Skip to content

Chapter 9 — Automated, Remote, and Real-Time Monitoring Systems

9.1 When to automate

Automation pays off when:

  • The dam is remote or hazardous to visit frequently.
  • High frequency data is needed (during filling, after quakes/floods).
  • Event capture matters (a storm or seismic event no inspector could attend).
  • Alarms must trigger immediately on threshold exceedance.

9.2 System components

A modern automated system has four layers:

  1. Sensors — the instruments of Chapters 5–8.
  2. Dataloggers — scan sensors, time-stamp, and store readings.
  3. Telemetry — cellular, satellite, radio, or fiber transports data off-site.
  4. Software — stores, validates, plots, and alarms.

9.3 Power and communications

Remote sites need power (solar + battery is common) and a communications path. Each adds failure modes: dead batteries, lost cellular coverage, and rodent-damaged cables are routine. Design with redundancy and self-diagnostics.

9.4 Thresholds and alarms

Automation is most valuable when it alerts. A good alarm system defines:

  • Action levels — values or rates of change that trigger review.
  • Escalation — who is notified, and how.
  • False-alarm management — filtering spikes from real trends to keep trust in the system.

Alarms that cry wolf get ignored

Over-sensitive alarms train operators to ignore them. Calibrate thresholds to meaningful changes and suppress known noise sources (temperature, tide, power dips).

9.5 Data validation at the edge

Automated systems should flag obviously bad data — out-of-range, frozen (no change), or dropouts — before it reaches the engineer. Validation rules catch sensor and communications failures early.

9.6 Real-time during events

During a flood or earthquake, real-time data lets the owner operate the reservoir and dispatch inspection where the data shows movement — turning a blind emergency into a managed one.

9.7 Key takeaways

  • Automate for remoteness, frequency, event capture, and alarms.
  • A system has sensors, dataloggers, telemetry, and software — each a failure point.
  • Design thresholds that alert on real change without crying wolf.
  • Validate data at the edge and use real-time data to manage events.