F. J. Baynes, consulting engineering geologist, Perth, Australia
This brief case history is of the investigation, design, construction and operation of some of the World's major iron ore railways during the last two decades in the ancient cratonic region of Western Australia.
The overall geo-approach was based on the development of a thorough understanding of the geological and geomorphological history of the area and adopting strategies such as staged investigation; definition of investigation objectives; investigations to answer specific questions; an emphasis on geo-mapping; establishment of reference conditions; the development of geo-models and the application of the observational method.
In general, this approach is not restricted to railways. Other linear structures that can traverse wide ranging geological conditions, such as tunnels, roads and pipelines as well as projects involving extensive ground engineering (hydroelectric schemes for example), have been built efficiently using similar approaches.
Around the world, only a few countries have the long and detailed history of geological mapping by a national survey, together with records of numerous site investigations and other ground activities, which give considerable foresight and the ability to develop fairly realistic models before a foot has been set on the ground. This approach is well discussed in detail by Culshaw (2005). The method outlined in this paper is the basic scientific approach that creates a conceptual model indicative of likely ground conditions and then tests it by the ground investigation.
This approach applies where there is no subsurface data, as on many sites in remote locations, but it also equally applies where there is a mass of subsurface data, for example as described by Culshaw (ibid). However, it is important to note that whatever geological (and geomorphological) observations are available, such as mapping and subsurface information, the data always has to be presented within a geological interpretation. This is where a competent geologist uses knowledge and experience to create surfaces in three dimensions (and sometimes the fourth dimension of time) that bound the data. It is the combination of geological observations and geological interpretations that go to form the Total Engineering Geological model. The significance of these distinctions should not be underestimated, specially when involved in contractual matters or litigation.
The advantages of the approach outlined in this case history was that it was particularly relevant to the fairly rugged terrain with little background information apart, perhaps, from large scale photo or satellite imagery. The geological / geomorphological knowledge was acquired more quickly and more cost effectively and better decisions were made about ground engineering issues which are often a principal source of project risk. There was also a more equitable and transparent sharing of ground engineering risk between owners, designers and contractors. This meant less conflict and easier project implementation - but greater flexibility and management effort was required.
The iron ore mined in the Pilbara is destined for export and requires transport to shipping ports on some of the longest and heavily loaded railway trains in the world. Ore trains can typically consist of up to 240 wagons and can be 2.6km long with each wagon carrying about 130t of iron ore. As a result, each wagon imposes individual axial loads of about 35t onto the railway track.
The track has a gauge of 1.435m and is supported on precast concrete sleepers 650mm apart, placed on a 200mm to 300mm thick layer of 40mm ballast. Ruling grades for trains are typically an absolute maximum of 0.5% and the system is currently estimated to consist of about 1400km of track.
The central Pilbara is a semi-arid upland with a base elevation 700m above sea level and with ranges of hills that extend up to a total 1km height. The relief leads to the development of a variety of active slope processes that can impact on the railways, such as debris flow and rock fall. Summer day temperatures range well into the 40°Cs and winter frost can occur.
Annual evaporation is well in excess of annual rainfall which averages from 180mm to 350mm, but it is highly irregular due to periodic cyclonic events and thunderstorm cells that produce localised intense falls. Rainfall up to 200mm in 24h can occur during the cyclonic season. Such events exert a major control over current slope process rates and the occurrence of damaging floods.
The design of the ground engineering (box 1) and related aspects mainly involved:
- Optimisation of the horizontal and vertical alignment and the associated excavation and earthwork volumes as these represent
the principal cost variables;
- Definition of construction materials both from cuttings and borrow areas;
- Design of cut and fill slopes, bridge foundations and drainage to achieve target performance.
Inputs from the geo-team into the ground engineering design includedthe choice of cutting slopes angles, excavation and borrow assessments,and assessments of any geo-hazards.
At the design stage both the owner and the engineer wanted to refine the cost estimate to ±10%. This was seldom achievable for the ground engineering components because of the large degree of uncertainty that still existed. However, the cost uncertainties associated with non-geological aspects of the project, such as steel, concrete sleepers and many of the fixed costs, could be evaluated much more accurately, often to considerably less than ±10%. This meant the overall inaccuracy of the cost estimates by the ground engineering components were typically balanced by the better accuracy of the non-ground engineering components to achieve the target.
Engineering design occurred towards the end of the main investigation phases and included the finalisation of the documents for the contractors. At this stage the considered results of the geo-team's work was generally summarised into reports.
Geology and geomorphology - Bedrock
During 2003, ore production was about 180Mt, representing around 9% of the world's total production and about 35% of the seaborne
trade (limited to high grade production). Deposits comprised ancient Precambrian bedded ores, mid-Tertiary fluvial channel ores and late Tertiary colluvial fans (box 2).
Landform evolution governs the distribution of superficial deposits, weathering profiles and slope processes and thus exerts a profound influence on the engineering characteristics of the near surface materials in which the railway earthworks are built. Understanding this evolution requires an understanding of the regional tectonic history of the area as this has controlled palaeo-latitudes and rates of uplift, and so past climates, weathering processes and cycles of erosion.
Table 1 summarises the total engineering geology/geomorphology history of the Pilbara and demonstrates the engineering value of the landform evaluation approach (box 3).
The geo-team approach and work done - Aims
Finance is generally limited at the pre-feasibility or initial design stages of any project and the collection of low cost information that will support very broad decisions has to be the main objective. This means investigations during these stages usually involve desk studies, studies of aerial photographs, site reconnaissance, ground truthing of aerial photographic interpretation by site visits, geo-mapping and field inspection.
At this stage, in the Pilbara, the alignments were effectively target corridors of several tens to hundreds of metres wide. Investigations involving detailed mapping, test pitting and drilling of boreholes to obtain information in deep cuts, laboratory testing, and so on, are far more expensive and in the Pilbara were only carried out as the alignments became more fixed and greater detail was required for design and costing.
A high level of geomorphological and geological understanding was not sufficient for successful project implementation of the various railways. Geological information had to be delivered to the right people in the project team at the right time, in an easily understandable manner with an explanation of the significance of the information. The use of the Total Engineering Geology Approach (Fookes, Baynes and Hutchinson, 2000, 2001) in which different types of information were collated and communicated to each of the schemes, proved an effective contribution to the railway projects. Table 2 illustrates typical objectives and activities at each stage.
The overall aim of the investigations was to evaluate the ground to the level where unforeseen project scale ground conditions were unlikely to be encountered. Some balance had to be established between expenditure on investigations and resulting benefits of the information obtained. As such:
- All of the observations were presented, using a standard descriptive system to minimise confusion among design or construction engineers (in prac tice AS1726 (1993) on site investigation, supplemented by other interna-
tionally well known standards);
- Emphasis on non-written information, for example colour photographs, maps and drawings, so that those with little or no understanding of the engineering geological/geomorphological descriptive terms could gain appreciation of the ground conditions by looking at the photographs or models. This was particularly helpful where the information related to machine performance such as ripping trials and backhoe excavations;
- All information relevant to the projects was assembled and made availableeither as reports provided to prospective tenderers, or assembled in a roomfor viewing by prospective tenderers (cf. The Guidelines of the Construc-tion Industry's Committee, 1987);
- A clear distinction was drawn in the reports between observations, inter- pretations and opinions, to minimise confusion as to the nature of infor-mation being provided. All of the site investigation was compiled into afactual report which consisted of observations but not interpretations anda separate interpretations/evaluation report in which interpretations and opinions were documented. Both kinds of reports were provided to pro-spective tenderers.
Geo-mapping was not a significant component of the investigations of the older (pre-1990's) Pilbara railways where efforts were directed immediately to more costly subsurface investigations, principally lines of boreholes on the assumed route. This has been shown elsewhere in the world to be extremely ineffective and is typical of the traditional approach adopted on many heavy civil engineering projects throughout the world.
It is particularly ineffective in investigating linear structures such as railways, roads and pipelines. This is because carrying out subsurface investigations too early in such projects almost inevitably results in having to carry out additional expensive subsurface investigations due to changes in alignment required by the ongoing design.
In contrast, early geomorphological and geological mapping of a route corridor is far less expensive and far more productive as it involves no expensive plant and is more likely to provide information relative to whatever final alignment is chosen. Geo-mapping can be regarded as the essential and defining strategy of the Total Engineering Geology Approach (box 4).
The different mapping units were logically differentiated in the field on the map legend or logs using the process of division described by Varnes (1974). The mapping units provided the framework for systematically making observations, collecting information, development of models and the reference condition.
Pilbara is a cratonic area underlain by massive Archaean granites and gneisses around 3bn (3Ga) years-old overlain in the south area by volcanics and sediments of about 2.5Ga on the eroded cratonic surface during the late Archaean and Proterozoic. Government geological maps on 1:250,000 scale exist.
A distinctive characteristic of the bedrock iron ore of the Proterozoic Hamersley Group is a lateral stratigraphic continuity of the BIF (bedded iron ore formations). These can extend over hundreds of kilometres which, during regional mapping, allowed the recognition of individual marker horizons and conspicuous landform patterns associated with specific formations. These formations, which have been gently folded, control the landscape development with the more resistant BIF forming a range of hills with distinctive curestas or mesa shapes. Broad valleys are typically associated with ancient shales and on occasion limestones subject to karst development. Thick dolerite sills of the
Capricorn orogeny intrude the area which has been subject to medium grade regional metamorphism.
Most bedrocks have high uniaxial compressive strengths together with a presence of through-going discontinuities of all types. The so called shales, which have been sheared during folding, are weaker and more easily weathered and can have bedding planes with the residual shear strengths as low as 15° and zero cohesion. The wide range of engineering characteristics is further accentuated by the influence of geological structures such as faults and joints and in particular by the development of deep ancient weathering profiles.
Five different fold phases have been identified of which two were important for the railways. One consists of open large scale regional fold styles leading to dips of 10° to 30°. The other consists of much tighter localised folds, which are restricted to structural corridors, possibly associated with regional faults or intrusions. These localised folds in particular can lead to small scale local instability due to adverse dipping planes.
Fault styles are mainly low angle thrusts associated with regional compression and are rarely encountered in railway excavation. Joints largely reflect the structural attitude of adjacent folding and faulting. Steeply dipping persistent master joints typically exert a fundamental control on the cliff line stability, particularly where the rock mass is prone to toppling failure, and characteristically form a saw tooth pattern on cliff lines.
In essence, the Pilbara was probably glaciated during the Permian when it was part of the polar super continent, Gondwana. Rifting started during the Jurassic to form the continent of Australia and during the Mesozoic, progressive erosion produced an extensive relatively planar land surface cutting across the Craton.
At the end of the Mesozoic, the climate is believed to have been warmer and more uniform globally in comparison with the present, due to the restrictive oceanic circulation. Despite being far south at this time, the Craton was subject to several cycles of deep chemical weathering and the formation of duricrusts that eventually led to the formation of the prominent Hamersley surface, remnants of which can now be traced across the Pilbara.
Slopes locally reflect the underlying lithology and structure and the overprinting of different phases of weathering, duricrusts formation, erosion and deposition. The upper parts of the slopes are often remnants of former erosion surfaces and usually include rounded duricrusted portions or flat surfaces subparallel to the bedding.
Below the upper slopes, steep cliff lines controlled by persistent subvertical joints are often developed due to rejuvenation of the landscape by uplift and incision. Cliff lines are actively degrading with aprons of rockfall debris forming transportational mid-slopes below the cliffs and at the exit point of gullies and canyons there are widespread alluvial/colluvial fans and slopes mantled with such debris.
Towards the base of the slopes both alluvial deposits and canalised debris flows form low angle fans. Within broad valleys there are alluvial flats and intermittent flood prone river systems. Simple example figures, based on the Observational Method, were used particularly for explaining geological/geomorphological evaluation and the distribution of materials for engineers.
CLICK HERE FOR TABLE 1
The use of reference conditions (RC) was another essential strategy (CIRIA Report, 1978) which consists of identifying groups of geological/geomorphological materials of similar engineering characteristics with a depiction of the full range of geological/geomorphological conditions that could be reasonably anticipated or foreseen in the project area. The RC were used to describe and communicate the geological and geomorphological conditions to the project engineers.
In contrast to the process of division used to establish the mapping units, a process of grouping was used to establish the RC. Table 3 gives an outline of selected reference conditions, the geomorphological input being particularly important in the first four conditions (see Baynes, Fookes and Kennedy (2005) for more detail). Table 4 gives an example of one of the specific reference conditions (blocky BIF outlined in Table 3) in detail, and Table 5 is an example conceptual design related to reference conditions and shows the precedent designs for cut batters.
The engineering and geomorphological maps produced with the objective of differentiating geological units with characteristic engineering behaviour (Fookes, 1969; Dearman, 1991) required the identification of mapping units at four different scales in the Pilbara for which the following terminology was adopted:
- Terrain units produced with a high geomorphological input consisting of distinct assemblage of bedrock, superficial deposits and landforms with recognisable engineering characteristics mapped to the scale of 1:50,000 to 1:250,000;
- Engineering geological formations mapped at 1:5000 and 1:50,000;
- Engineering geological members mapped at 1:1000 and 1:5000;
- Engineering geological types which are units that have effectively homogeneous characteristics and were described using standard descriptive terms, usually in logs of boreholes, test pits and exposures.
The definition of the RC was also a part of developing the models that were central to the understanding and communication process. In most projects, the geological/geomorphological models are based upon observation or sampling of only a small part of the ground and alternative interpretations may be possible. Where a project is to be constructed under contract and where there is uncertainty in the level of understanding particular features of the ground, which could have significant influence on the contractor's method or cost, the use of RC indicates the assumptions that have been made for the basis of the contract.
The models that were developed for the railway projects took the form of simple geological and geomorphological maps and sections, combinations of geological, geomorphological and geotechnical information in engineering geology mapping sections, evolutionary diagrams and three dimensional block models. The block models were presented to show three different kinds of information. These were conceptual models (for example figure 1), observational models (for example figure 2) and evolutionary models.
CLICK HERE FOR TABLE 3
The Observational Method
It was not possible to completely investigate every detail of the geology and geomorphology of any of the railway routes prior to construction for logistical and financial reasons so the design and attached engineering cost estimates were always necessarily based on limited information. To overcome the inherent uncertainty that will always be present in any ground engineering, the Observational Method was routinely adopted during railway construction.
The method was found to be particularly useful for the following:
- The short time frame within which many of the projects had to be imple- mented, for example some railways of several hundred kilometres of heavy duty track had to be completed within three-years from commence-ment of the feasibility study.
- The cost benefit of carrying out investigations prior to project approval, i.e. the investigation of many tens of deep cuttings with boreholes in ruggedcountry would have taken a long time and cost several million Australian dollars - time and money that were not necessarily available.
- An awareness of the relative ineffectiveness of any subsurface informa-tion technique, when compared with full scale excavation, in finding what was actually in the ground. While 20m of good quality 83.1mm diameter oriented PQ3 core can be extremely useful, it is very expensive to acquire and can never provide as much geological detail as a 20m deep, 100m long cutting at the same location.
CLICK HERE FOR TABLE 4
The human resources required
It is always difficult to estimate how long it will take to carry out types of studies that have been described in this case history. Table 6 was based on a number of different projects during the 1990s and 2000s. It indicates in very broad terms the typical number of experienced engineering geologists and approximate time it took for the different project stages involved in building some 50km to 100km of railway across relatively flat ground. Similar resources were required for about 10km to 30km of track across more deeply incised hilly country.
The engineering geologists concerned each had a minimum of ten years' experience and had worked in the Pilbara on a number of occasions. All had geomorphological experience although were not trained geomorphologists. On occasions, specific geomorphological input was required and a professional geomorphologist was hired as appropriate, as were specialist structural geologists, geotechnical earthworks engineers, blasting consultants and so on. Table 6 illustrates the typical resources other than the specialist input.
In addition, there was a regular specialist international review as indicated in Table 2, principally by a very experienced engineering geologist and by very experienced consultant bridge engineers. Local reviews to national standards were also carried out on occasions. The typical review period by the international consultants was about four or five days on site and a week locally in report preparation.
In very round terms, construction of individual railway projects was typically spread over about three years. This included a year or more in investigation and design, and a year or more in the run up to construction and active construction. Very round costs of each railway project at prevailing prices (about ten years ago) were about AUS$100M of which probably investigation costs were less than 2%. All of the projects were, give or take a little, within budget, on time and cost, and were successfully implemented. The two projects using the Total Geology Approach both won various Australian Engineering and Railway Excellence Awards.
The events described in the case history are an amalgam of three or four projects of different sizes but all in very generally similar Pilbara terrain. It can not be said that the investigation, design and construction processes were all identical but they were similar, and there was a longer learning curve in the investigation procedures for the older projects, and less for the newer projects with experience accumulated from each successive project.
During the design and construction, the principal areas requiring a geo-teams input were aggregate soundness, blasting and excavation, borrow areas, cut slope stability, danger to health fibrous minerals, environmental issues (which were given a high profile), fills/embankments, floods, granular collapsing soils, karst, reactive clays, seismic risks and settlement due to dewatering.
As an estimate, between a quarter and a third of the geo-team effort was probably directed towards geomorphological understanding and evaluation. It is considered that the Total Engineering Geology/Geomorphology Approach can be best used in all ground engineering works if it is appreciated that it is a practical philosophy that aids and promotes successful engineering in the face of geo-uncertainty. In a nutshell, communication is the essence of success with the early implementation of the Observational Method, Stages and Reference Conditions and with the use of experienced geo-personnel.
Baynes, F., Fookes, P.G., Kennedy, J.F., 2005. Total engineering geology approach to applied railways in the Pilbara, Western Australia. Bull. Eng. Geol. Environ. 64. 67-94.
CIRIA, 1978. Tunnelling - Improved Contract Practices, Construction Industry Research And Information Association Report 79, CIRIA, London.
Construction Industry Committee, 1987. Guidelines for the provision of geotechnical information in construction contracts. The Institution of Engineers, Australia. pp. 19.
Culshaw, M.G. 2005. From concept towards reality : developing the attributed 3D geological model of the shallow subsurface. Quarterly Journal of Engineering Geology and Hydrogeology. 38. 231-284
Dearman, W.R., 1991. Engineering geological mapping. Butterworth-Heinemann. pp. 387.
Fookes, P.G., 1969. Geotechnical mapping of soils and sedimentary rocks for engineering purposes with examples of practice from the Mangla Dam Project. Geotechnique 19. 52-74.
Fookes, P.G., Baynes, F.J., and Hutchinson, J.N., 2000. Total geological history: a model approach to the anticipation, observation and understanding of site conditions. Conference Proceedings, GeoEng2000, An international conference on geotechnical and geological engineering, 19-24 November 2000, Melbourne, Australia.
Fookes, P.G., Baynes, F.J., and Hutchinson, J.N., 2001. Total geological history: a model approach to understanding site conditions. Technical note. Ground Engineering, Vol. 34, No. 3. 42-47.
Varnes, D. J., 1974. The logic of geological maps, with reference to their interpretation and use for engineering purposes. United States Geological Survey Professional Paper 837. United States Government Printing Office Washington. pp. 48.