INTRO: Field tests are to start shortly into methods of improving the stiffness of the subgrade without disrupting commercial services. The Isert project is being carried out by the University of Birmingham with Railtrack and a group of suppliers. Methods to be tested include mixing or injection of chemical stabilisers, densification and novel drainage technologies

BYLINE: Matthew Brough MEng MSc Eng PhD

Research FellowSchool of Civil EngineeringUniversity of Birmingham

SIGNIFICANT savings in long-term track maintenance costs are promised by research now under way in Britain. The Isert project (Improving the Stiffness of Existing Railway Track) aims to develop a better understanding of the relationship between the design of track substructures and subsequent maintenance.

Although excessive differential settlement can be expected on poor formations, and is accounted for when planning maintenance, differential settlement and subsequent geometry deterioration is often under-predicted by current theoretical models.

Relating track parameters to life-cycle costs was one of the objectives of the Eurobalt programme (European Research for an Optimised Ballasted Track, RG 12.00 p813). This focused on understanding and measuring track stiffness and settlement. As a result, it is now possible to perform an optimisation study and determine optimum track parameters critical to long-term maintenance requirements. Consideration was also given to methods of remedying poor formations where the existing track design is sub-optimal.

The study concluded that variations in formation properties and condition are the most significant factors governing stiffness and deterioration of vertical geometry. Integrating this finding into previous track geometry deterioration models showed a dramatic improvement in modelling accuracy. It also confirmed the existence of an optimum stiffness, at which maintenance costs are considerably reduced. Isert also assessed formation and substructure treatment methods, and identified several that may prove beneficial 1.

Due for completion in September 2002, Isert focuses on track stiffness, the load:deflection ratio resulting from the combined resistance of the rail, ballast and subsoil to loading, and its importance in predictive models of rail deflection and subsequent geometry deterioration, aiming to devise a method for assessing and improving subgrade stiffness without disrupting services.

Isert is the second Epsrc Link project to be undertaken by the Railways Research Group of Birmingham University’s School of Civil Engineering under the Inland Surface Transport programme. It is jointly funded by the British government and industrial partners Railtrack, GTRM, Scott Wilson Pavement Engineering, WS Atkins Rail, Keller Ground Engineering, Fugro and Serco Railtest.

Track settlement problems

Full-scale trials are being undertaken to confirm that a consistent level of subgrade stiffness results in lower maintenance costs. These trials form the main focus of Isert, applying existing and modified soil strengthening techniques without initial track removal. Long-term changes in stiffness, track geometry and formation condition are recorded, using measurement techniques such as falling weight deflectometry, ground penetrating radar, automatic ballast sampling, cone penetration testing, continuous surface wave systems and other hand-held devices 2. The best means of routine measurement needs to be identified, so that long-term in-situ remote monitoring can take place accurately, consistently and safely.

Repeated dynamic loading, fine-grained soils and high moisture content all contribute to problems with the subgrade directly beneath track components. Progressive shear failure (the plastic flow of soil caused by excessive repeated loading of the subgrade interface) leads to heave at the trackside and ballast pockets. Excessive plastic deformation caused by soil compaction and progressive consolidation causes non-uniform settlement. Finally, high moisture content and repeated loading of the interface leads to attrition and mud pumping, contaminating ballast and causing drainage problems 3.

These problems (Fig 1) are recognised as direct causes of differential settlement, erosion and the necessity for repeated maintenance 4. Differential track settlement can exacerbate underlying subgrade problems, leading to a rapid deterioration of track quality in the direction of traffic. Removal of the causes is the first step towards formation improvement.

Three test sites are being established, and traditional ground improvement techniques used to solve different trackbed support problems. Site investigation has commenced at a possible weak formation clay subgrade site at Leominster.

Current solutions

In order to select techniques suited to in-situ optimisation of formation stiffness within the physical and time restrictions of the British rail network, traditional subgrade improvement methods must first be assessed. Subsequent problems can then be fully accounted for in any techniques to be implemented in the field trials.

Traditionally the formation and subgrade has not been targeted in works schedules, and track components are maintained with little regard for formation properties and subsequent effects on continued asset life. There are many examples where expensive remedial work has been carried out without eliminating the initial causes.

When acceptable track geometry is difficult to maintain due to a weak or unstable subgrade, repeated ballast addition is commonly used as a short-term preventative measure 5. Aiming to lessen the dynamic loading in the formation layer, it compensates only temporarily for settlement problems. It can also exacerbate ballast pocket and water entrapment problems, leading to frequent maintenance.

Alternatively, a sand blanket and geotextile (SBG) can be installed. The sand blanket acts as a filter barrier and increases shear resistance, whilst the geotextile directs water away, preventing slurry migration and ballast contamination 4. SBG, with proper drainage established, has proved to be an effective remedial measure at poor formation sites. However, installation requires the costly and time-consuming removal of track components. Furthermore, there is scepticism about its performance in an industry facing increasingly stringent restrictions on possession times and demands for higher loads and line speeds.

Grounds for improvement

Economic analysis of tamping on soft formations suggests that modified substructure designs could be widely applicable, provided that sufficiently cheap, effective treatment methods can be devised 6.

There is a range of possible ground improvement technologies (Table I). These can be categorised into six types: densification; consolidation; reinforcement; chemical stabilisation; thermal stabilisation; and biotechnical stabilisation.

Each technique has been assessed for its application to a railway environment and its suitability for in-situ use. This theoretical study focused on the application of the technique without ballast removal, the possibility of interference with track components, and any technology-specific problems. Although excavation, modify and replace (EMR) methods such as admixtures are not in-situ techniques, they could be integrated into renewals or maintenance works where the removal of track components cannot be avoided 7.

Compacting and grouting

Following the initial study, grouting, slurry pressure injection, chemically-stabilised soil mixing, vibrocompaction, stone columns and soil nailing have been identified as the most suitable ground improvement techniques. Although not traditionally classed as a ground engineering technology, for completeness drainage was also considered.

Vibrocompaction and stone columns have been used in Germany. This requires the complete removal of track components and ballast, and the technique would be more applicable for granular soils rather than the cohesive soils typical in Britain.

Grouting is the injection of materials such as lime, cement, fly-ash, silicate or chemical stabilisers into soil to change its physical characteristics. Slurry pressure injection (SPI) forces slurry into voids or fissures in the soil. Lime slurry pressure injection (LSPI) was used extensively to treat expansive clay soils on US railways during the 1970s using multi-probe injection rigs (Fig 2a).

The use of substandard fly-ash and incorrect procedures attracted criticism to the method 8. Industry restructuring led to a loss of expertise, allocation of funding elsewhere and a reduction in LSPI use. Today US attention is focused on cement slurry injection, partly as a result of insufficient data from many successful LSPI case histories. More recent applications of LSPI in Australia have demonstrated its success as a technique for reducing overall track maintenance costs 9.

Isert has looked at critical assessments, limitations and possible solutions, concluding with a review paper highlighting possible plant and procedure modifications to overcome some of these limitations.

SPI could prove applicable if grout mixes are used for stabilisation and modification. British applications could be possible if many of the limitations associated with the current design rationale, slurry distribution and post-injection performance and acceptance testing can be overcome. Distribution control and breakout is a major concern, with perceived slip failure and grouted ballast or drain contamination problems. Examples of slurry breakout and grouted ballast are shown in Figs 2b and 2c.

In-situ mixing of soils with cement grout or other chemical stabiliser slurries using multiple-axis augers creates overlapping stabilised soil columns 10. By changing the column configuration, this can be used to strengthen soft soils or for groundwater and liquefaction control. A range of mixing tools could prove suitable, but some could be subject to installation and operation problems due to the physical working restrictions between sleeper cribs and ballast penetration resistance.

Auger rig penetration load, ballast contamination with grouts, ballast disturbance, track heave, stabiliser distribution and early strength development are major issues of concern when considering trackbed improvement using soil mixing augers. Consequently an auger design has been chosen to overcome these problems, with modified flat blade and standard helical and paddle mixing augers demonstrating the best potential for development.

Laboratory trials of modified and purpose-built flat blade augers are currently being undertaken to assess their operational limitations through ballast. The results suggest penetration resistance is within the operational limitations of rigs. Ballast heave and breakdown is also minimal. Installation and operating performance will be assessed further in pilot trials in a clay bed overlain with compacted ballast. Grouts used will be chosen on the basis of testing using clay taken from the Leominster site and a range of binders.

Following pilot-scale trials and a subsequent on-track safety case review, a live trial will be performed at Leominster. Following implementation, a long-term monitoring programme will be set up to measure changes in stiffness, track geometry and formation condition.

Drainage implications

Investigations at Leominster showed that the shear strength of the subgrade should be sufficient to resist the applied dynamic loading. It was the interaction of water, ponding and subsequent softening of the upper formation layer that was cause for concern. A range of shallow drainage systems will be installed in situ beneath the track components later this year. Long-term monitoring will be continued until the completion of the project next year.

Further trials will be conducted at other poor formation sites across Britain, and proposals have been made for formation treatments to be integrated into existing maintenance work where ballast excavation is unavoidable.

Details of the Railways Research Group, partners and relevant links can be found at

TABLE: Table I. Traditional ground improvement techniques

Improvement method Treatment type Treated soils Main applications

1. Admixture (cement) Ch Sand, collapsing silt B, St, Cc

2. Admixture (lime) Ch Plastic or swelling clay B, St, Sw, Cc, Pr

3. Cement grouting Ch Coarse soils, rock B, St, Se

4. Chemical grouting Ch Fine sand B, St, Se

5. Jet grouting Ch, Re All soils B, U, Se, L, St, En

6. Deep soil mixing Ch Fine-grained soils St, E, Se, En

7. Lime columns Ch Saturated clay Se, En, St

8. Preloading/drains Co Saturated clay B, St, En

9. Electro-osmosis Co Saturated clay B, St, L

10. Vibrocompaction De Sand and gravel B, St, L

11. Dynamic compaction De Granular soils, fills B, St, L, En

12. Blasting De Saturated sand B, St, L

13. Compaction grouting De Loose soil with voids St, Ex

14. Mechanically stabilised fill Re Granular soil B, St, Sl

15. Soil nailing Re Granular soil, clays Sl, Ex

16. Micropiles Re Granular soil, clays U, Sl, En, B, St

17. Stone columns Re Clay, liquefiable soil Sl, St, B, L

18. Bio-stabilisation Bi Fertile granular soil Sl, Ex

19. Heating Th Loose, collapsible silts Sl, B, En, Cc

20. Freezing Th Underwater sand Ex, Se, En

Bi = Biotechnical; Ch = Chemical; Co = Consolidation; De = Densification; Re = Reinforcement; Th = Thermal.B = bearing capacity; Cc = collapsing control; En = Environmental control; Ex = Excavation support; L = Liquefaction resistance; Pr = Plasticity Reduction; Se = Seepage control; Sl = Slope stability; St = Settlement control; Sw = Swelling reduction; U = Underpinning.Source: Adapted from Munfakh 10

Output objectives

Guidelines are to be produced that relate ballast, sub-ballast and soil conditions with remedial techniques and application methods. These will be established according to three principal output objectives: