Civil engineering projects in Kenya demand rigorous quality assurance to ensure safety, durability, and compliance with standards. Preparing samples for testing is a critical step in material evaluation, influencing project outcomes from foundations to superstructures. This chapter explores how to correctly prepare and collect samples for testing, emphasizing methods suited to the Kenyan civil engineering context, such as construction materials testing for aggregates, concrete, and soils.

3.1 Sample Tests

Sample testing forms the backbone of quality control in civil engineering projects. It provides objective data on material properties, enabling engineers to make informed decisions on material suitability and project compliance. In Kenya, adherence to recognized testing standards ensures that infrastructure like roads, bridges, and buildings meet safety and durability requirements.

Testing has two halves, and both must be done correctly or the result is worthless. First the material is sampled, so that a small portion truly represents the whole delivery or source. Then that sample is tested, to measure the property the specification calls for. A perfect test on an unrepresentative sample gives a confident wrong answer.

Definition of Terms

• Population or lot: The whole quantity of material being judged, such as one delivery of aggregate or one day's concrete pour.
• Sample: The portion taken from the lot for examination.
• Increment: One scoop or portion taken from the lot. Several increments are combined to make a sample.
• Representative sample: A sample whose properties match those of the lot it came from.
• Reduction: Cutting a large field sample down to the smaller quantity a test needs, without changing its character.
• Specification: The document stating the properties the material must have, and the tests used to prove them.
• Compliance: Meeting the specification. Material that fails is rejected or accepted with a reduction in payment.
• Chain of custody: The written record of who held the sample, from the field to the laboratory.

3.1.1 Methods of Sampling

Sampling methods in civil engineering are designed to obtain representative material portions that reflect the whole batch or source. Choosing the correct method affects the reliability of test results, impacting decisions on material acceptance or rejection.

Random Sampling

Random sampling involves selecting samples without any bias towards any part of the material batch or source. This method ensures every unit has an equal chance of selection, reducing systematic errors.

• Unbiased Representation: By giving all material units equal selection probability, random sampling captures the overall variability within the batch.
• Applicability: Commonly used for bulk materials like aggregates at quarries or stockpiles in construction sites.
• Implementation: For example, a contractor at a Nairobi construction site may divide a stockpile into grids and randomly select samples from different grids.
• Limitations: It may miss localized defects if the sample size is too small.
• Quality Control: Random sampling supports statistical analysis for material acceptance criteria.

Systematic Sampling

Systematic sampling selects samples at fixed intervals from a material batch or production line. This method is simpler to implement than random sampling and can detect patterns of variability.

• Consistent Intervals: Samples are taken at regular intervals, such as every 10th bag of cement delivered to a site.
• Ease of Use: Construction supervisors at county government projects use systematic sampling for routine checks on concrete mixes.
• Pattern Detection: Helps identify recurring quality issues linked to specific production times or sources.
• Risk of Bias: If there is a hidden pattern matching the sampling interval, results may be skewed.
• Efficiency: Often preferred for continuous processes like asphalt mixing.

Stratified Sampling

Stratified sampling divides the population into homogenous subgroups or strata, then samples are drawn from each stratum. This technique improves representativity when material batches are heterogeneous.

• Subgroup Identification: For example, soil samples from different depths or zones in a site are treated as separate strata.
• Enhanced Accuracy: Ensures that all material variations are adequately represented in testing.
• Practical Use: At a coffee cooperative constructing a processing plant, stratified sampling of soil ensures foundation design accounts for varying soil properties.
• Sample Allocation: Samples per stratum may be proportional to the stratum size or importance.
• Complexity: Requires prior knowledge of material variability and extra planning.

Composite Sampling

Composite sampling combines multiple individual samples into one aggregate sample for testing. It is cost-effective and useful when materials are uniform.

• Sample Mixing: For instance, several small concrete samples taken from different batches can be mixed to form a composite sample.
• Cost Reduction: Reduces the number of tests needed, saving resources.
• Uniformity Assumption: Works well when material properties are consistent.
• Risk of Dilution: Localized defects may be masked in the composite sample.
• Use Cases: Common in testing cement quality delivered to a large hotel construction site.

Taking the Sample in the Field

Whichever method is chosen, the physical act of taking the sample follows a set procedure.

From a stockpile:

1. Do not sample from the surface or the base of the pile. Fine material washes down the sides and collects at the bottom, and the surface is weathered and contaminated.
2. Cut into the pile at several points, at least three, spread around it and at different heights.
3. Remove and discard the outer 150 mm of material at each point before taking the increment.
4. Combine the increments into one field sample.

From a lorry or wagon:

1. Take increments from at least three separate positions across the load.
2. Dig down at least 300 mm below the surface at each position.
3. Avoid the corners and edges, where segregation is worst.

From a conveyor belt:

1. Stop the belt.
2. Insert a sampling frame across the full width of the belt.
3. Remove all the material between the frames, including the fines swept from the belt surface.

From bagged material such as cement:

1. Select bags at random through the stack, not only from the top layer.
2. Sample with a sampling tube pushed diagonally through the full depth of the bag.

From fresh concrete:

1. Take the sample from the discharge of the mixer or truck, never from the beginning or the very end of the discharge.
2. Take at least four increments spread through the discharge.
3. Remix the combined sample on a clean, non-absorbent surface before filling any moulds.
4. Begin testing within a few minutes, since concrete starts to stiffen at once.

Reducing a Sample to Test Size

A field sample is far larger than any test needs. It must be reduced without changing its grading, since simply scooping off the top takes the fines and leaves the coarse.

Quartering method:

1. Tip the sample onto a clean, hard, level surface.
2. Mix it thoroughly by turning the whole heap over three times.
3. Flatten the heap into a circular cake of even thickness.
4. Divide it into four equal quarters with a straight edge.
5. Discard two diagonally opposite quarters and combine the remaining two.
6. Repeat until the sample is the required size.

Riffle box (sample divider): The sample is poured evenly across a box of alternating chutes that discharge into two pans. Each pan receives half the sample. One half is kept and the other set aside, and the process is repeated until the right quantity remains. This is quicker and more accurate than quartering.

Labelling, Handling and Storage

A sample without a label is not evidence. Each sample is sealed and labelled with:

• Project name and contract number
• Material description and source, including quarry or supplier
• Sampling location and the point on the lot it came from
• Date and time of sampling
• Quantity the sample represents
• Name and signature of the person sampling
• Sample reference number
• Test or tests required

Samples are protected from rain, sun, contamination and loss of moisture in transit. Soil samples for moisture content go into airtight containers and are weighed as soon as possible. Concrete cubes are kept damp and undisturbed for the first 24 hours before being taken to the laboratory. A duplicate sample is often kept sealed on site, so that a disputed result can be checked against it.

Tests on Aggregates

Aggregate makes up around three quarters of the volume of concrete, so its quality largely sets the quality of the concrete.

Sieve analysis procedure: A dried, weighed sample is passed through a nest of sieves arranged coarsest at the top. The sieves are shaken mechanically for a set period. The mass retained on each sieve is weighed, and the cumulative percentage passing each sieve is calculated and plotted as a grading curve against the specification limits.

Worked example. A 1000 g sample of fine aggregate is sieved and the mass retained on each sieve recorded.

Percentage passing = 100 − cumulative percentage retained. Plotting these values gives the grading curve, which is compared with the envelope in the specification. This sand is well graded, with a spread of sizes and no single size dominating.

Silt content (field test): Place 50 mm of sand in a clear jar, fill with water, shake and leave to stand for one hour. The silt settles as a distinct layer above the sand. Silt content is the depth of the silt layer as a percentage of the sand layer. More than about 6 to 8 per cent means the sand must be washed.

Tests on Cement

On site, cement is also judged by colour, by the absence of hard lumps, by the bag feeling cool rather than warm, and by the manufacture date printed on the bag.

Tests on Fresh Concrete

These are carried out at the point of placing, before the concrete stiffens.

• Slump test: A standard cone 300 mm high is filled in three layers, each rodded 25 times, then lifted vertically. The distance the concrete settles is the slump in millimetres. It measures workability and, more usefully, consistency between batches. A sudden change in slump between loads signals a change in water content or grading.
• Compacting factor test: Concrete falls through two hoppers into a cylinder. The ratio of the mass of partially compacted concrete to fully compacted concrete gives the compacting factor. More sensitive than the slump test for dry mixes.
• Vebe consistometer: Measures the time for a concrete cone to remould under vibration. Suited to very dry mixes.
• Air content: Measures entrained air, which affects both durability and strength.
• Temperature: Recorded because hot concrete stiffens faster and gains less ultimate strength.

Tests on Hardened Concrete

• Compressive strength (cube test): The standard acceptance test. Cubes 150 mm are cast from the sampled concrete, cured in water at a controlled temperature, and crushed at 7 and 28 days. The 7-day result gives early warning, and the 28-day result decides compliance.
• Tensile splitting test: A cylinder is loaded across its diameter until it splits, giving indirect tensile strength.
• Flexural strength: A beam is loaded to failure, used mainly for concrete pavements.
• Non-destructive tests: The rebound hammer estimates surface hardness, and ultrasonic pulse velocity indicates uniformity and voids. Both are used on structures already built, where cubes are no longer possible, but neither replaces the cube test.

Cube test procedure:

1. Fill the mould in three layers, compacting each layer with 35 strokes of a standard rod or on a vibrating table.
2. Level the top surface, and mark the cube with its reference and date.
3. Leave the mould covered and undisturbed for 24 hours.
4. Strip the mould and place the cube in a curing tank at about 20 degrees Celsius.
5. Remove at the test age, wipe surface water, weigh, and measure.
6. Place in the compression machine with the trowelled face to the side, not against the platens.
7. Apply load at a steady rate until failure, and record the maximum load.

Compressive strength = failure load ÷ cross-sectional area

Worked example. A 150 mm cube fails at a load of 675 kN.

Area = 150 × 150 = 22 500 mm²
Load = 675 kN = 675 000 N
Strength = 675 000 ÷ 22 500 = 30 N/mm²

This satisfies a specified grade of C25 concrete, whose characteristic strength is 25 N/mm², since individual results are expected to exceed the characteristic value by a margin.

Tests on Soils

Moisture content = (mass of water ÷ mass of dry soil) × 100

Worked example. A soil sample weighs 250 g wet and 220 g after oven drying at 105 degrees Celsius.

Mass of water = 250 − 220 = 30 g
Moisture content = (30 ÷ 220) × 100 = 13.6 per cent

Note that the divisor is the dry mass, not the wet mass. Dividing by the wet mass is the commonest error in this calculation and gives 12 per cent, which is wrong.

Plasticity index = liquid limit − plastic limit. A high plasticity index means the soil swells and shrinks greatly with moisture, which makes it unsuitable as a subgrade without treatment.

Tests on Steel Reinforcement

• Tensile test: A sample bar is pulled to failure to find yield strength, ultimate tensile strength and elongation. Elongation matters as much as strength, since a bar that snaps without warning is dangerous.
• Bend and rebend test: A bar is bent through a specified angle around a former and straightened, then examined for cracking. It proves the steel can be bent on site without fracturing.
• Mass per metre: A measured length is weighed and compared with the nominal mass for that diameter. Underweight bars are a common form of substandard supply, since a bar rolled slightly undersize looks correct but carries less load.
• Dimensional and rib check: Confirms the diameter and the rib pattern that gives bond with the concrete.

Tests on Bricks and Blocks

• Compressive strength: Units are crushed in a testing machine and the average strength reported.
• Water absorption: Units are weighed dry, soaked for 24 hours and reweighed. The gain as a percentage of dry mass indicates porosity.
• Dimensional tolerance: Twenty units are measured end to end and the average compared with the nominal size.
• Efflorescence: A unit is stood in water and allowed to dry, then examined for white salt deposits.

Reporting Test Results

A test result that cannot be traced back to a specific lot of material is of no use in a dispute. Every report records:

• Sample reference, date sampled and date tested
• Material description and source
• Test method and the standard followed
• All measured values, not only the final figure
• The specification requirement and whether the sample complies
• Name and signature of the person carrying out the test
• Any departure from the standard method, and the reason

Where a result fails, the material is not simply rejected on one test. The specification normally allows a retest on a fresh sample from the same lot, and only a second failure condemns it.

Chapter Summary

This chapter explored the fundamental aspects of preparing samples for testing, beginning with an examination of various sample tests and the methods of sampling that ensure representative and reliable results. It then detailed the role of standard manuals and procedures, providing an overview of the sampling process and emphasizing the critical steps of preparation before collection alongside techniques for effective sample collection. The discussion advanced to the tools and apparatus used in sample testing, highlighting the different types and their specific applications in achieving accurate analysis. Further, the chapter underscored the importance of standard manuals in guiding sample collection, clarifying their purpose in maintaining consistency and reliability. It also outlined the different types of standard manuals available to practitioners and identified key organizations responsible for publishing these standards. Together, these topics provide a comprehensive framework for understanding the preparation of samples that meet both technical and regulatory requirements in testing environments.