How Is Building Settlement Monitoring Designed and Installed?

Settlement monitoring establishes survey points on structures and stable benchmarks, reads them on a schedule with precise leveling or automated sensors, and compares movement against engineer-set trigger levels — detecting foundation movement early enough to act before damage occurs. Foundations move. The engineering question is never whether there is movement but whether it is within what the structure tolerates. Monitoring converts that question from a philosophical debate into a plotted line with thresholds — and plotted lines settle arguments that opinions cannot.

Point layout comes first: monitoring points on the structure at corners, mid-spans, and anywhere movement is expected, plus deep benchmarks founded below the zone of influence that serve as the stable reference. Without a truly stable benchmark, every reading is relative to ground that may itself be moving — the most common design flaw in amateur programs. The reading schedule and the trigger levels are designed together: frequent readings during active work, tapering as readings stabilize, with alert and action thresholds tied to the structure's specific tolerance. Automated sensors get specified where continuous data justifies the cost — the decision is economic, weighing the sensor cost against the cost of the damage an early warning prevents. Every program I design includes the response plan: who gets notified at each trigger level, and what authority they have to stop work.

Failed programs share traits: no baseline readings before work started, benchmarks that were never verified stable, trigger levels nobody defined, and data that nobody reviewed until the cracks appeared. Monitoring hardware without engineering process is expensive decoration. What I build into every program.

• Take baseline readings before any causing activity — without a baseline, movement cannot be measured
• Found benchmarks below the zone of influence and verify they are truly stable
• Set alert and action trigger levels in advance, tied to the specific structure's tolerance
• Define the response plan: who is notified and who can stop work at each trigger
• Review the data on schedule and plot trends — a single reading is noise, a trend is information