Overflow monitoring every event.

What is overflow monitoring?

Overflow monitoring is the continuous measurement of water level at the points where a sewer or stormwater network is designed to spill: combined sewer overflows, storm outlets, pump-station sumps and weirs. When the level rises above the weir crest, water leaves the network into a river, a fjord or the ground. Monitoring means knowing that it happened, for how long, and how far above the threshold the level went.

The reason to instrument it is rarely curiosity. Discharges have to be reported, the reports have to be defensible, and “the operator thinks it spilled twice that autumn” is not a number. A level sensor and an event log turn an argument into a record.

How it works on Divako

1. Sensor. A pressure, ultrasonic or radar sensor is installed at the weir or in the chamber. Battery-powered LoRaWAN or NB-IoT devices run for years; pump-station controllers can also feed level and status over Modbus or OPC UA.
2. Network. The sensor reports every few minutes. On LoRaWAN it reaches your own gateways; on NB-IoT it rides the operator’s network. Both run side by side in the same account as your metering.
3. Platform. The payload is decoded and the per-device calibration applied – sensor offset, weir crest height – so the stored value is a level you can compare with a threshold. Each device carries four: HH, H, L and LL.
4. Alarms and events. Crossing the overflow threshold opens an event and raises an alarm to email, webhook, MQTT or your incident system. When the level falls back, the event closes with its duration and peak.
5. Integration. Events and level series leave over the same REST and MQTT interfaces as your meter readings, for SCADA, GIS or annual reporting.

What to watch out for

• A threshold without hysteresis produces a hundred events. Water sloshing at the weir crest crosses and re-crosses the line, and each crossing is a new alarm at three in the morning. Set a return threshold slightly below the trigger, and a minimum duration before an event counts.
• Wastewater fouls sensors, and fouling looks like a reading. Rags on an ultrasonic transducer and grease on a pressure cell drift the level quietly instead of failing loudly. Compare neighbouring sites after a rain event, and put sensor cleaning on the same schedule as the rest of the chamber.
• A storm is when every site alarms at once. The hour you most need the system is the hour it sends four hundred messages. Group outlets, route the HH tier differently from the H tier, and decide in advance which alarms a duty engineer must see during heavy rain.
• The event log is only as good as the clock and the calibration. A duration is the difference between two timestamps on a device whose clock drifts, measured against a crest height someone wrote down once. Re-check both after any work in the chamber, before the number ends up in a report.