OR/13/006 Geotechnical properties
|Entwisle, D C, Hobbs, P R N, Northmore, K J, Skipper*, J, Raines, M R, Self, S J, Ellison, R A, and Jones, L D. 2013. Engineering geology of British rocks and soils - Lambeth Group). British Geological Survey. (OR/13/006).|
* Geotechnical Consulting Group (GCG)
The geotechnical data reported in this section are derived in the main from routine laboratory testing (BSI, 1990; Head, 1980, 1982 and 1986 and subsequent editions, Brown, 1981). Geotechnical tests on soils and rocks may be broadly sub-divided into ‘index’ and ‘mechanical’ property and engineering chemical tests. The term ‘index’ implies a simple rapid and repeatable test, the equipment and procedure for which are recognised worldwide (e.g. liquid limit); or a test, which measures a fundamental physical property of the material (e.g. particle density). Mechanical property tests measure a particular behaviour of the material under the imposed conditions (e.g. a triaxial strength test). If conditions are changed the result of the test will be different. Equipment and methods for these tests tend to vary worldwide, and note should be taken of the test methodology, particularly where no standard exists. Mechanical property tests tend to require carefully prepared specimens. Index tests tend to be used to characterise a formation and to plan further testing, whereas mechanical property tests may be used for design calculations. For mechanical properties where little or no data are available, such as swell-shrink, permeability and durability, index tests are often used as a guide if correlations have been established elsewhere. In some cases, however, such correlations may not be appropriate. The parameters are in the glossary.
Water content and density
The database contains 4,165 natural water content values (w %) for Lambeth Group samples. The values range between 1% and 74%, however, most values fall between 13% and 28%. The median values for the Reading, Woolwich, and Upnor formations are 19%, 24% and 19% respectively. The distribution of water content values for the Upnor and Reading formations are similar and generally lower than those of the Woolwich Formation. However, the reasons for this may be different. For the Upnor Formation, the lower water contents probably reflect the greater proportion of coarse material present, whilst for the Reading Formation; past subtropical weathering has probably resulted in desiccation and alteration leading to reduced water content values. The Woolwich Formation is the least altered formation and may be slightly or very organic, most notably in Area 2.
Water content data are presented in a series of depth profiles (water content vs. depth) differentiated by formation and by area (Figure 6.1). The wide variation reflects the varied lithologies and degree of alteration (e.g. by subtropical weathering), but some of the higher and lower values are may be due to sampling disturbance and sample storage. This is indicated, for example, by water content values a little below or above the liquid limit for samples described as stiff or very stiff.
The natural water content of the Lambeth Group of the different areas by lithostratigraphical unit are in Figure 6.2 to Figure 6.5.
Very low water content determinations on samples occur for a number of reasons, for example:
- Near surface. Many low water content values are from the near surface and probably reflect sampling during an extended period of dry weather. This is most noticeable in sand samples of all the formations and in the gravel deposits of the Upnor Formation;
- At depth. Shell beds and some very shelly deposits from the Upnor and Woolwich formations and gravel beds of the Upnor Formation often have low water contents, as do some very dense sands. This may be due, at least in part, to sampling disturbance. Pedogenic alteration and formation of calcium carbonate, iron oxide or silica concretions in parts of the Upnor Formation and the Reading Formation (mostly the Lower Mottled Clay).
High water content values, above 40%, near surface may be due to sampling after extended rainfall. However, a majority of those below 3 m are mostly from the Lower Shelly Clay in Areas 1 and 2. Theses samples are mostly described as lignite, which is found at the base of the Lower Shelly Clay most notably in the Shorne inlier, north of Cobham [TQ675 695], Kent.
Some very high water content values (water content > liquid limit) are suspect. These may have been obtained from poorly executed or flooded cable percussion boreholes. The natural water content depth profile is plotted with points differentiated for sampling method in presented in Figure 6.6. The terms used for sampling are those in generally use. ‘Undisturbed’ is taken with cable percussion sampler (U100) and ‘core’ is from rotary coring. The plot shows that nearly all the high values, greater the 40%, are from ‘bulk’, disturbed or ‘undisturbed’ cable percussion samples. However, there are many fewer rotary core samples tested and they do not include such potentially high water content materials such as lignite. Natural water content of clean sand and gravel are not generally useful as they are prone to sample disturbance. The data for the site investigations presented are from before the sampling requirements as stated in Eurocode 7 (BSI, 2007) had come into use, hence the variety of different sampling methods and material types tested.
The database contains 1,565 values for bulk density, r, (Mg/m3). This parameter may also be expressed as the total unit weight, g kN/m2, where g is equivalent to the bulk density multiplied by the acceleration due to gravity (9.807 m/s2). The values of bulk density vary between 1.43 Mg/m3 and 2.375 Mg/m3, although most fall between 1.64 and 2.36 Mg/m3. The median values for the various formations are similar, lying in the range 2.04 to 2.08 Mg/m3, however, overall values may vary considerably locally, both vertically and laterally. The bulk density depth profile for all data differentiated by formation (Figure 6.7) shows a broad spread of values with a generally reduction in variation with depth as there is a reduction in lower bulk density with increasing depth. The London Clay Formation (Figure 6.8) shows a more consistent increase of density with depth. The Lambeth Group has many values between 2.0 Mg/m3 and 2.2 Mg/m3 at all depths particularly within the Reading Formation and the Upnor Formation. The scatter of the Lambeth Group data probably reflects changes due to depth but also lithology, pedogenic effects (including cementation) and other structural features. Whereas, scatter of data for the London Clay Formation is probably associated with depth, the different units within the formation and structural features. The comparison is crude being as the depth is from ground level without any information of units above but it does show the gross differences of the bulk density and depth relationship between the two units. Some low values may be the result of sample disturbance resulting in de-saturation due to stress relief (Hight et al., 2004). Values of bulk density for the Upnor Formation, Lower Shelly Beds, and Laminated Beds may be unreliable due to sample disturbance (Hight et al., 2004), for example where samples are taken by driven tube sampling that can readily disrupt the soil fabric leading to reduced density measurements that are not representative of the in situ state. The CIRIA report (Hight et al., 2004) states that values of less than 1.95 Mg/m3 beneath central London should be treated with caution. The bulk density vs. depth profiles for each of the area locations recognised in this study are presented in Figure 6.7 to 6.12.
The dry density, rd, is the density (or dry unit weight, gd), of the oven-dried soil, i.e. with no ‘free’ water contained in the voids. Measured dry densities are poorly represented in the database, but can be calculated from the bulk density, r Mg/m3, and water content, w%, by the following relationship:
r d = (100 / 100 + w%)´ r
The overall median value of dry density from 1,569 values is 1.71 Mg/m3,with values for the Upnor, Reading and Woolwich formations of 1.71 Mg/m3, 1.736 Mg/m3 and 1.669 Mg/m3, respectively. Median values for Areas 1, 2, 3 are close to the overall dry density median for all data.
Particle density (or specific gravity) is poorly represented in the database (109 values), the majority of these being from the Reading Formation. Of these the overall median value is 2.65 Mg/m3. There are insufficient particle density data values for Woolwich or Upnor formations to draw any meaningful conclusions concerning their statistical distributions.
The results of all Atterberg, or liquid and plastic limit, test data are shown as Casagrande A- line plasticity plots (liquid limit, wL vs. plasticity index, IP), differentiated by formation in Figure 6.13, and by area in Figure 6.14. Plasticity plots of data for each area (Areas 1 to 4) are shown in Figure 6.15 to Figure 6.18 differentiated by formation or unit.
A total of 2,883 liquid limit (wL %) results are contained in the database. Values for liquid limit range from 16 to 123% with an overall median of 53%. Median values for each of the formations range from 40 to 56%. Where distinguished, the Upper Mottled Clay of the Reading Formation, and the Lower Shelly Clay of the Woolwich Formation, have the highest median liquid limits. The data for the undifferentiated Reading Formation, Mottled Clay, is intermediate between the Upper and Lower Mottled Clays in Area 1. The lowest liquid limit values occur in the Upnor and Reading formations of Area 2, reflecting their higher sand content. A few samples in all the formations have extremely high liquid limit values, although this makes up only about 1% of all data.
A total of 2,769 plastic limits (wP %) and plasticity indices (IP = wL- wP) are in the database. The plasticity indices range from 2 to 92%, the overall median being 31%. Median values for the Upnor, Reading, and Woolwich formations are 23%, 36%, and 31%, respectively.
There is a wide range of plasticity values within the Lambeth Group, and within its formations. Overall, the data fall within the ‘low’ to the ‘extremely high’ range, although the great majority fall within the ‘low’ to ‘very high’ range. The clays of the Upper Mottled Clay tend to be of higher plasticity than those of the Lower Mottled Clay. The clay mineralogy of the Lower Mottled Clay and the Upnor Formation is sometimes dominated by the active clay mineral smectite, which suggests that they should have the higher plasticity. However, the Lower Mottled Clay and Upnor Formation contain significant silt and sand that tends to ‘dilute’, or reduce, the plasticity. In the Upnor Formation clays are often present in thin bands and retrieved samples are often mixed with coarser sandy material that results in lower plasticity determinations during laboratory testing. The Laminated Beds in Area 1 are generally more plastic than those in Area 2, probably because they contain more clay. The liquid limit of the Lower Shelly Clay in Area 2 is more variable than in Area 1 as Area 2 contains samples of lignite or highly organic clay, which have higher liquid limit values. Plasticity of the Upper and Lower Mottled Clays and Laminated Beds may increase towards the base. This is may be due to changes in clay mineralogy, and it has been considered that it may also be due to deposits coarsen upwards (Hight et al., 2004), however, the Upper and Lower Mottled Clays generally fine upwards.
Liquidity index, IL, is a ratio which gives an assessment of the ‘position’ of the in situ condition of a soil in its consistency range related to the Atterberg limits.
Values of liquidity index may be used as a guide to desiccation or, where equilibrium water content is established, the degree of over-consolidation of a soil. A value of 0 indicates that the natural water content (w) equals the plastic limit (wP). A value of +1.00 indicates that the natural water content equals the liquid limit (wL). There are 2,591 liquidity indices for the Lambeth Group in the database. Values range from -0.6 to +0.86, with overall median and mean values of zero (that is, water contents are predominantly equal to the plastic limits). The median values of all the formations lie very close to zero. The liquidity index is variable near-surface depths because of seasonal changes in water content. The liquidity index profile for all the data, differentiated by formation and shown in Figure 6.19, tends to show a general trend of decreasing liquidity index with increasing depth. However, there are occasionally relatively high values at depth in samples from all formations, but particularly the Woolwich and Upnor formations. Low values occur near-surface but also at depth in all formations.
Liquidity indices for all data differentiated by area are shown in Figure 6.20, and for each area differentiated by lithostratigraphy (Figures 6.21 to 6.24). Area 1 (Figure 6.21) shows a general trend of reducing liquidity index with depth for the majority of the data but with a significant number of random higher or lower data points. Area 2 (Figure 6.22) shows little change in liquidity index with increasing depth. In this area the Reading Formation consists of the Lower Mottled Clay, which tends to be sandy, as does the Upnor Formation; the Lower Shelly Clay contains lignite, which tends to have a higher liquidity index. Areas 3 and 4 (Figure 6.23 and Figure 6.24) comprise the Upnor and Reading formations and show a weak trend of decreasing liquidity index with increasing depth. However, in both areas there are a number of low values in the Reading Formation and, in Area 3, high values in the Upnor Formation below 10 m. It has been suggested (Hight et al., 2004) that unusually high and low values may be spurious as a result of sample disturbance resulting in mixing of sands and clays (which may be the case for some parts of the Laminated Beds and the Upnor Formation), or redistribution of water during sampling. High liquidity index values are generally measured on cable percussion samples rather than on rotary core samples (Hight et al., 2004). However, some of the unusual values may be due to the variation in particle size where the sample has a large >0.425 mm component, the water content being measured on the whole sample and the plasticity being measured on part of the sample. This effect could be partly removed by correcting the plasticity values but the percentage of <0.425 mm particles has not been recorded in many cases. Also, the Reading Formation may have lower liquidity values due to its desiccation by pedological soil formation processes shortly after deposition.
A total of 1,858 particle-size grading curves are contained in the database. The particle size distribution data are presented as percentile plots including the minimum (coarsest sample) and maximum (finest sample). The different percentiles shown depend on the number of data for each different category, as shown in Table 6.1. Figure 6.25, illustrates the different percentile boundary lines and colours to depict percentile ranges. Whilst the percentile plots do not represent individual grading curves they conveniently summarise large amounts of data and show the distribution of different proportions of the data.
|Number of samples||Percentile values shown|
|25 to 99||10th and 90th|
|100 to 499||2.5th and 97.5th|
|500 or more||0.5th and 99.5th|
The particle size distribution of the Lambeth Group for all data and by formation are shown in Figure 6.26. The Lambeth Group as a whole is very varied ranging from clay to slightly sandy, gravelly cobbles. However, most of the Lambeth Group ranges between slightly sandy clays and silty sands and gravels. The Upnor Formation is the most variable but a majority of samples are coarse grained with over 10% gravel. The Woolwich Formation is more often finer grained than the other formations with fewer samples containing gravel. The Reading Formation also contains a wide spread of lithologies with a higher proportion of clay samples, and less than 10% gravel samples.
Fifty percent of the Upnor Formation (Figure 6.27) varies between a clayey or silty fine to medium sand and a slightly silty slightly gravelly fine to medium sandy clay with a little coarse sand. Nearly 90 percent of all samples are predominantly coarse-grained varying from clayey or silty fine sand to gravel and cobbles. Between 5 and 10% of samples are fine-grained.
Grading by area of the Upnor Formation shows that for the finer sand to clay fractions the median values and interquartile ranges (i.e. the area shown between the 25 and 75 percentiles, representative of the central half the data distribution) are similar in Areas 1, 2 and 3 and indicative of clayey fine to medium sands with a little fine to medium gravel. However, gravel content is variable, with Area 1 having a greater proportion of gravel samples followed by Area 2. Flint gravel is often present in the basal part of the Upnor Formation in all the areas, but in Areas 1 and 2 additional gravel occurs in the ‘pebble beds’ in the upper part of the formation and as gravel-sized particles of hard bands, most commonly calcrete. The sandy gravels ’pebble beds’ might not be recovered or be only partially recovered by either rotary or cable percussion drilling methods, and this may reduce the representation of these materials within the dataset.
The grading of the Reading Formation (Figure 6.28) is almost as variable as the Upnor Formation, however, it is generally finer-grained, the majority being clay, sandy clay or clayey/silty fine to medium sand. Gravel is a major component of less than about 5% of samples.
The particle size distributions of the Reading Formation differentiated by stratigraphical unit for different areas shows some differences between the Upper and Lower Mottled Clay in Area 1 and 2. No such distinction can be made in Areas 3 and 4 and the grading data are presented for undifferentiated Mottled Clay. In Area 1 the Lower Mottled Clay are coarser than the Upper Mottled Clay. The Upper Mottled Clay is predominantly clay with sand layers that are almost always fine-grained, and it is essentially gravel free. Less than half the samples contained more than 10% sand. In contrast the Lower Mottled Clay generally contains a greater proportion of sand, particularly in Area 2, where these beds generally grade into a sand to the east. A majority of the Lower Mottled Clay samples contain over 10% sand; most samples comprising sandy clay or clay/silty fine to medium sand, with some samples containing significant proportions of gravel, which is generally composed of calcium carbonate or iron concretions.
In Area 3, the undifferentiated Reading Formation generally varies between clay and fine to medium sand, and contain much less gravel in comparison to the Lower Mottled Clay. The grading distribution appears to reflect a ‘mix’ or amalgamation of the Lower and Upper Mottled Clay of the Reading Formation in Area 1, but with much less gravel. The gravel particles, where described, comprise calcium carbonate concretions (calcrete), which are less common in Area 3.
In Area 4, the undifferentiated Reading Formation is generally more variable, as shown by the wider interquartile range. These deposits tend to be more gravelly than both the Reading Formation of Area 3 and the Upper Mottled Clay of Area 1. Here the gravel may consist of flint or calcium carbonate concretions.
The Woolwich Formation as a whole appears to be, generally, more well-graded than the Upnor and Reading formations (Figure 6.29) and the Laminated Beds are more well-graded than the Lower Shelly Clay. This may be a true reflection of the particle size or could be due to the mixing of laminations of different lithology before testing to provide suitable sample size, or during drilling and sampling, for instance in disturbed samples. If the particle size distribution of the Laminated Beds is affected by mixing then the resulting particle size will depend on the lithologies and thickness of the laminations.
Gradings for the Laminated Beds in Areas 1 and 2 show some differences, Area 2 is generally coarser, containing a greater proportion of fine sand and more gravel. The gravel is predominantly shelly but may also comprise iron-rich concretions such as siderite.
The Lower Shelly Clay in Area 1 is predominantly a slightly sandy clay or, less frequently, clayey sand often with some shell gravel; whereas, in Area 2 it tends to contain coarser material as it is sandier further east. The increase in gravel in this area is predominantly shell but also lignite particles in the southeast.
Summary of particle size
The Lambeth Group contains a wide variety of deposits varying from clay to gravel and cobbles. The different formations and units do, however, show certain characteristics that may limit the variability. The Upnor and Reading Formation are the most variable.
- Highly variable,
- Generally coarser than the other formations, more than 75% of samples contained more than 50% coarse fraction,
- A majority of samples contain some gravel and over 25% of samples contained more than 20% gravel,
- The gravel is generally flint but may also be shell or calcium carbonate or iron concretions,
- Gravel in the upper part of the formation is an important component in Areas 1 and 2 (‘pebbles beds’), however, they may be under represented in the dataset as they might not be recovered during drilling and sampling.
- Highly variable,
- The Lower Mottled Clay is generally coarser than the Upper Mottled Clay, containing more fine to medium sands and gravels, and becoming sandier to the east in Area 2,
- The Reading Formation in Areas 3 and 4 have similarities with both the Upper and Lower Mottled Clay of Area 1 but contain more sand,
- Gravels in the Reading Formation generally consist of calcareous or iron oxide concretions, and are predominantly found in the Lower Mottled Clay. Flint gravel may occur in Area 4,
- The particle size distribution of the Reading Formation is consistent with fine-grained overbank deposits with mostly fine and sometimes medium sand representing channel infill. In some places, generally in the lower part, gravel has formed as calcium carbonate and iron oxide cemented concretions as a result of subtropical soil formation processes.
- Highly variable,
- Generally more well graded than the other formations,
- Acquired Laminated Beds test samples may often comprise a mix of different coarse and fine laminae,
- Generally the east (Area 2) is coarser than the west (Area 1),
- Most of the gravel is shell but may also comprise iron concretions such as siderite, and, in the Lower Shelly Clay in Area 2, lignite.
Sulphate, pH, and other chemical tests
A small group of relatively simple chemical tests for soils is usually included in geotechnical testing. These are: pH, water soluble sulphate (2:1 water/soil extract), acid soluble sulphate, water soluble chloride (2:1 water/soil extract), and organic content (BS1377 [BSI, 1990a]; Head, ; BRE Special Digest 1, ). In addition, there are tests for the sulphate content of ground water used in modern chemistry laboratories. Other chemical tests that relate to environmental assessments, including total sulphur, water soluble magnesium (2:1 water/soil extract), ammonium ion, water soluble nitrate (2:1 water/soil extract), are not considered here as they are usually a result of site-specific contamination studies and are normally of very local occurrence.
Organic matter is derived from a wide variety of animal and plant remains so there may be a wide range of compounds. The organic compounds present depend on the origin and maturation, which is usually a result of burial and heat. Organic matter, particularly peat or recent organic-rich deposits have reduced bearing capacity, higher and more long term compressibility, lower acidity, and may produce and contain gas.
Excessive acidity or alkalinity of groundwater can have detrimental effects on concrete below ground level. Even moderate acidity can corrode metals. Some soil stabilisation agents may be unsuited to alkaline conditions. The pH also affects the solubility of many metal ions. The measurement is usually carried out on groundwater samples whenever the sulphate content is measured.
Groundwater and pore-water containing sulphate can attack concrete and other materials containing cement. A reaction takes place between the sulphate and aluminium compounds in the cement, causing crystallisation of complex compounds resulting in expansion and build up of internal stresses in the concrete and softening of the concrete. The values obtained from the sulphate tests are used primarily for concrete specification during construction design.
Classification and testing recommendations for sulphate content in soil and groundwater have developed and changed in recent years. A former classification for sulphate in soils given by the Building Research Establishment (BRE Digest 250, [BRE 1981, 1986]) is shown in Table 6.2. This would have informed many of the test regimes carried out prior to 1991 and hence a significant proportion of the sulphate data in the Lambeth Group database.
Concentration of sulphate as SO3
|Total by acid extraction %||2:1 water:soil extract g/l|
|2||0.2 to 0.5||1.0 to 1.9||0.3 to 1.2|
|3||0.5 to 1.0||1.9 to 3.1||1.2 to 2.5|
|4||1.0 to 2.0||3.1 to 5.6||2.5 to 5.0|
In 1991 the classification changed to BRE Digest 363 (BRE 1991; 1995). This BRE classification (Table 6.3) requires assessment of total sulphate, then if above the threshold for Class 1, the aqueous sulphate test is carried out to decide on the appropriate cement type. The classification concentrations for the water extraction and groundwater sulphate are the same as the previous classification but multiplied by 1.2 as the values are now expressed as SO4 and not SO3.
Concentration of sulphate as SO4
|Total by acid extraction %||2:1 water:soil extract g/l|
|2||If >0.24 classify on
the basis of 2:1 extract
|1.2 to 2.3||0.4 to 1.4|
|3||2.3 to 3.7||1.4 to 3.0|
|4||3.7 to 6.7||3.0 to 6.0|
The current classification system in use is BRE Special Digest 1: 2005 (BRE, 2005). This BRE classification uses different classification schemes depending on the category of the site. The classification system used for natural ground locations is presented in Table 6.4. The four site categories are:
- Natural ground locations except those containing pyrite
- Natural ground locations that contain pyrite
- Brownfield locations except those containing pyrite
- Brownfield locations that contain pyrite
In addition to specifying a Design Sulphate (DS) Class (based on water soluble sulphate and total potential sulphate), the scheme also specifies an Aggressive Chemical Environment for Concrete (ACEC) Class (based on groundwater mobility and pH). For natural ground locations water soluble sulphate and pH of soil and water are assessed against criteria in Table 6.4. If pyrite is present in significant amounts and concrete will be exposed to disturbed ground then assessment is also made against total potential sulphate (conservatively estimated as three times the total sulphur). For brownfield locations water soluble sulphate, pH and total potential sulphate of soil and water are assessed. Where concentration of sulphate are high and pH low, additional criteria for magnesium, chloride and nitrate are also assessed. See BRE Special Digest 1: 2005 (BRE, 2005) for full assessment methodology.
|Design Sulphate Class||
Concentration of sulphate as SO4
|Groundwater (SO4 g/l)||Static water||Mobile Water|
|Total potential sulphate (SO4%)||2:1 water:soil extract (SO4 g/l)|
|DS-2||0.24–0.6||0.5–1.5||0.4 to 1.4||>3.5
|DS-3||0.7–1.2||1.6–3.0||1.5 to 3.0||>3.5
|DS-4||1.3–2.4||3.1–6.0||3.1 to 6.0||>3.5
Results from eight ‘chemical’ laboratory test parameters are contained in the database: total (solid) sulphate, aqueous extract (solid) sulphate, sulphate in groundwater, pH, organic content, total chloride (solid), aqueous extract (solid) chloride and chloride in groundwater. Total sulphate or total chloride are the acid-soluble sulphate content, whilst aqueous 2:1 water/soil extract sulphate or chloride is the water-soluble sulphate content. Both are obtained from liquid extracts but give the content of the soil itself rather than of the ground water, and are expressed as a percentage by weight and as grams per litre, respectively. Sulphate and chloride content data may be quoted as below detection level. This greatly complicates statistical assessment of the raw data as these data cannot be included. If they are not included, then the dataset is not wholly representative and may be slightly biased. However, the use of classes for cement type for sulphate content does not have this problem as the ‘below detection level’ data will all be DS Class 1. The number of values used in statistically analysing the chemical data and those below detection level, for each test type, are presented in Table 6.5. Assessment has only been made for natural ground using only pH, total sulphate by acid extraction, water soluble sulphate and groundwater sulphate. There is little data for chloride and no data were available for total potential sulphate, magnesium, or nitrate.
The percentage of samples in the different ‘Design Sulphate Classes’ for each Formation using the accepted classification schemes (Table 6.4) are given in Table 6.6. Aqueous extract sulphates are represented by 140 data values that range from below detection level to 9.21 g/l. About 10% of the Woolwich Group and 6% of the Lambeth Group are classified as DS Class 3 or more. The Reading Formation has the lowest DS Class, nearly always DS Class 1.
Groundwater sulphate tests show similar trends to the total sulphate and aqueous sulphate results. Less than 20% of all the samples are classified as DS Class 2 or more, the great majority being from the Woolwich Formation.
|Test type||Lithostratigraphy||Number of values used in statistical analysis||Below detection level|
|Sulphate, Acid extraction or Total Sulphate||Lambeth Group||488||40|
|Sulphate, 2:1 Aqueous extraction||Lambeth Group||136||4|
|Sulphate, Groundwater||Lambeth Group||92||5|
|Organic content||Lambeth Group||58||-|
|Chloride, Aqueous extraction||Lambeth Group||30||3|
|Chloride, Groundwater||Lambeth Group||124||-|
|Test type and formation||Percentage of samples in each sulphate class|
|Aqueous soluble sulphate|
The total sulphate, aqueous soluble sulphate and ground water sulphate versus depth profiles (Figure 6.30 to Figure 6.32) do not show a discernible trend with depth; unlike the London Clay Formation (Figure 6.33 and Figure 6.34), which shows a general decrease in maximum sulphate contents (total and aqueous) with increasing depth. Typically, sulphate in clay deposits is associated with subaerial weathering, hence the reduction of sulphate with depth in the London Clay Formation. Low sulphate values in the Reading Formation are most likely the result of sub-tropical weathering shortly after its deposition, when oxidation of sulphide to sulphate would have happened and the sulphate removed by dissolution as a part of the weathering process. In contrast, the Woolwich Formation and much of the Upnor Formation were not subject to similar weathering to any great extent and usually retain iron sulphide and calcium carbonate until the iron sulphide is oxidised by contemporary weathering. This may occur at depth due to air ingress when the water table is depressed. This is much less likely in the thick clay sequence of the London Clay Formation.
The analysis of pH is based on 664 data values ranging between 3.7 and 9.9. The median values of all the formations and units vary between 7.3 and 8.1, that is, slightly alkaline. About 70% of all values indicate that the samples have pH of 7 to 8. Most of the more acidic samples (pH <6) are from the Upnor and Woolwich formations, particularly from Area 2. This is probably related to organic content and/or the oxidation of pyrite. All the samples with pH of less than 6 are from the top 10 m, those with values in excess of pH 9 may be from any depth (Figure 6.35).
The water-soluble chloride is generally less than 0.2 g/kg with two values from the Upnor Formation of 0.4 and 0.6 g/kg, which were from over 15 m below ground level in the Canary Wharf area of London. Groundwater chloride content is generally less than 0.5 g/l, with 6 values greater than 1.0 g/l. Four of the the higher values, from the Woolwich and Upnor formations, are from boreholes drilled in the Thames River or in the docks on the Isle of Dogs. The other two higher values, both greater than 2 g/l were from the Reading Formation from a site in Brey, north west of Windsor, Berkshire.
Acquired site investigation data indicates that organic content is not regularly carried out on material from the Lambeth Group, even where the organic content is likely to be high. The Reading Formation will tend to have very little or no organic content, although there may be some at or above the mid-Lambeth Group Hiatus in the west. Organic material is described in the Laminated Beds and in some of the sand channels in the Upper Mottled Clay in Area 1 and occasionally in the Upnor Formation. The Woolwich Formation may contain significant organic content, most notably in the Lower Shelly Clay, the highest values being for lignite in Area 2.
Most of the higher sulphate values occur within the Woolwich Formation, therefore, appropriate tests to assess the need for sulphate resistant cement are particularly required in this formation. The Woolwich Formation has the greatest potential to form sulphates due to its high pyrite content, often associated with organic material, and calcium carbonate (shell) content, which has been preserved by anoxic conditions of deposition. Sulphate contents appear to show little change with depth, unlike marine clays such as the London Clay Formation.
The few low values of pH (below 6) are all found in the upper 10 m, whereas, high values (greater than 9) are found at any depth.
Organic content is highest in the lagoonal and estuarine Woolwich Formation, most notably at the base of the Lower Shelly Clay in the southeast of the London Basin and the eastern part of the Hampshire Basin, where significant thickness of lignite are present. It is also present in the Laminated Beds and sometimes in the sand channels in the Upper Motteld Clay and occasionally in the Upnor Formation.
The strength of a soil or rock is a measure of its capability to withstand a stress (or stresses) in a particular direction or configuration. Strength is not a fundamental property of a soil or rock, but is dependent on the condition of the soil/rock and the type of stresses applied to it. The measured strength of soils is particularly sensitive to the drainage conditions and duration of the test, in addition to specimen characteristics such as density and fissuring. If drainage is allowed the test is capable of measuring effective strength parameters, which are usually required for the assessment of ‘long-term’ strength. These are usually determined from tests that include consolidated drained triaxial (CD) tests, consolidated undrained (CU) triaxial tests with pore pressure measurements, and drained shear box tests. If the conditions are undrained the test is assumed to measure total strength parameters, unless pore-water pressures are measured, in which case the effective stress parameters may be calculated. Total strength parameters are generally determined using unconsolidated undrained (UU) triaxial tests, shear vane and penetrometer tests. All effective and total strength tests reported here have been acquired from intact ‘undisturbed’ laboratory specimens.
Total shear strength (τ) is usually defined by the Mohr-Coulomb failure criteria, the equation of which is as follows:
t = c + s tan f
Where: c = cohesion, s = normal (perpendicular to shearing) stress, and f = angle of internal friction.
For a fully saturated, intact specimen, prevented from draining at all stages of the test, the value of the internal friction angle, f, is zero. The undrained shear strength, su, thus equals the undrained cohesion, cu. However, if triaxial test specimens are consolidated at each stress level by allowing drainage, as in the consolidated-undrained (CU) or consolidated-drained (CD) tests, effective shear strength may be measured if pore-water pressure is measured and subtracted from the total stresses. This is reported in terms of the ‘effective’ cohesive and frictional strength parameters c’ and f’. The effective shear strength, s’, is then calculated from the Coulomb equation as follows:
s' = c'+(s - u) tan f'
The residual strength is the minimum strength of rocks and fine-grained soils after initial shear failure has occurred and may be determined on intact or remoulded samples in a shear box or a ring shear apparatus. The residual strength is usually determined to assess the strength along a pre-existing shear plane (e.g. on samples from a landslide slip surface), or in certain highly fissured clays.
It is difficult to give typical or average values of strength for the Lambeth Group or individual formations and members within it, because of the variability of lithology, fabric, structure, and cementation and the post-depositional processes of weathering and consolidation it has undergone. This results in variable depth profiles for intact strength on a scale of metres or centimetres, whether these are determined in situ or in the laboratory.
Undrained shear strength
Undrained (total) triaxial strength data are reported in site investigations either with the assumption that the friction angle,f, is zero, or that it has a positive value, despite this being contrary to the principles of the test (Head, 1992; Head, 1998). Undrained strength data containing a positive friction angle have been omitted from the database.
The 1,338 undrained cohesion (cu) values analysed show variable undrained strengths within the Lambeth Group. Median strength values range between 112 kPa and 164 kPa, with overall values ranging between c.10 kPa to over 800 kPa ,with the Reading Formation tending to have the highest and the Woolwich Formation the lowest values. The data show undrained strengths to be particularly variable in central London and Hight et al. (2004) comment that the project-wide variability in undrained strength is similar to that found at a single location. This is the case for all formations and units.
The profile of Lambeth Group undrained strength values with depth (Figure 6.36) shows a great deal of scatter with an indistinct, trend of increasing strength with depth. Samples of extremely high strength and stronger (>300 kPa) occur near surface and increase in number with depth. There are also low strength samples at depth. In comparison to the Lambeth Group data, the undrained strength profile based on 2,100 data values for the London Clay Formation, from all areas, shows a clear overall trend of increasing strength with depth, but with generally less scatter of the data at all depths (Figure 6.37). The contrast between the Lambeth Group and the London Clay Formation results reflect the differences in their depositional environments and the post-depositional processes in particular pedogenic processes (cementing and fissuring) that affected some of the Lambeth Group deposits (Hight et al., 2004).
Depth profiles of undrained strength values for each area are given in Figure 6.38 to Figure 6.41. Again, a high scatter of results is evident but an overall trend of increasing strength with depth can be seen.
- In Area 1 (Figure 6.38) the undrained strength vs. depth plot indicates a general increase in undrained strength with depth for all units in the top 10 m, but with increasing variability for all units at depths greater than 10 m. The Upper Mottled Clay has the greatest increase in strength with depth but also the greatest variability. Below 10 m, the Upnor Formation shows no clear trend of increasing strength with depth;
- Area 2 (Figure 6.39) shows a similar general trend of increasing undrained strength with depth for all formations/units to about 30 m, with the exception of the of the Lower Shelly Clay where no such trend is discernible;
- Area 3 (Figure 6.40) a trend of increasing strength with depth for the Reading and Upnor formations is seen to about 10 m below ground level. Below this depth the variability of the strength data for the Reading Formation increases markedly with no clearly discernible trend to 40 m;
- Area 4 (Figure 6.41), based on a more limited dataset, shows a similar general trend of increasing undrained strength with depth for both the Reading Formation and undifferentiated Lambeth Group samples to 35 m.
Median effective strength data, (c´ and f´ values) for the Lambeth Group formations are shown in Table 6.7.
|Group/Formation||Number of samples||Dominant lithology||Test type||c´ (kPa)||f´ (°)|
As would be expected, the data show that the effective cohesion, c´, values are greater for clay samples, with sand samples having higher median angles of internal friction. However, a number of data values for Upnor Formation sand are similar to Upnor Formation clay, which may reflect the presence of clay laminae in the tested ‘sand’ samples or differences in lithology between the field description and test sample.
A plot of effective shear strength (s’) vs. depth (Figure 6.42), calculated using effective triaxial test data combined with estimated overburden stresses obtained from median densities, shows an overall well-defined trend of strength increase with depth for the Upnor and Reading formations, but with the latter showing some scatter of high data values within 10 m of the ground surface.
The dominantly sandy Upnor Formation data is distinguished by test type (shear box and triaxial tests) and by major lithology (clay or sand). The Reading and Woolwich Formation envelopes are distinguished by test type and also by different plasticity classes as suggested by Hight et al. (2004).
The effective strength envelopes for the Upnor Formation (Figure 6.43) show that, in general, there is little difference between most of the clay and sand envelopes, which probably reflects the mixed lithology of this formation (most of the samples tested being described as sandy clays or clayey sands). The triaxial and shear box test results are similar, as factors that may provide a difference in results between the two test types (e.g. fissuring) are rare in the Upnor Formation. However, two sand samples have high cohesion values due to cementing.
The effective stress envelopes of the Reading Formation (Figure 6.44) from both triaxial and shear box tests show a reduction in angles of internal friction in samples with higher plasticity index (>35%). Effective cohesion values are more variable and there are major differences between values from shear box tests and triaxial tests. In general, effective cohesion values determined from shear box tests vary little between the different plasticity index classes, whereas for the triaxial tests samples in the higher plasticity classes generally have lower effective cohesion, with over 50% of values recording 0 kPa.
Effective stress envelopes for the Woolwich Formation (Figure 6.45. Effective strength envelopes for the Woolwich Formation) indicate that samples in the lower plasticity classes tend to have higher effective cohesion values.
Residual shear strength is the minimum strength of a soil reached after continuous shearing along a pre-determined shear plane, and can be tested on intact or remoulded samples in the laboratory. The results are expressed in terms of the residual angle of internal friction, fr´, and residual undrained cohesion, cr´, from a plot of effective normal stress vs. shear stress, which is generally considered to be a straight line. The values of cr' should be very low or zero but this can only ascertained if tests are carried out at a series of low normal stresses.
Only a few residual shear strength values were available for the Lambeth Group mostly from shear box tests. Residual angles of internal friction, fr´, ranged from 8° to 36° with an overall median value of 27°, and a median value for clays of the Reading Formation of 18°. Residual shear strength (angle of internal friction) has been plotted against plasticity index in Figure 6.46. This shows an inverse correlation between residual shear strength and plasticity index as demonstrated for Lambeth Group data by Lehane et al. (1995), Lupini et al. (1981), and Voight (1973). It also suggests a change in residual strength behaviour from turbulent to sliding (Lupini et al., 1981) at a plasticity index, Ip, between 20 and 25%. Lehane suggested that for clays with plasticity indices of greater than 30% the residual angle of internal friction, fr´, was approximately 11 ± 3°. The plot of fr´ and plasticity index indicates that a majority of values with plasticity index values of greater 25% where the sliding mode is expected to occur had fr´ values are above the bounds based on Vaughan et al. (1978). This may be due to insufficient shearing strain during the shear box tests.
Factors affecting the strength of the lambeth group clays
Hight et al. (2004) considered the factors affecting strength, such as the effects of plasticity, fissuring and sample disturbance, in some detail. Much of the findings from the work presented here corroborate those described by these authors.
The undrained shear strength of the Lambeth Group is very variable and despite overall trends, increase in strength is not necessarily related to depth below ground level. Hight et al. (2004) considered that Lambeth Group samples with undrained strengths greater than 500 kPa were probably cemented. However, the high strength may also be due to desiccation during the sub tropical climatic conditions of the Palaeocene. Low undrained strength values occur in samples described as stiff or very stiff. This may be due to failure along fissures, which are commonly described in the clays of the Lambeth Group, most commonly in the Reading Formation, and/or sample disturbance.
For the Reading Formation clays samples of higher plasticity tend to be weaker than low and intermediate plasticity materials (<50%) of similar liquidity index (Figure 6.47). This is may be due to a greater concentration of fissures in the high plasticity soils or more cementing in low plasticity soils; cementing agents such as iron oxides and hydroxides and calcium carbonate are inactive minerals and are likely to reduce plasticity. In some cases the peak and residual angles of internal friction are similar (Figure 6.46), indicating that the samples are, to some extent, shearing along preformed fissures.
Under a subtropical climate with pronounced seasonal rainfall, as prevailed during the deposition of the Lambeth Group, higher plasticity clays are most prone to shrink and swell. They are, therefore, more likely to fissure during the dry season. In contrast, low plasticity clays are less likely to shrink and swell to the same extent, will tend to form fewer fissures to a more limited depth, and develop fissure surfaces likely to be less smooth than those formed in high plasticity clays. Also, pedogenic cements, in particular calcium carbonate, will tend to reduce plasticity and increase strength.
Consolidation is the process whereby pore water is expelled from a soil as the result of applied, static, external stresses, resulting in structural densification of the soil. For most purposes, the external stress is considered to be unidirectional, and usually vertical. Swelling strain data may also be obtained from the oedometer test. The oedometer is a simple laboratory apparatus, which applies a vertical load to a small disc-shaped soil specimen, laterally confined in a ring. The consolidation test is normally carried out on undisturbed specimens by doubling the load at 24-hour intervals, and measuring the resulting consolidation deformation (BS1377: BSI, 1990; Head, 1998). This test is only suitable for fine-grained samples.
The rate at which the consolidation process takes place is characterised by the coefficient of consolidation, cv, and the amount of consolidation by the coefficient of volume compressibility, mv. Consolidation data derived from the oedometer test on undisturbed specimens are used in the calculation of likely foundation settlement, and may also provide information on the stress history, geological history, state of disturbance, permeability, and elastic moduli of clay soils.
The consolidation data considered here is from laboratory oedometer tests. Only those tests where the load has doubled at each stage are included in the National Geotechnical Properties Database.
This section includes information on the following:
- Initial voids ratio,
- Change of volume (coefficient of volume compressibility, mv) with consolidation stress,
- Change of the rate of consolidation (coefficient of consolidation, cv) with consolidation stress.
There are results from 370 oedometer tests in the database. A majority, 244, are on Reading Formation clays, 91 on Woolwich Formation clays and silts, 33 on the Upnor Formation and 2 on undifferentiated Lambeth Group.
The initial voids ratio, e0, of the Lambeth Group samples are plotted against depth in Figure 6.48. Maximum voids ratio reduced slightly with depth, whereas there is little increase in the minimum values.
Coefficients of volume compressibility and consolidation data for estimated in situ stress +100 kPa, similar to the data given in Hight et al. (2004), are summarised as minimum maximum and percentile values in Table 6.8, and presented as graphs with respect to consolidation stress in Figure 6.49 and Figure 6.50.
Estimated in situ stress +100 kPa
Coefficient of volume compressibility,
Coefficient of consolidation,
The data show that the clays and silts of the Lambeth Group have very low to medium compressibility, with the Reading Formation tending to have lower values than the Upnor and Woolwich formations. This may be due to the effects of subtropical weathering, typical of the Reading Formation decreasing water content, and voids ratio and increasing stiffness. The coefficient of consolidation values are variable for all the units but the Lower Mottled Clay generally have slightly lower values and the Upnor Formation and Laminated Beds slightly higher values. The variation probably reflects differences in the initial voids ratio, particle size and plasticity and perhaps the changes in particle size within the sample, i.e. the samples of Laminated Beds and Upnor Formation may contain laminated clay and silt reducing the time to consolidate as water drains more rapidly within the silt laminae.
Summaries of the coefficient of volume change (mv) and coefficient of consolidation (cv) with stress (Figure 6.49 and Figure 6.50, respectively) show that, in general, coefficient of volume change of all materials tends to reduce with increasing stress. The coefficients of consolidation of the Reading Formation and the Lower Shelly Clay samples from Area 1 reduce with increasing stress, but there is little reduction with stress for the Upnor Formation, Lower Shelly Clay from Area 2 and Laminated Beds. This may be due to the greater clay content of those that do show this reduction with stress.
Deformability (the terms compressibility and stiffness may also be used) is a measure of the strain undergone by a soil or rock subjected to a particular level and direction of stress. This strain may be unidirectional or volumetric. Deformability may be measured in both laboratory (intact) and field (rock or soil mass). Usually, test data are interpreted from stress-strain plots, with several parametric variants of deformability available. The elastic properties of a material are defined by the fundamental properties: bulk modulus, K, and shear modulus, G. Bulk modulus represents the change in all-round stress per unit change in volume, whereas shear modulus represents the change in shear stress per unit change in shear strain. The simplest form of deformability measurement is that of Young’s modulus, E, which is derived from a uniaxial compression test and is defined as follows:
where: σ1 = major principal stress
- ε1 = strain in direction of major principal stress
The relationship between strain in the direction of stress and strain at right angles to it is defined by the Poisson’s ratio, n, as follows:
where: σ1 = major principal stress
- ε1 = strain in direction of major principal stress
- ε2,3 = strain at right angles to major principal stress E = Young’s modulus
Shear modulus, G, is defined as:
- E = Young’s modulus
- n = Poisson’s ratio
- G = shear modulus
- E’ = drained Young’s modulus
- n’ = drained Poisson’s ratio
Shear modulus may be measured in a variety of ways from the stress vs. strain plots. The most commonly quoted are the initial shear modulus, Gi, and the unload/reload modulus, Gur.
No deformability data are contained in the database. However, Hight et al. (2004) gave an account of case studies where small-strain stiffness determinations were made in the laboratory using triaxial tests. Undrained Young's moduli at 0.1% strain, normalised for effective overburden, were quoted as 820 and 1265 for Upper Mottled Clay samples obtained from tunnels at the Angel, Islington (London). These values were considerably higher than those for the London Clay. The low plasticity clays of the Upper Mottled Clays were found to be stiffer than the higher plasticity clays of the Upper Mottled Clay (Hight et al., 2004). Also stiffness reduces almost by an order of magnitude with increasing strain form 0.001% to 1% strain from consolidated anisotropic undrained triaxial compression and extension tests.
Permeability, in the geotechnical context, is a measure of the ability of soil or rock to allow the passage of water subject to a pressure gradient. The permeability measured on intact specimens in the laboratory is usually distinct from that measured in the field, as a result of the huge scale difference, and the influence, in the field tests, of discontinuities and lithological variations. The database contains 241 permeability determinations covering the major lithostratigraphic units. The medians for the Reading Formation members range from 3.5 x 10-8 to 2.5 x 10-7 m/s. The Upnor Formation gave medians of 8.3 x 10-7 m/s (Glauconitic Sand) and 5 x 10-7 m/s (Pebble Beds). The Woolwich Formation medians were 3.5 x 10-7 m/s (Laminated Beds) and 6.0 x 10-7 m/s (Lower Shelly Clay). The overall minimum and maximum for the Lambeth Group were 1.8 x 10-10 m/s and 2.0 x 10-4 m/s. The permeability medians, maxima, and minima for each Formation were unexpectedly similar.
Hight et al. (2004) gave ranges of in situ permeability for the Lambeth Group from the Crossrail and Channel Tunnel Rail Link (CTRL) projects as shown in Table 6.9.
|Formation||Permeability, k (m/s)||Permeability, kH (m/s)|
|RBUMC||5 x 10-7 to 5 x 10-9|
|RBLMC||1 x 10-8|
|WLLB||2 x 10-7 to 3 x 10-8|
|UPRGS||1 x 10-8 to 4 x 10-8|
|UPR||1 x 10-4 to 1 x 10-8|
where: kH is horizontal permeability
The permeability figures quoted for the Reading Formation Mottled Clays are probably influenced by the presence of sandy layers and possibly, fissure flow clays. The permeability of the mottled clays material is probably lower than the figures quoted. Sand layers within the Reading Formation Mottled Clay members impart a distinctly anisotropic element to the permeability. The contrast in permeability between these two component lithologies is considerable. The general ranges shown in Table 6.10 may be used for comparison:
|Very fine or silty sands||10-5–10-8|
|Fissured and weathered clays||10-4–10-8|
Compaction, california bearing ratio and moisture condition value
Compaction is the process whereby soil is densified, usually by reworking in layers, in order to produce an engineering fill of known properties. This is achieved by applying dynamic forces, using special plant, such as rollers, vibratory rollers, rammers, or by special ground improvement processes. The densification is achieved by the soil particles packing closer together to:
- increase shear strength and, therefore, bearing capacity,
- increase stiffness and, therefore, reduce future settlement,
- decrease voids ratio and permeability, thus reducing the potential for frost heave.
The water content of the placed fill and the amount of energy input are critical to the density that can be produced. The process is not the same as consolidation.
The Lambeth Group is variable source of fill material for earthworks due to the scale of its lateral and vertical lithological variation and water content changes. The acceptance criteria of the different materials encountered and the potential for blending will be needed as part of the planning and designing of earthworks. Also, some of the material, such as lignite, is unsuitable. For these reasons the Lambeth Group is often considered to be a difficult material to use as an engineered fill. However, the increasing financial and environmental costs of importing and removing material to and from site have increased the use of such materials. Engineering fill requires identification of suitable material from the site investigation, appropriate specification and control of the material and its emplacement. If this is done then it is possible to keep importation and waste to a minimum. In some areas, for instance near Orsett of south Essex and Upnor in north Kent, the sand and gravels of the Upnor Formation dominate the Lambeth Group providing suitable material for some earthworks.
Most of the compaction data are from road schemes in Areas 3 and 4. In Areas 1 and 2 the data are from investigations for roads and railway construction projects (e.g. the Channel Tunnel Rail Link, CTRL). Of the 139 optimum water content and maximum dry density data test values available 20 were for California bearing ratio (CBR). Most of the data are for light or heavy compactive effort, and a few vibro-compaction tests. Compaction data are generally presented as plots of optimum water content vs. maximum dry density as shown for different lithologies and compactive efforts (Figure 6.51). The data show fairly typical behaviour;
- The clays have higher optimum water contents than the sands.
- Light compactive effort tests have lower dry density and higher optimum water contents.
Unusually, the vibro-compaction values are similar to light or heavy compactive effort results.
California bearing ratio
California bearing ratio is plotted against optimum water content and maximum dry density (Figure 6.52 and Figure 6.53 respectively). The two plots show a tendency of increase in CBR with decreasing optimum water content and increasing optimum dry density for each lithology.
Moisture condition value
The moisture condition value (MCV) is a laboratory or field test, and is a means of selection, classification, and specification of fill material (BSI, 1990; Highways Agency, 1991; Caprez and Honold, 1995). The test aims to determine the minimum compactive effort required to produce near-full compaction of a 1.5 kg sample of soils passing a 20 mm sieve. The test differs from the traditional Proctor compaction test in that the compaction energy is applied across the entire sample surface, and compaction energy can be assessed as an independent variable. A total of 38 MCV data for Reading Formation Mottled Clays are contained in the database. The MCV values range from 0 to 18%, with a median of 9.3%. MCV values less than 7% tend to indicate very poor trafficability.
Swelling and shrinkage
Swelling and shrinkage are two mechanical properties of a soil, which though driven by related physio-chemical mechanisms, are usually treated separately in the laboratory. Swelling is mainly a function of the clay minerals present in the soil or rock. The engineering phenomenon of heave may be caused by factors other than swelling of clay; for example, by stress relief. The geological processes affecting swelling and shrinkage were described by Gostelow (1996). Assessment of swelling and shrinkage usually does not involve direct measurement, but rather indirect estimation of volume change potential from index tests on reworked samples. Swell/shrink tests are not well catered for in British Standards.
The wide variety of test methods applicable to swelling/shrinkage is described by Hobbs and Jones (1995). Laboratory tests may be carried out on undisturbed or disturbed samples. Undisturbed samples are as near to their in situ condition as possible, whereas disturbed samples may be reworked, reconstituted, or compacted depending on the engineering application. Swelling tests usually measure either the strain due to swelling, resulting from access of a sample to water, or the pressure produced when the sample is restrained from swelling (zero strain test). Swelling strain samples may be disc-shaped oedometer type samples for one-dimensional (1D) testing of soils and slaking rocks, or cubes for three-dimensional (3D) testing of non-slaking rocks. The 1D samples are laterally restrained. Swelling pressure samples are usually oedometer discs and may be mounted in a normal oedometer or a special swelling pressure apparatus. Two shrinkage tests are specified by British Standards (BSI, 1990). These are the shrinkage limit test, carried out on undisturbed or disturbed samples, and the linear shrinkage test, carried out on reworked soil paste (prepared as for Atterberg limits). It should be noted that the shrinkage limit is a specific water content below which little or no volumetric shrinkage occurs, whereas the linear shrinkage is a percentage reduction in length (strain) on oven drying.
The Lambeth Group is generally considered to be of ‘low’ to ‘medium’ swell/shrink potential depending on lithology and mineralogy. Whilst no directly determined swelling or shrinkage data are held in the database, a small number of tests were carried out at the BGS on undisturbed samples of Reading Formation, Mottled Clays from various locations. The results found a good positive correlation between 1-D swelling strain, e1-D, and swelling pressure, Psw, as follows:
The tests have also shown that the laterally confined vertical swelling strains (1D swelling strain test) are typically between 1 and 2 times unconfined maximum vertical swelling strains (3D cube test). Volumetric swell strain ranged from 0.4 to 14.0%. Vertical swell (i.e. perpendicular to bedding) was found to always exceed horizontal swell in the 3D cube test, by up to 4 times. This swell anisotropy was greatest for the Knoll Manor 2 and Whitecliff Bay 1 samples. Maximum swell was typically achieved between 10 and 100 hours. Free swell test data had a range of 18 to 80%. No correlations were obtained between free swell, volumetric swell strain, and the index parameters (liquid limit, plasticity index, liquidity index) and activity.
The BRE Digest 240 (BRE, 1993) gives a scale of susceptibility to volume change (i.e. swelling or shrinkage) for over-consolidated clays in terms of a modified plasticity index, Ip' (Table 6.11).
|Ip’||Volume change potential|
where: Ip’ is a modified plasticity index:
The purpose of the modified plasticity index is to take account of the proportion of fines in relation to the total sample and to reduce the measured plasticity index in proportion. Many Atterberg limit data in the database do not include <0.425 mm results. This may be because the sample did not require sieving, or that a small number of coarse particles were removed by hand, without sieving. The modified plasticity index and volume change susceptibility data are shown in Table 6.12.
|Formation||Number of samples||Median Ip´ (%)||Median volume change potential||Samples with high or higher volume change susceptibility (%)|
The BRE Digests 240 and 241 (BRE, 1993 and 1990, respectively) classifications do not indicate the actual volumetric shrinkage to be expected for each of the volume change potential categories. Net volume changes depend on the initial saturation condition of the test sample. In the case of the shrinkage limit test this is usually natural water content, whereas in the case of the linear shrinkage test it is close to the liquid limit. All the Mottled Clay samples tested at BGS for swell/shrink gave a 'medium' volumetric susceptibility according to the BRE classification.
Standard Penetration Test (SPT) results
The Standard Penetrometer Test (SPT) is a long-established method of in situ geotechnical testing, which was initially designed to measure the density of coarse-grained deposits but is now commonly used on most materials encountered during cable percussion drilling. This dynamic method employs a falling weight to drive a split-sampler and cutting shoe (or solid 60° cone in the case of coarse soils or soft rock) 300 mm into the ground from a position 150 mm below the base of a borehole; the initial 150 mm being the 'seating’ drive. The use of the test is described in British Standard 5930 (BSI, 1999; BSI, 2010) and the methodology in British Standard 1377: Part 9: Clause 3.3 (BSI, 1990), which has been superseded by BS EN ISO 22476-3 (BSI, 2005). There has been much discussion concerning the test method, test apparatus, and test interpretation (Stroud and Butler, 1975; Stroud, 1989).
It was recommended (Clayton, 1995; BSI, 1990) that test results be reported in the form of six 75 mm penetration increments; the first two representing the 'seating’ drive and the final four the 'test’ drive, the sum of the latter providing the SPT 'N’ value. This is often not the case in site investigation reports, though it does form part of the Association of Geotechnical Specialists (AGS) digital data transfer format.
The Standard Penetration Test (SPT) may be regarded as crude, but it is inexpensive and effective. In most cases, site investigation reports included a record of the incremental blows and penetrations. These have been entered into the geotechnical database for analysis. The summaries presented for the SPT are derived from over 3,000 tests.
For a small proportion of tests the incremental data were not available. In these cases, tests giving a full N value were accepted, together with those in which a partial main test drive could be distinguished from a seating drive.
A total of 3,010 completed SPT N-values were assessed statistically and another 1,511 incomplete tests were considered. Incomplete tests were stopped before the full depth of 0.45 m had been completed (0.15 m seating blows and 0.30 m test). Historically, tests were often aborted if the N value was >50 blows, indicating very dense sand or gravel or material that could not be driven into. However, higher values have been reported where the information has geotechnical importance but cut off values are, generally, still used; such as, for example, 140 blows for the investigations at Farringdon Station, 150 for the Jubilee Line Extension and 200 for the Crossrail, Channel Tunnel Rail link (CTRL) and the Newbury Bypass (Hight et al., 2004). Table 6.13 shows the SPT completion data and percentage of tests over 50 blows for the Lambeth Group, its formations as well as the London Clay and Thanet formations for comparison.
The Woolwich and Reading formations have similar percentages of incomplete tests and tests of over 50 blows (54% and 53%, respectively), whereas the 70% of tests on the Upnor Formation were greater than 50 blows. Although this is significantly more than the rest than of the Lambeth Group, it is markedly less than the Thanet Formation, which is usually a very dense sand (Table 6.13). The London Clay Formation has only 13% of test values greater than 50 blows and most tests are completed.
|Lithological unit||Number tests (including incomplete tests)||Completed tests||Completed tests %||% of tests >50 blows (including incomplete tests)|
|London Clay Formation||2941||2866||97.4||13|
The medians and ranges of SPT N-values for the Lambeth Group formations are similar, with the Upnor Formation tending to have slightly higher values (median 47 blows) and the Woolwich Formation slightly lower values (median 40 blows). The range of values for each unit, each area and for each unit in an area, is very large, usually greater than 100 blows. The highest values tend to occur in the Upnor and Reading formations of Area 1 and the lowest in Area 4, which may be due to the greater depth of test in Area 1. The depth vs. SPT N-value plots for the Lambeth Group by formation and for each area by unit are given in Figure 6.54 to Figure 6.58. In all cases there is a trend of increased N-value with depth, which is most marked in the upper 20 to 25 m. Higher values (N-values >100) may occur at any depth. Increases in the minimum and maximum values with depth are more obvious in Areas 3 and 4. These variations were considered by Hight et al. (2004) to be due to:
- Cementing due to pedogenic processes,
- Shelly limestone,
- Plasticity and fissure texture.
Figure 6.59 shows the SPT N value vs. depth profile for data from the 1:10k map sheet TQ38SE in east London, which area includes Hackney and Poplar, for Thanet Formation, Lambeth Group and London Clay Formation. The graph shows an increase in N-values with depth for the London Clay Formation, typically high values for the Thanet Formation and a large scatter of the Lambeth Group tests.
Brief summary of geotechnical properties
Despite a generally wide variation in geotechnical properties, analysis of parameters held in the geotechnical database has revealed some recognisable trends in the properties and engineering behaviour of the principal formations within the Lambeth Group. Data have been analysed by stratigraphical formation and unit (where available) and according to arbitrarily defined geographical areas. Some areas give contrasting parameter values compared with others. However, the data are by no means equally distributed across the areas, and in many cases variation within a single area is similar to that across multiple areas.
The densities of the Reading Formation Mottled Clays are consistently high (bulk density median = 2.1 Mg/m3).
The particle size distribution of the Lambeth Group varies from coarse, cobbly gravel or sandy gravel in the Upnor Formation to nearly pure clays within the Reading and Woolwich formations. All the formations were variable. The Upnor Formation generally varied between a sandy gravel to a slightly sandy clay but is primarily coarse-grained. The gravel within the Upnor Formation is primarily flint but, in some areas is calcrete and occasionally silcrete of ferricrete. The Reading Formation is a sandy gravel to a clay but a majority is a clay, sandy clay or clayey/silty fine to medium sand. The gravel comprising mostly of calcrete. Of the units within the Reading Formation the Upper Mottled Clay tends to be finer than the Lower Mottled Clay partly due to the presence of calcrete gravel in the latter unit. The Woolwich Formation is generally more well graded than the other formations and the Laminated Beds is more well graded than the Lower Shelly Clay. Both the Lower Shelly Clay and Laminated Beds tend to be coarser in Area 2; either because of more sand of gravel composed of shell in the case of the Lower Shelly Clay or iron-rich concretions in the Laminated Beds. The Laminated Beds and Lower Shelly Clay, Woolwich Formation, gave the highest total sulphates. This is because of oxidation of pyrite due to near surface weathering and it reaction of the sulphate ions with calcium ions primarily from shell.. They range from zero to 1.5 g/l with a median of 0.08 g/l overall. The Reading Formation medians range from 0.05 to 0.30 g/l, whereas the Woolwich Formation median is 0.57 g/l. Organic contents ranged from 0 to 60% with an overall median of 0.34%. The highest values are from the Woolwich Formation, most notably the lignite of the Lower Shelly Clay Member which has a median of 16%.
The undrained strength, cu, for the Lambeth Group is relatively high, ranging from 82 to 223 kPa. The Reading Formation Upper Mottled Clays show the highest median (223 kPa), and the Upnor Pebble Beds the lowest (82 kPa). The remaining units represented lie within a relatively narrow band between 122 and 151 kPa with the exception of the Reading Formation Lower Mottled Clay at 178 kPa. Undifferentiated Reading Beds Mottled Clays gave a median cu of 151 kPa. The profile of cu with depth is highly scattered although a general trend of increasing strength with depth can be discerned. A negative correlation is found between residual shear strength and plasticity index. The median residual friction angle for Reading Formation Mottled Clays was 18°.
The results of oedometer consolidation tests place the Lambeth Group in the 'low' to ‘medium' rate of consolidation cv class, typical of medium and high plasticity soils. The overall median values for coefficient of volume compressibility, mv, reduce consistently with increasing stress from 0.09 to 0.004 m2/MN The mv results place the Lambeth Group in the 'low' to 'very low' category.
The permeability medians, maxima, and minima for each formation were unexpectedly similar. The medians for the Reading Formation members range from 3.5 x 10-8 to 2.5 x 10-7 m/s. The Upnor Formation gave medians of 8.3 x 10-7 m/s (Glauconitic Sand) and 5 x 10-7 m/s (Pebble Beds). The Woolwich Formation medians were 3.5 x 10-7 m/s (Laminated Beds) and 6 x 10-7 m/s (Lower Shelly Clay). Overall, values ranged from 1.8 x 10-10 to 2.0 x 10-4 m/s.
Compaction data gave maximum dry density and optimum water content formation medians in the range 1.80 to 1.99 Mg/m3 and 10 to 13%, respectively. A total of 38 MCV data for are contained in the database. The range of MCV results for the Reading Formation Mottled Clays was 0 to 18, with a median of 9.3%.
Swell/shrink test data obtained from Reading Formation Mottled Clay samples at BGS showed a good relationship between 1D swelling strain and 1D swelling pressure. Vertical swell (i.e. perpendicular to bedding) was found to always exceed horizontal swell in the 3D cube test, by up to four times. Maximum swell was typically achieved between 10 and 100 hours. Overall (volumetric) swell in the 3D cube test ranged from 0.4 to 14%, but was typically around 7%. Linear shrinkage ranged from 12 to 16%.
The majority of Standard Penetration Test (SPT) data represent the Reading Formation Mottled Clays and Upnor Formation Glauconitic Sands. The median N value for the undifferentiated Reading Formation Mottled Clays is 35 blows/mm with a range of 3 to 232 blows/mm.
- HEAD, K H. 1980. Manual of soil laboratory testing: Soil classification and compaction tests. Volume 1. Pentach Press. UK.
- BROWN, E T (ed). 1981. Rock characterisation testing and monitoring: ISRM (International Society of Rock Mechanics) suggested methods. Pergamon Press, Oxford, UK.
- BSI, 2007. BS EN 1997-2 Eurocode 7. Geotechnical design. Ground investigation and testing. British Standard Institute, London, UK. 196pp.
- HIGHT, D W, ELLISON, R A, and PAGE, D P. 2004. The engineering properties of the Lambeth Group. Report RP576 Construction Industry Research and Information Association (CIRIA), London.
- BRE. 1981. Concrete in sulphate-bearing soils and groundwaters. BRE Digest 250. BRE, Watford, UK
- BRE. 1986. Concrete in sulphate-bearing soils and groundwaters, Revised. BRE Digest 250. BRE, Watford, UK.
- BRE. 1991. Sulphate and acid resistance of concrete in the ground. BRE Digest 363. BRE, Watford, UK.
- BRE. 1995. Sulphate and acid resistance of concrete in the ground. Revised. BRE Digest 363. BRE, Watford, UK.
- BRE. 2005. Concrete attack in aggressive ground (Special Digest 1). Building Research Establishment, Garston, Watford.
- HEAD, K H. 1992. Manual of soil laboratory testing: Permeability, shear strength and compressibility tests. Volume 2. Pentach Press, UK.
- HEAD, K H. 1998. Manual of soil laboratory testing: Effective stress tests. Volume 3. Wiley Publishing.
- MAYNE, P W. 1980. Cam-Clay predictions of undrained strength. Association of Soil Mechanics and Civil Engineering, GT106, 1219–1242.
- LEHANE, B M, CHAPMAN, J P, PAUL, S P, and JOHNSON, J G A. 1995. The apparent variability of the Woolwich and Reading Beds. Proceedings of the XI European conference on soil mechanics and foundation Engineering, Copenhagen, 8, 95–102.
- LUPINI, J F, SKINNER, A E, and VAUGHAN, P R. 1981. Drained residual strength of cohesive soils. Geotechnique, 31, 181–213.
- VOIGHT, B. 1973. Correlation between Atterberg plasticity limits and residual shear strength of natural soils. Geotechnique, 23, 265–267.
- HIGHWAYS AGENCY. 1991. Specification for Highways Works. HMSO. London.
- CAPREZ, M, and HONOLD, P. 1995. Moisture condition as a basis for compaction. Proceedings of the XI European Conference on Soil Mechanics and Foundation Engineering — The Interplay between Geotechnical Engineering and Engineering Geology, Copenhagen, 3, 25–30
- GOSTELOW, T P. 1996. Some geological aspects of clay swelling and shrinkage. British Geological Survey, Engineering Geology Series, Technical Report, WN/95/16. British Geological Survey, Keyworth, Nottingham, UK.
- HOBBS, P R N, and JONES, L D. 1995. Methods for testing for swell and shrinkage of soils. British Geological Survey Technical report WN/95/15. Keyworth, Nottingham, UK.
- BRE 1993. Low rise buildings on shrinkable clay soil Part 1, BRE digest 240. Building Research Establishment, Garston, Watford.
- BRE 1990. Low rise buildings on shrinkable clay soil Part 2, BRE digest 241. Building Research Establishment, Garston, Watford.
- BSI. 2005. BS EN 22476-3. Geotechnical investigation and testing. Field testing. Standard Penetration test. British Standards Institute, London, UK. 26pp.
- STROUD, M A, and BUTLER, F G. 1975. The standard penetration test and the engineering properties of glacial materials. Conference on the engineering behaviour of glacial materials. University of Birmingham, UK. 124–135.
- STROUD, M A. 1989 ‘The Standard Penetration Test-its Application and Interpretation’. Institution of Civil Engineers Conference on Penetration Testing, Birmingham, United Kingdom. Thomas Telford, London, pp.29–49.
- CLAYTON, C R I. 1995. The standard penetration test (SPT): Methods and Use. R 143. Construction Industry Research and Information Association, London.