A monitoring system is only as good as what it can sense. The art of instrumenting a structure is choosing the right combination of sensors — each measures a different physical quantity, with its own strengths, blind spots and cost. No single device sees everything, so SHM almost always blends families.
The major sensing families
Broadly, SHM sensors fall into a handful of groups defined by what they measure: motion, deformation, stress waves, position, and full-field imagery.
Motion: accelerometers
Accelerometers measure how a structure vibrates. They are the workhorse of vibration-based SHM because the dynamic response — natural frequencies, damping, mode shapes — is rich in information about global stiffness. Modern MEMS accelerometers are cheap, small and low-power enough to deploy in dense networks; piezoelectric devices remain the choice when very high bandwidth or sensitivity is needed.
Deformation: strain and fiber optics
Where accelerometers see the whole structure move, strain sensors see local deformation at a specific point. Foil and vibrating-wire gauges give a calibrated, direct measure of stress at known critical details — exactly where fatigue cracks tend to start.
Fiber-optic sensing extends this idea dramatically. A single optical fiber can host hundreds of fiber Bragg gratings (FBGs), or act as a continuous distributed sensor along its entire length. Because the signal is light, it is immune to electromagnetic interference and well suited to long, buried or hard-to-reach assets such as pipelines, tunnels and dams.
Stress waves: acoustic emission
When a crack grows, it releases a burst of elastic energy. Acoustic-emission (AE) sensors listen for these transient stress waves, making AE one of the few techniques that detects damage <em>as it happens</em>, rather than inferring it after the fact. With an array of sensors, the source of an event can be triangulated — valuable for catching active cracking in steel and pressure vessels.
Position and imagery: GNSS and vision
Some movements are too slow and too large for accelerometers. Satellite positioning (GNSS) resolves the slow drift, settlement and thermal sway of tall buildings and long-span bridges over long baselines. Vision-based sensing, meanwhile, recovers full-field displacement and crack maps from ordinary cameras — non-contact, retrofittable, and increasingly powerful as computer vision improves.
A mature deployment might combine accelerometers for global dynamics, FBGs for distributed strain, AE around fatigue-prone joints, and GNSS for slow drift — each covering the others' blind spots.
Choosing well
Sensor selection is an engineering judgement, not a checklist. It starts from the failure modes you actually worry about, then asks which quantities those failures perturb, how fast, and over what area — and balances that against environment, power, data bandwidth and budget. The best system is rarely the one with the most sensors; it is the one whose sensors are pointed at the right questions.