Whole Life Cost-Benefit Analysis for Median Safety Barriers - Executive Summary

Introduction

TRL Limited has been commissioned by the Highways Agency to examine the performance and cost effectiveness of median barriers installed on major roads in Great Britain. This report consists of two inter related parts - a review of median and median barrier accidents which resulted in casualties, and the relative costs associated with median steel safety fences and concrete barriers.  A detailed whole life costing survey for the most common types of safety fences and barriers currently in use on the median of major roads, is also included.

Executive summary

TRL Limited has been commissioned by the Highways Agency to examine the performance and cost effectiveness of median barriers installed on major roads in Great Britain. This report consists of two inter related parts - a review of median and median barrier accidents which resulted in casualties, and the relative costs associated with median steel safety fences and concrete barriers.  A detailed whole life costing survey for the most common types of safety fences and barriers currently in use on the median of major roads is also included.

In order to obtain detailed information relating to accident statistics and whole life costs associated with median safety barriers, one area of the Highways Agency's Network, the M25 Sphere, was be identified, and pertinent information obtained and analysed.  In addition a letter was also distributed to all fourteen of the Highways Agency's regional Traffic Operations Departments, their Agents and their Term Maintenance Contractors.  The letter was also sent to Britpave (representatives of UK concrete barrier manufacturers) and a similar letter sent to UK metal safety fence manufacturers.

Three distinct types of accident were investigated - those in which a vehicle crossed the median, those in which a vehicle was redirected across the carriageway, and those in which the vehicle remained close to the median after impact.  This was applied to accidents in which a vehicle struck either an object in the median or, in a more specific search, struck a median safety barrier.  This analysis was carried out for cars, HGVs and 'other vehicles'.  Data were extracted from the STATS 19 accident database for the years 1990 to 2002 inclusive.  The main findings were as follows:

When considering vehicle trajectories following an initial impact with a median barrier, it was seen in the STATS19 data that whilst a crossover accident is less probable than one in which a vehicle is rebounded or retained close to the barrier, it is almost three times as likely that a fatal injury will result. 

The data have also shown that when a car impacts a median safety barrier it is more probable that it will be contained, rather than breach it.  However, if the impacting vehicle is an HGV, it is more probable that the vehicle will be retained close to the barrier.  It is equally probable that an HGV will cross through a median barrier, than be rebounded by it.

For impacts by cars and HGVs, the highest percentage of serious injuries result from crossover accidents, the highest number of slight accidents resulting from those types of accident in which the vehicle remains close to the median barrier.

For impacts by 'other vehicles', the highest percentage of serious accidents comes from those in which the vehicle is retained close to the median barrier.  Slight accidents are more prevalent with the crossover accident.

An analysis of accidents occurring on the M25 Sphere has shown that:

  1. Of the 373 km of median safety barrier installed within the M25 Sphere,
    • Metal safety fences (including wire rope safety fencing)constitute 87.3% of the total length,
    • Concrete safety barriers contribute 12.7%.
  2. No fatal casualties have resulted from an impact with a concrete barrier;
  3. The number of serious casualties per kilometre is comparable between steel safety fencing and concrete safety barriers;
  4. Concrete barriers result in a lower rate of slight casualties and total accidents per kilometre than metal safety fences.

Following this accident data analysis, it was decided to compare the relative whole life costs (WLC) of different classes of safety fence and barrier.  The 'normal containment' safety barrier systems examined were the wire rope safety fence (WRSF), double sided and two parallel rows of tensioned corrugated beam (TCB) and open box beam (OBB) and vertical concrete barrier (VCB).  The WLC of barriers with greater levels of containment were also examined, these being the Higher Vertical Concrete Barrier (HVCB), the Dutch Step (concrete) Barrier and the Double Rail Open Box Beam (DROBB) safety fence. 

A total of seven items of cost were examined, these being safety fence and barrier installation, general maintenance, repairs (following an accident), removal costs, accident costs, and traffic management and traffic delay costs associated with any works to the safety barrier.  These were then consolidated on a whole life costing spreadsheet, to enable the WLC for 1km of each safety fence or barrier type to be calculated over a service life of 50 years.  After a period of 25 years it was assumed that the metal safety fences would be removed from site and replaced with an identical system. There were a number of items of cost excluded from the whole life cost calculations due to their complex and/or site specific nature.  These included the relocation of services (such as lighting columns and signs), the cost of consequential structural damage, and the costs associated with the complete closure of the carriageway during the recovery of vehicles.

Whole life cost calculation shave shown that the Dutch Step and Vertical Concrete Barriers have the lowest whole life costs. The Dutch Step Barrier also offers H2 containment, and hence, this increase in containment is provided with a reduced whole life cost.

Calculations have also shown that the introduction of a very high containment safety barrier would only be cost effective if there were 15 accidents occurring on the 1km length of median barrier in its 50 year life.  Furthermore, calculations regarding HGV impact only have shown that very high containment median safety barriers may be economically viable if 3 HGV accidents occur on a 1km length of barrier during a working life of fifty years.  However, this report has not examined costs associated with structural consequences (i.e. the costs associated with damage to the item of roadside furniture being protected [for example a bridge pier]).  In such cases, these structural consequences may be sufficiently high that the use of a very high containment safety barrier would be warranted.

Double sided TCB is seen to be the least expensive metal safety fence system, slightly cheaper than wire rope safety fence.

It was however, assumed in these calculations that the accident rates and severity would mirror those seen in the historical accident data ranging from 1990 to 2002.

Those safety barriers with a containment level greater than 'normal' will deflect less than a normal containment safety barrier when impacted with the small (900kg) vehicle as their rigidity enables them to contain heavier vehicles.  Hence it follows that the severity indices (such as ASI, THIV and PHD) for safety barriers with a 'high' or 'very high' containment level will be more onerous than those of normal containment.  (European studies such as the RISER and ROBUST projects are currently being carried out to try to correlate these severity indices with 'real life' injuries; however no link, to date, has been found).

Safety barriers with a greater level of containment are less likely to be breached.  In the case of central reserve safety barriers this may, in turn, reduce the number of fatalities occurring on the opposite carriageway.  However, if the barriers are of a higher and solid construction, a driver's visibility of activity on the opposite carriageway will be diminished.

Due to the rigidity of these high or very high containment barriers, it has been seen through full-scale testing that there is a general tendency for some vehicles with a higher centre of gravity (such as HGVs) to roll over the barriers if they are not of sufficient height. Whilst it is preferable for a vehicle to traverse the median on its wheels rather than on its side, a greater level of impact energy is required to roll an HGV than for it to breach a deformable safety fence.

During the life of a median safety barrier, there may be occasions which arise when legitimate access through the barrier is required.  Whilst this action can be quite speedily effected with a metal safety fence, this is not true for the concrete systems. In situations where concrete barriers exist, the use of removable/movable/ demountable sections of barrier would be necessary, or a detour around the barrier would be required.

Implementation

Following the investigations carried out under this commission further studies are recommended:

Quick Links

Find out how to make sure you're seeing the latest version of the site

Further Technical Documents

All the material previously located in the ‘Quick Links’ area of this page can now be found in the new ‘Further Further Technical Information’ section.