Network Rail has an ever more pressing need to manage infrastructure assets so as to minimise their whole life cost.
That clearly includes the cost of failures and the resulting train delays and cancellations as well as the cost of responding to incidents and remedying the damage they cause.
The company is working increasingly closely with its customers, and so it is also very aware of the non fiscal aspects of failure and disruption, such as the reputational damage to its image and the image of train operators, as well as the loss of custom to alternative modes.
Discussions about these matters go back to the Railtrack days and even earlier, when senior managers wanted to know how to avoid unexpected disruption from embankment failures on the WCML. There was no easy answer, given that there were far too many miles of known suspect banks for it to be economic to strengthen them all before they failed.
Today, Graham Birch is Network Rail Senior Asset Engineer (Geotechnics), based at Croydon. He and his colleagues have been carrying out important work to improve the management of railway earthworks assets, in the south-east of the country in particular.
His team is now applying measures which, for his area of the country at least, are beginning to offer ways to economically monitor and manage embankments and cutting slopes.
The first objective is to ensure that likely trouble spots are identified and remedial action is taken before a damaging failure occurs.
The second is to have in place monitoring systems on suspect sites not yet rectified that can give warning of failure before it causes an incident with traffic, ensuring rail safety. Third is the need to understand the reasons for instability so as to develop the most cost effective remedial strategy for each site.
It is necessary to understand the geological history of the area concerned before one can begin to understand the earthworks constructed there. The history of the earthworks themselves is equally important.
The natural materials exposed in cutting faces when the railways were built were laid down millions of years ago. The railways in the UK were built around 150 years ago, so the cuttings were dug and the embankments built at that time.
Consequently, there are marked differences in geotechnical properties, and thus engineering behaviour, between the materials exposed in the cutting faces and those within the embankments.
In this country, there was opposition to the construction of railways from wealthy landowners and operators of the existing canal system.
This caused difficulty in carrying out topographic surveys of the routes, requiring, in some cases, the employment of prize fighters to protect the survey teams, and even necessitating surveying by moonlight to avoid gangs of objectors.
These constraints, in conjunction with the pressure to minimise land-take, resulted in side slopes being overly steep on embankments and in cuttings. A further consequence is in the necessity for diversion of the lines from the optimum routes to skirt around estate boundaries; Hatfield Curve and the bend in Oxted Tunnel may both be examples, but there are many others.
Embankments were usually built by end tipping the material arising from the nearest cutting with no prior surface preparation. The fill was therefore a random mix of whatever came along, placed without consolidation onto an inadequate base.
Not surprisingly many banks failed during construction, and those that didn’t often still give trouble today in poor track geometry and worse.
UK railway builders were pioneers, and in many ways they were learning as they went along; in engineering, and in managing the political climate, the challenges were novel and errors were made.
When construction moved abroad the climate was generally politically more favourable; the railways were able to use more land and choose better alignments. Lessons had been learned about the engineering too, and so better design and construction standards were possible.
The surface physiography is dictated by the underlying geology of the area. In the UK, south and east of a line roughly between the Severn and Humber estuaries, over-consolidated clays have a significant impact on the earthworks.
These clays are sensitive to moisture change and respond by shrinking, as they dry, and swelling when re-wetted. London Clay and Weald Clay are the most widespread, but Gault Clay is the most moisture sensitive.
The orientation of rail lines in SE England relative to the geological structure or ‘grain’ is a factor in determining how likely it is that a route will be affected by the moisture sensitive clays.
Wessex’s primary routes run southwest-northeast and are largely clear of the clays, but Kent’s primary routes are heavily susceptible because they run east-west, parallel to the axis of the Wealden Anticline, and tend to stay on either chalk or clay throughout. Sussex’s primary routes run north-south, roughly at right angles to the grain and so cross the clays quickly and pass onto sounder geology.
The winter of 2000/2001 was exceptionally wet, causing 160 failures of earthworks in the South East. Initially it was the clay embankments that failed due to excessive moisture content but, as the rainfall continued, cutting slopes outside the clay areas also began to fail. This period was effectively one of ‘destructive testing’ and provided fresh insight into the preparatory processes and triggering factors in earthworks failures.
GISmos and chains
Asset management of Network Rail’s earthworks begins with examination in accordance with Standard NR/L3/CIV/065, Examination of Earthworks. The examinations are carried out by consultants using hand-held electronic data gathering devises linked to satellites and loaded with bespoke software such as “GISmo” (Geographical Information System Mobile).
Examiners use drop-down menus to ensure that the data gathered is as objective as possible, so they can make like for like comparisons between routes.
Every five chain lengths (~100m) of route is categorised by condition as Serviceable, Marginal or Poor. Those in the first category are re-examined every 10 years, and those in the second are re-examined every five.
Poor earthworks are subjected to further evaluation, to confirm the consultants’ score, and then prioritised to determine which sites need to be passed on for monitoring or remediation. This is how the business plan is populated.
LiDAR (Light Detection And Ranging) surveys are used to obtain detailed surface contour data and subsurface information is obtained by Ground Investigation using state-of-the art slope-climbing rigs for sampling the steep faces of earthworks, and road/rail truck-mounted Cone Penetrometer Test rigs for drilling through the 4-foot, such as that operated by Lankelma, which is capable of doing a 20 metre test in half an hour.
Monitoring of Poor earth structures may use conventional techniques, such as boreholes equipped with piezometers and inclinometers, but more sophisticated means of obtaining real-time condition data remotely are under development.
Especially steep slopes can be instrumented to give real time warnings of hazards as well as to provide engineering data. Pull wires attached to slope netting are used at Hooley Cutting, Sussex, to detect the accumulation of debris behind the containment netting.
Alarm thresholds are set to alert selected managers to an incident via the mobile phone system and webcams can be checked from any computer as a back up.
Between Folkestone and Dover in Kent, where the railway runs along the seaward side of the 150 metre high Chalk sea cliffs, there are signal wires attached to a rockfall detection fence which automatically set the signals to danger in the event of a cliff fall incident.
Trees hold up banks – don’t they?
Vegetation has an important influence on earthwork behaviour. Contrary to common belief, trees do not hold up the banks.
Whilst this may applicable to natural slopes or engineered highways earthworks, it is not applicable to the over-steepened cutting and embankment slopes on the railway infrastructure. Field trials on clay cored embankments have demonstrated ground disturbance to be 10 times greater where trees are present compared to grass.
Also, tree roots typically penetrate five times further than those of grass. Given that clays shrink and swell in response to seasonal variations in moisture content, it is easy to appreciate how the presence of trees can exacerbate the effects of seasonal moisture variations, in particular at the desiccation part of the cycle when the presence of trees can give rise to poor track geometry and ultimately speed restrictions.
Vegetation has other detrimental effects for railways, particularly in the case of trees on cutting faces, as we well know. It is thus important to manage vegetation, remove trees from slopes and discourage their re-growth.
Climate change is tending to increase the number and potential severity of weather related incidents. Dryer summers, wetter winters and more days of heavy rain clearly imply more problems with earth structures.
As mentioned before, weather has a major influence on earthworks. Clay behaviour is affected by its moisture content and the greater the deviation from normal, the greater is that influence, be it extremely wet or extremely dry.
Of particular interest is the relationship between rainfall, Soil Moisture Deficit (SMD) and earthwork behaviour. Rainfall information is obtained from the Meteorological Office and compared to the local Long Term Average (LTA).
The SMD parameter was originally developed by the Met Office for agricultural use. However, research at Imperial College into the failure of London Underground’s clay embankments highlighted the potential for its use in monitoring the condition of clay-cored embankments which are subject to seasonal shrink-swell processes.
“Earthworks Watch” is a system developed by Graham and his colleagues to inform asset managers, maintainers, emergency response contractors and others in the SE area about the likely condition of their earth structures.
It allows them to better understand their assets and to plan and respond more effectively to likely changes. It is based upon 12 years of monitoring in the south east of England by Network Rail, and is proving an effective management tool.
The system looks at five key variables:
- Asset type (cutting…at grade…embankment)
- Geology (granular/rock…cohesive/clays)
- Condition (serviceable…marginal…poor)
- Moisture content (SMD) (saturated…normal…desiccated)
- Vegetation type (trees…grass).
Rainfall is also recorded relative to the long term average for the site. The system presents users with three indicators: ground condition, condition trend and earthwork response. These are presented on the Network Rail portal weekly in map form and monthly as a full graphical display of the data.
The system is, in Graham’s view, of greatest benefit south of the imaginary line between the Severn and Humber estuaries as elsewhere the bedrock types are different and not generally susceptible to the same sort of analysis.
However, Hydrologically Effective Rainfall (HER), a measure of excessive water after the SMD value has reached zero (i.e. saturation), is a further parameter which is being assessed for its potential to predict the propensity for flooding and scour in the non-clay materials prevalent elsewhere in the UK.