Dam safety monitoring · Live demo

Telling real dam movement from the daily rise and fall

Every dam breathes. It leans as the reservoir fills and drains, and it expands and contracts with the sun and the seasons. That constant, harmless motion hides the movement that actually signals danger. Senmos uses correlation analysis and virtual sensor channels to strip out the water-level and temperature effects — so alarms fire on genuine structural change, not on the daily tide.

Open the live demo Book a walkthrough

What actually matters for dam safety

A dam on a 3D map in Senmos with accelerometer, displacement, inclinometer, pressure, temperature and water level sensors pinned across the structure and canyon
Every instrument on the asset, in one place — the starting point for reading a dam.

A dam fails through a small number of well-understood mechanisms: seepage and rising pore pressure in the body or foundation, settlement and deformation of the embankment or structure, and displacement of the crest. The thing that drives almost all of the normal day-to-day movement is the reservoir water level — followed by temperature. Read raw, those drivers swamp the safety signal. Model them out, and the real story is left standing alone.

“We have been very happy with Senmos, it has been a welcome change onto a more modern and cheaper solution after working with Vista Data Vision in the past.”

Remigijus AbromavičiusCEO, GPS partneris — geotechnical monitoring

Isometric cutaway of an arch dam showing a reservoir water level gauge, a crack meter spanning a joint on the downstream face, and a wireless chain of inclinometers down the abutment slope

How a dam is instrumented

Sensors are placed where the failure mechanisms show themselves: in the body and slope to catch deformation, on the joints to catch cracking, and on the crest to catch displacement. The reservoir level gauge matters just as much — it is the signal everything else is correlated against.

Each instrument answers a different failure question — the slope string for deformation, the crack meter for joint movement, the level gauge for the driver everything else is measured against.

From a noisy sensor mix to a single safety signal

The full sensor mix into Senmos

Inclinometers/IPI strings, crack meters, GNSS/GPS, piezometers and the reservoir-level gauge stream into one Senmos project — automatic ingest, no manual import.

Correlation analysis

Senmos quantifies how strongly each safety signal tracks the reservoir level and temperature — identifying the drivers and how much of the daily movement they explain.

Virtual sensor channels

A calculated channel models the expected movement from water level & temperature, then subtracts it: residual = measured − modelled. What’s left is the movement those drivers can’t explain.

Automated alarms on the residual

Thresholds sit on the clean residual, not the raw signal — so an alarm means genuine structural change, not a full reservoir on a hot afternoon.

The intelligence layer — what makes the alarm trustworthy

Crack, GPS and displacement signals correlate with both water level and temperature at the same time — so plotted against either one alone, the scatter is wide and far from linear (medium correlation). Senmos models the temperature out and re-tests movement against water level: the cloud collapses onto a tight, straight line (high correlation). That clean relationship is what the alarm is built on.

Movement vs water level — raw medium correlation — temperature is mixed in R² ≈ 0.55 water level → movement → Movement vs temperature — raw medium correlation — water level is mixed in R² ≈ 0.50 temperature → movement → Senmos models the temperature out → then re-tests movement against water level Movement vs water level — temperature removed high correlation — a clean, linear relationship R² ≈ 0.95 water level → movement →
Figure — Movement plotted against water level (left) and against temperature (right) is a wide, non-linear scatter, because both drivers are mixed in (medium correlation). Once Senmos models the temperature out and re-tests against water level, the points fall onto a tight straight line (high correlation) — the clean relationship the alarm relies on.

See it live in Senmos

The demo is a fully working Senmos dam project with simulated data, so you can explore freely. Log in and you can:

Explore the dam in 3D

The dam model with every inclinometer, crack meter, GNSS unit and piezometer pinned where it sits.

Watch the slope deform

The IPI string and slope-monitoring view showing body and downstream-face movement.

Compare measured vs residual

The correlation/residual chart — raw signal, modelled effect, and the clean residual side by side.

Trace seepage & cracks

Crack-meter and piezometer history, with the reservoir-level overlay that drives them.

Close 3D map view of the dam crest in Senmos showing sensor pins and displacement direction arrows across the archThe Displacement overview dashboard in Senmos with live value panels for a mechanical displacement sensor and a joint crack meterLine plots in Senmos showing reservoir level against three pore pressure channels over 30 days, above a seismic acceleration plot with event spikes

Zoomed to the crest — displacement vectors across the arch

Open the live demo project Calculate your price

Demonstration data. This is a Senmos demo using simulated data to illustrate the platform — the dam, the sensors and the readings are synthetic. It is not live monitoring of an operational dam.