During an excavation, as the ground redistributes stress, buildings in the surrounding area can tilt to a small degree. Tilt monitoring is used during these works to establish whether these changes stay in the norm or if the buildings require intervention as the movement becomes too much to absorb and the building's health becomes compromised.

The baseline

Since most existing buildings have a natural lean, each sensor mounted needs to be levelled and zeroed against a surveyed reference before the monitoring begins. A masonry building from the 1920s can have a historic tilt of 10 or 20 mm/m, and this sort of pre-existing tilt is not important for excavation monitoring.

A baseline is also a characterization of the building's normal movement with temperature and humidity changes. For example, on a sun-exposed concrete face with a daily temperature swing of 15 to 20 °C, the tilt signal can reach 0.3 to 0.6 mm/m peak-to-peak before compensation, structure and sensor combined. That's why the data from the first forty-eight hours is not to be used as actual tilt data, but should be read as the sensor's thermal stabilization after installation. One to two weeks of pre-excavation logging, ideally spanning at least one weather change, gives the diurnal amplitude, the phase lag between air temperature and structural response, and the weekday-versus-weekend pattern if the building is occupied.

Since monitored buildings are usually not the contractor's, it is best to use wireless sensors for excavations. A battery-powered LoRaWAN tiltmeter is installed in one visit, transmits for years without any intervention, and doesn't require a cable that would need to cross the property. A similar advantage is applied to settlement monitoring, which you can read more about in our article on monitoring Miami's high-rises with wireless sensors.

The minimum scheme for monitoring a building near an excavation is one biaxial sensor near the foundation, one at mid-height on the excavation-facing side, and one on the far side to detect differential behavior. On taller buildings, sensors at multiple levels on the near facade let a tilt profile be reconstructed with height, which distinguishes foundation rotation from racking of the upper storeys. The Tiltmeter Chain Tool in MyMove builds these profiles.

Trigger values

Trigger values are set by engineers against the condition of a specific building, its construction type, its foundation system, its pre-existing damage. Unlike vibration monitoring, which has DIN 4150-3 with published PPV guideline values, tilt monitoring has no standard values that can be used universally.

Monitoring plans are usually three-tier structures that show the classical serviceability limits. Skempton and MacDonald's 1/500 of angular distortion, which is 2 mm/m (1 mm/m equals 1 mrad), is still the traditional design limit against cracking in ordinary structures, so alert levels are usually below it, around a third to a half of the limit. Action levels are usually at or near 2 mm/m, while alarm levels are set for any value above the action level.

Tilt needs to be monitored at several points. Boscardin and Cording showed in 1989 that rigid-body tilt is relatively innocuous, while angular distortion combined with horizontal strain can cause cracks. So a building rotating uniformly at 3 mm/m may be fine while one distorting at 1.5 mm/m across a single bay is not, which is why tilt is monitored at several points.

The alert level must have a margin above the baseline changes of the structure. For example, if the facade has an amplitude of 0.4 mm/m every afternoon and the alert goes off at 0.5, the system could send alarms daily, which will then desensitize the monitoring team to the alarms. Such a situation could lead to legal issues, since the alarms would be documented but no action would be taken by the company. Every cumulative trigger should also be paired with a rate trigger, because the same cumulative reading carries a different meaning at different speeds. Rotation can accumulate slowly and in step with the excavation program, but it can also be the same total in the 48 hours after a failure of a dewatering pump.

The response plan

A response plan is a short document, agreed and signed before excavation starts, that states for each trigger level who is notified, within what time, and what happens on site.

The first action at any exceedance is to verify the instrument. Sometimes a tilt spike can happen because of temperature and not a general tilt of the structure. An isolated jump on one node, or thermal drift, can happen and might mean the sensor detected a problem with the mounting. Some tiltmeters send a temperature reading along with the tilt measurement to account for thermal drift in the analysis of the situation.

The escalation ladder in most plans:

• For alerts: Notify the monitoring engineer and the structural engineer the same working day, raise the logging frequency, and record what the excavation was doing at the time (stage, dewatering status, surcharge near the boundary).
• For actions: Notify the owner of the monitored building under the pre-construction agreement, pause or modify works on the affected boundary.
• For alarms: Stop the works on the affected zone until the engineer of record has reviewed the data, inspected the building, and cleared continuation in writing.

The plan should define its own end. When excavation and basement works are complete, a closure comparison between the pre- and post-construction baselines documents whether permanent rotation occurred.

Frequently Asked Questions

Is a 48-hour baseline ever enough?Only for the sensor, never for the building. Two days tells you the tiltmeter has thermally stabilized; it cannot capture the structure's diurnal envelope across changing weather, which is what triggers are set against. If the pre-works window is compressed, keep logging through early low-risk stages and refine the triggers before bulk excavation reaches the sensitive depth.

Can trigger values be taken from a standard, the way PPV guideline values come from DIN 4150-3?No. Vibration limits generalize across buildings well enough for standards to publish tables; tolerable tilt depends on the specific structure's stiffness, foundation, and existing condition. The published mm/m ranges are starting points for the engineer of record, not compliance values. Citing "1.0 mm/m per industry practice" without a condition assessment of the building is the most common defect we see in monitoring plans attached to tenders.

Do the tiltmeters come off the building when excavation ends?Usually after the closure report, not immediately. Ground movement continues after bulk excavation as pore pressures re-equilibrate, particularly in fine-grained soils, and basement construction itself loads the ground. Keeping the network live through structural completion costs little with wireless sensors already in place.

‍