eduKate Singapore

VIEW THIS AS

Auto mode follows the Route Engine until you choose a viewpoint.

YOU ARE HERE

ROUTE CHECK

CONNECTED TO

WHAT NEXT

Use the canonical route for this room, or HELP if you are unsure.

Next useful routeCanonical hubHELP

The Gold Standard of Geotechnical Monitoring

—

by

eduKate Asia

An instrument beside a deep excavation reports that a retaining wall moved six millimetres. Is that safe, unsafe or simply a normal response anticipated by the design? Without the reference, location, time, rate of change and expected behaviour, the number alone cannot answer. Geotechnical monitoring exists to connect ground movements with informed engineering decisions.

The gold standard of geotechnical monitoring is a programme that measures the right physical quantities, preserves their meaning, tests design assumptions and gives qualified people time to respond. This guide develops eighteen quality controls, two original fictional examples, practical review questions and a linked route through the geoscience and engineering library.

This is part of The Gold Standard. “Gold standard” is an editorial framework for evidence and professional care, not an engineering certificate, approval of a construction method or declaration that any real site is safe.

Quick information: what reliable ground monitoring requires

  • An identified structure, geological model and credible mechanism of concern.
  • Sensors suited to actual movement, pressure or force questions.
  • Stable references, baseline measurements and documented accuracy.
  • Appropriate observation frequency and quality checks.
  • Site-specific review triggers and named engineering decision-makers.
  • A traceable response and correction path when observations change.

Chapter index

1. Begin with a ground-behaviour question and its consequence

Back to contents

Geotechnical monitoring observes how soil, rock, groundwater and structures interacting with the ground behave before, during and after works. It is essential where a design assumption may be uncertain or construction can change conditions. The goal is not to install the largest number of sensors. It is to detect meaningful changes early enough for competent people to interpret and act.

Specify what needs protection, which deformation or pressure change could matter, who makes engineering decisions and what evidence they will need. An excavation beside an occupied building presents different questions from an embankment settling on soft soil. The gold standard links a known engineering concern to measurements, reviewed limits and a credible response pathway.

2. Write the ground model before choosing instruments

Back to contents

A ground model describes soil and rock layers, groundwater, relevant structures, possible slip surfaces, loads and uncertainties. Sensors selected without this model may record parameters that have little relationship to the most consequential failure mode. A row of settlement points cannot directly reveal the pore-water pressure distribution that affects effective stress.

Start with boreholes, geological interpretation, previous ground movement, nearby foundations and the planned construction sequence. Record what has been measured and what remains inferred. The instrument plan should test the assumptions whose failure would change the design or construction decision.

3. Use ISO 18674 within its stated scope

Back to contents

ISO 18674-1:2015 lays out general rules for geotechnical field instrumentation and performance monitoring. ISO’s catalogue still identifies it as current, although specialised parts of the series cover distinct instruments. Extensometers are addressed by Part 2, inclinometers by Part 3 and piezometers by Part 4.

A standard provides a common technical frame; citing it does not certify an installed system or determine every project threshold. Match each method and its current edition to the instrument, ground condition and engineering task. Local approvals and project-specific professional requirements remain necessary.

4. Differentiate displacement, settlement, strain and tilt

Back to contents

Settlement is vertical movement relative to a specified reference; displacement can be vertical or horizontal, and tilt describes change in orientation. Strain concerns deformation relative to an original dimension. These quantities can be related but are not interchangeable, and a sensor measuring one cannot simply be relabelled as another.

State the measurement direction, reference point, sign convention and units. A displacement of ten millimetres over ten metres is different from a ten-millimetre differential over one metre. The spatial distribution of movement can matter more than an isolated largest reading.

5. Use settlement points and geodetic control correctly

Back to contents

Survey points can track relative or absolute elevation changes, depending on the control network and reduction method. A datum that itself moves can make every measured building point appear stable or produce a spurious trend. Independent control outside the affected zone is therefore important for appropriate projects.

Preserve survey method, benchmark identity, observation timing and expected measurement precision. Avoid treating repeated decimal places as proof of accuracy. When several points move together, investigate whether the reference has shifted before deciding the ground has settled uniformly.

6. Use inclinometers for subsurface movement profiles

Back to contents

Inclinometers can help locate and characterise lateral ground movement along an instrumented line, such as the behaviour of a retaining wall or slope. A local bend in the profile may suggest a depth range where movement is concentrated, but installation geometry and instrument error can alter the apparent result.

Compare sequential readings with the baseline and inspect the full profile, not only the largest displacement number. Verify instrument orientation, casing condition, reading method and whether the interpreted shear plane is independently supported. A reading suggesting change should trigger qualified review rather than an automatic claim of imminent failure.

7. Use extensometers to measure change along a line

Back to contents

Extensometers measure displacement between selected points or along a defined path, depending on their design. ISO 18674-2:2016 covers line-displacement measurement with extensometers and was confirmed current in the ISO catalogue. The method can help examine settlement, convergence or displacement within soil, rock and structures.

Identify which anchors are assumed stable and whether the installation spans the zone of interest. An extensometer outside the active movement region may show almost nothing while other parts of the site change. The location of the measurement is part of its meaning.

8. Use piezometers to understand pore-water pressure

Back to contents

Piezometers measure pore-water pressure or hydraulic head within their stated installation and response characteristics. Under suitable assumptions, pressure can be related to an equivalent column of water. In saturated soil, pore pressure affects effective stress and therefore can influence stability and settlement.

A piezometer is not a direct measurement of soil strength. Screen location, response time, calibration and changing groundwater conditions matter. Instruments at different depths or in different strata may show real differences that should not be averaged into one generic water table.

9. Consider loads and earth pressures separately

Back to contents

Load cells, pressure cells and strain measurements can describe structural or soil interaction under the relevant measurement configuration. Total earth pressure is not the same as effective stress or pore-water pressure. Interpreting these together may help test design assumptions, but a single gauge cannot fully characterise forces throughout a wall or foundation.

Record sensor geometry, installation conditions and calibration. Load redistribution can occur without a simple one-to-one change in the monitored pressure. Compare observations with realistic structural and geotechnical models rather than assuming every surprising reading proves one particular failure mechanism.

10. Establish the baseline before site works change conditions

Back to contents

A baseline is a set of observations collected under known conditions before the intervention of interest. It should be sufficient to estimate ordinary fluctuations and instrument behaviour where practical. A first reading taken after an excavation has advanced is not a clean pre-excavation baseline.

Collect dates, construction stage, groundwater conditions, weather and instrument status with each measurement. Seasonal changes can resemble project effects, and construction effects can be missed if observation intervals are too long. A transparent baseline helps separate actual change from drifting measurement systems.

11. Plan observation frequency around plausible change rates

Back to contents

A slowly settling embankment and a rapidly advancing deep excavation may need different monitoring intervals and response expectations. Automated systems can support frequent measurement but also create large volumes of noisy or invalid data. A high sampling frequency does not guarantee timely interpretation.

Set observation timing, quality-control review, personnel availability and escalation according to the engineering question. Define how missed transmissions or inaccessible sensors will be handled. The interval between meaningful observations should be short enough for the approved response plan, not merely convenient for monthly reporting.

12. Distinguish accuracy, precision, resolution and drift

Back to contents

Sensor resolution describes the increment it can display; precision concerns repeatability; accuracy concerns agreement with a suitable reference; drift is a change in the measurement system over time. A device showing 0.01 millimetre increments is not automatically accurate to that level under site conditions.

Use calibration records, reference checks, redundant observations and environmental corrections where appropriate. A sudden change confined to one sensor can be a real local effect or a fault. Quality control should flag questionable data while keeping the original record for investigation.

13. Convert raw signals into correctly referenced measurements

Back to contents

The reading displayed by an instrument may require a calibration curve, temperature adjustment, geometry correction or conversion from electrical units. Every conversion can introduce error. Data pipelines should preserve the raw input and the formula or coefficient version used to calculate the engineering quantity.

Test representative values by independent calculation and make units explicit. A reversed inclinometer axis or wrong piezometer calibration can create a persuasive-looking but physically misleading time series. Silent edits to past measurements destroy the ability to reconstruct the site history.

14. Interpret patterns across instruments and site events

Back to contents

A credible conclusion often uses more than one signal: changing wall movement, pore pressure, ground settlement, excavation stages and rainfall may form a coherent pattern. Correlation alone cannot establish that one activity caused the movement, and stable readings do not prove a structure has no hidden risk.

Plot relevant observations against actual construction and environmental events, preserving when measurements were taken rather than merely when data were received. Seek spatial consistency and independent confirmation. An engineer should state competing explanations and which new observation could distinguish them.

15. Set trigger levels with engineering authority

Back to contents

Review levels and actions need to be developed by qualified professionals from design assumptions, site risks, expected ground behaviour and regulatory conditions. They are not universal settlement or tilt thresholds applicable to every building. A trigger may require increased measurement, inspection, design review, work modification or a protective action.

Write who receives an alert, how quickly it is reviewed, which data are checked and who has authority to act. The numerical trigger itself is only one element. An automated email cannot replace an engineering decision, and a system that lacks a staffed response path is not dependable.

16. Keep a credible stop-work and emergency route

Back to contents

If observed conditions indicate possible danger, the response must follow the approved site safety plan and relevant statutory duties. Workers and nearby people need clear communication and access to competent on-site personnel. A technical dashboard should not discourage legitimate reports of visible cracking, movement or other hazards.

A good monitoring programme integrates instrument exceptions with inspections and professional judgement. Do not ignore a worrying observation because it has not yet crossed a generic percentage threshold. Conversely, a questionable sensor signal needs urgent validation appropriate to the consequences, not panic based solely on unreviewed data.

17. Audit the chain from sensor to corrective action

Back to contents

A monitoring report should show which instrument observed the change, when it occurred, how the signal was converted, what checks were completed, which design expectation was compared and which action followed. FHWA geotechnical guidance emphasises standardised data collection, timely interpretation and communication with the design team.

Maintain an audit trail of maintenance, anomalies, corrections, alerts and responsible reviewers. A successful monitoring system is judged partly by whether decisions improved as the data changed, not by the amount of information archived. The record should survive a handover to another site team.

18. Transfer the monitoring method to a different ground problem

Back to contents

Soft-ground settlement, unstable slopes, tunnel works and retaining walls involve different mechanisms and appropriate sensor placements. Their monitoring principles transfer, but thresholds, instruments and interpretation cannot be copied blindly. Geological structure and potential consequences should drive the plan.

Start again with failure modes, measurement quantities, stable references, baselines, review intervals and named decision owners. The gold-standard outcome is a project where meaningful ground behaviour is observable, understood and responded to responsibly—not a claim that an instrument can guarantee safety.

Worked fictional case A: movement speed versus total movement

An imaginary wall-monitoring point reads 2 millimetres of lateral movement at the start of day 1, 5 millimetres at the start of day 3 and 11 millimetres at the start of day 4, all relative to the same baseline. Over the first two days of elapsed time, change is 3 millimetres, or 1.5 millimetres per day on average. During the following one day, change is 6 millimetres, or 6 millimetres per day. The average rate has increased fourfold between these two selected intervals.

That rate difference is a diagnostic signal, not a universal stop-work threshold. The review must consider sensor error, actual measurement times, excavation stage, groundwater changes and the engineer’s approved action criteria. The points may not describe a uniform acceleration during each interval. All readings are fictional and should not be applied to any real worksite.

Worked fictional case B: pressure is not the same as elevation

A fictional pressure transducer in a saturated zone reports gauge pore-water pressure of 19.62 kilopascals. Under a simplified static-water assumption with unit weight of water 9.81 kilonewtons per cubic metre, the equivalent pressure head is 19.62 ÷ 9.81 = 2.0 metres of water. This pressure head is measured relative to the transducer elevation, not directly as an absolute site water-level elevation.

If the transducer reference elevation is 15.0 metres on a declared datum, the corresponding piezometric head would be 17.0 metres under this model. This does not establish the ground-water table everywhere, soil strength or any factor of safety. Temperature, salinity, pressure reference, calibration and geological connection can affect field interpretation.

Monitoring audit before releasing a conclusion

  • Define the ground movement or pressure mechanism and potential consequences.
  • Verify instrument type, installation zone, units and stable reference.
  • Retain raw readings, calibrations, environmental context and baselines.
  • Recheck outliers and missing telemetry without deleting source evidence.
  • Show temporal rates and spatial patterns as well as cumulative readings.
  • Compare observations against site-specific design expectations.
  • Record named alert recipients, review periods and competent authorities.
  • Document safety actions, later corrections and remaining uncertainty.

Frequently asked questions about geotechnical monitoring

What is geotechnical monitoring?

Observation of soil, rock and geotechnical structure behaviour using appropriately selected field measurements and inspections.

What is the purpose of a settlement marker?

To track vertical displacement relative to a clearly identified reference over time.

What is a piezometer?

An instrument used to measure pore-water pressure or hydraulic head at a defined location or zone.

What does an inclinometer measure?

It helps describe angular or lateral displacement along a monitored line, often across depth.

Are sensor readings sufficient to declare a slope safe?

No. They need context, engineering interpretation, inspection and an appropriate safety assessment.

Can one numerical settlement trigger be used everywhere?

No. Trigger values depend on site conditions, structural sensitivity, design assumptions and approved procedures.

Why is baseline monitoring important?

It shows conditions and instrument behaviour before the change being studied.

What if a sensor stops reporting?

The site needs an established exception and alternative review procedure matching the safety consequences.

What is instrument drift?

A change in a sensor’s output unrelated to an equivalent change in the physical quantity measured.

Does high sampling frequency guarantee early warning?

No. Signals must be reliable, timely reviewed and linked to responsible action.

What is the observational method?

An engineering approach that uses observations to verify assumptions and adjust decisions under preplanned controls.

What does a good monitoring report contain?

Instrument identity, reference, readings, quality checks, interpretation, thresholds, actions and unresolved uncertainty.

Connected Gold Standard reading

Geotechnical engineering supplies design and failure-mode analysis, hydrogeology explains pore pressures, geomatics supports positioning, and human factors explains whether alerts reach decision-makers. These complementary guides help a reader understand which part of the evidence needs further work.

Official technical sources and responsible use

ISO and FHWA source pages were checked on 29 October 2026. ISO 18674-1:2015 remains a published general-rules reference, while instrument-specific parts have their own editions. This guide does not prescribe universal alert limits. Both numerical scenarios are fictional educational calculations and not an engineering interpretation, permit or operational safety assessment.

Return to chapter index · The Gold Standard hub · How X Works