Structural Health Monitoring: What to Check Before You Commission a System

A monitoring proposal can look complete on paper and still leave the wrong system on the structure. The disappointments are rarely about sensor accuracy. They come from a system that measures something real but not the thing that governs, from data nobody is contracted to interpret, or from a platform holding five years of readings you cannot export. That is the quiet truth about instrumenting an asset: the best system is not the one with the most sensors, it is the one matched to a decision you actually need to make. This guide starts with the details that decide which is which.

What actually determines a good system, at a glance:

• The decision first: what you will do differently once the readings arrive
• The failure mode, not the catalogue: measure how the structure would genuinely deteriorate
• The sensor is the cheap part: installation, power, connectivity and interpretation carry the cost
• Someone has to read it: a monitoring contract without engineering is only a data contract
• Ownership and exit: raw data formats, platform access, and what happens at the end
• Site and regional realities: heat, dust, humidity, access and local approvals

Start with the decision, not the sensors

Monitoring begins with a question, not a specification. Before comparing systems, write down the decision the data will support. Whether to keep a building in service. Whether a crack is still moving. Whether a repaired element can be returned to full load. Which of forty bridges gets next year’s budget. If nobody can answer “what will we do differently when the readings arrive”, the system will produce data that nobody acts on, and the cost continues every year regardless.

This is also where structural health monitoring separates from assessment. An assessment gives you a verdict at a point in time. Monitoring gives you a trend, and a trend is only valuable if a threshold has been agreed in advance. ISO 13822, which sets out the basis for assessing existing structures, is the natural companion document here: it frames what you are trying to prove before you decide how to measure it.

Agree the thresholds before the first sensor is fixed

A monitoring scope should state, in writing, the alert levels, who is notified, how quickly, and what action each level triggers. Without that, an exceedance at two in the morning becomes an email that sits unread until Monday. Decide whether a threshold breach means “review at the next monthly report”, “attend site within 24 hours”, or “restrict occupancy now”. The engineering value of the system lives in that table, not in the sensor datasheet.

Be equally clear about the baseline. A structure instrumented after visible distress has no undamaged reference to compare against, so the first months of readings establish the baseline rather than judge it. That is a legitimate outcome, but it should be an expectation you set at the start, not a disappointment six months in.

Match the system to the failure mode

Deterioration mechanisms announce themselves in different measurements, and a system that is excellent for one can be nearly blind to another.

• Movement and settlement show up in tilt, displacement and levelling, sampled slowly but needing long-term stability.
• Crack activity needs displacement measurement across the crack itself, with temperature recorded alongside so thermal breathing is not mistaken for growth.
• Fatigue in steel connections requires strain measurement at the detail, at a sampling rate high enough to capture real stress cycles.
• Corrosion in reinforced concrete is followed with half-cell potential and resistivity, which describe the environment for corrosion rather than the section loss itself.
• Global stiffness change is inferred from vibration and modal behaviour, which is powerful at whole-structure scale and comparatively insensitive to small local damage.

That last distinction is worth dwelling on, because vibration-based methods are the most heavily marketed. They are genuinely strong for tracking a structure’s overall condition over years, which is why they suit bridges and long-span assets. The approach we set out in predictive bridge health monitoring depends on recognising a change in signature against a stable reference, and that is a different job from watching one crack in one wall. ISO 4866 and ISO 16587 cover measurement and performance parameters for this class of work.

The practical test for any proposal: ask the supplier to name the deterioration mechanism the system is designed to catch, and to explain what it would miss. A supplier who cannot answer the second half is selling hardware, not monitoring.

The sensor is the cheap part

Hardware is usually the smallest line in a monitoring budget. The cost sits in installation access, cable routing and containment, power provision, enclosures rated for the environment, calibration, commissioning, and the engineering time to turn readings into a conclusion somebody will sign.

Compare proposals on the full scope, not the sensor count. A quotation that lists sensors and a dashboard, and leaves installation, power and reporting as provisional sums, is not comparable to one that prices the whole delivery. Ask specifically what is excluded: scaffolding and access equipment, builder’s work, mains connection, network provision, annual recalibration and sensor replacement.

Who interprets the data?

This is the question that most changes the value of a monitoring contract. A dashboard shows you numbers. An engineer tells you what they mean and accepts professional responsibility for saying so. Those are different services and they are frequently sold under the same word.

Ask who reviews the data and how often. Ask what qualifications and professional indemnity cover sit behind the interpretation. Ask what a reporting cycle contains: raw plots only, or trend analysis against the agreed thresholds with a recommendation. Ask what happens on an exceedance out of hours. If the answer to all of these is a platform feature rather than a named responsibility, you are buying telemetry and will still need an engineer later.

Installation realities: power, connectivity and access

Power decides more of the design than most owners expect. Mains supply is dependable but needs a route and an isolator. Batteries free the layout but create a maintenance visit, and a sensor placed where access needs a mobile platform or a lane closure will quietly become the most expensive point on the structure. Solar helps outdoors, though panels need cleaning and the heat that comes with the sunlight derates them.

Connectivity is the second constraint. Wireless nodes reduce installation cost but need a gateway with real coverage, and basements, plant rooms, box girders and deep foundations are exactly where coverage fails. Confirm coverage by survey, not by assumption. In occupied buildings, also confirm what the works mean for tenants: drilling, hot works permits, out-of-hours access and making good.

Quick check: for every sensor on the layout, ask how somebody physically reaches it in three years to replace a battery or recalibrate. If that answer involves closing a lane, hiring a platform or evacuating a floor, the maintenance cost has just outgrown the hardware, and the sensor may be better placed somewhere slightly less ideal but reachable.

What it costs, and what shape that cost takes

Monitoring is rarely a single purchase. The realistic figure is a five-year total, and it usually combines an upfront capital element with a recurring one: the supply and installation, then platform or licence fees, data hosting, connectivity, periodic calibration, sensor replacement, and the engineering retainer for reporting.

The shape matters as much as the total. A capital-heavy model suits an owner with budget available now and a long horizon. A subscription model spreads cost and keeps the supplier engaged, but it should be checked against the monitoring duration you actually need. Short-term monitoring during construction, demolition next door, or a repair validation period has an end date, and paying for a platform beyond it is avoidable. Ask for the five-year total on every proposal, on the same assumptions, and compare those.

Regional factors in the UAE and the Gulf

Environment is not a footnote here. Summer temperatures affect both the equipment and the structure, and both effects need to be designed for.

On the equipment side, check the stated operating range of the electronics rather than the sensor alone, since gateways, batteries and power supplies commonly fail before sensors do. Confirm the ingress rating for dust, and for coastal and marine assets confirm the treatment of enclosures, fixings and cable glands against salt-laden humidity.

On the structural side, thermal cycling in the Gulf produces daily movements large enough to dominate a raw signal. A crack that appears to open and close by a substantial margin every day is usually reporting the weather rather than deterioration. Any credible system records temperature alongside the primary measurement and compensates for it, and any credible report separates thermal response from genuine trend. If a proposal does not mention temperature compensation, that is a significant gap rather than a detail.

Approvals and coordination deserve early attention too. Instrumenting an asset in service can involve the authority having jurisdiction, the operator, and existing warranties or maintenance contracts. Confirm who must approve fixings into structural elements before the layout is finalised.

Data ownership, platforms and exit

Monitoring generates a record that may outlive the contract, the supplier and the platform. Establish who owns it and in what form.

Ask whether you receive raw data or only processed output, and in which format. Ask whether there is export or an interface into your asset management, FM or building management systems. Ask what happens to the historic record if you change supplier, and whether the readings remain readable outside the original platform. A long baseline is one of the most valuable things a monitored structure has, and losing it at a contract change resets years of evidence.

Monitoring supports engineering decisions; it does not replace inspection or engineering judgement. Sensors report a small number of points on a large structure, and they report what they were placed to report. Treat an alert as a reason to investigate rather than a verdict, and keep periodic inspection and structural assessment in the programme alongside the instrumentation, particularly where a system presents its output in confident visual language.

A final check before you commission

Before signing a monitoring scope, confirm six practical points:

• Decision: the action each alert level triggers, written down, with named responsibilities.
• Mechanism: the deterioration the system is designed to catch, and what it will miss.
• Scope: the full delivered cost, with access, power, connectivity and reporting included.
• Interpretation: who reviews the data, how often, and who is professionally responsible.
• Environment: temperature compensation, ingress and corrosion protection suited to the site.
• Data: ownership, export format, and what survives a change of supplier.

Then give the engineering its proper place. A monitoring system is not a product you install and admire; it is an instrument that should quietly narrow uncertainty about an asset you are responsible for. The right one produces fewer numbers than you expected and more decisions than you thought possible, and it earns its keep on the day it tells you that something has changed while the change is still inexpensive to address.

If you are scoping instrumentation for a building, bridge or industrial asset, ESG advises on structural health monitoring from the decision stage through specification, interpretation and reporting. The companion piece on when instrumentation is justified is a useful next read if you are still deciding whether monitoring is the right response at all.