movement had a lesser impact than those running parallel to movement and provided better use of space.
1Traffic Advisory Leaflet (TAL 01/16): The Influence of Bollards on Pedestrian Evacuation Flow, DfT and NPSA, Nov 2016 (as amended Oct 2017).
2Traffic Advisory Leaflet (TAL 2/13): Bollards and Pedestrian Movement, DfT and NPSA, May 2013 (as amended Oct 2017).