IoT-Based Flood Detection and Early Warning System with Water Level Sensors

Rivers and reservoirs can rise metres in a few hours during heavy rain, yet most riverside settlements learn about the danger only when water is already at their doorstep. Manual gauge reading is slow and happens at fixed times, so a rapid overnight rise between readings goes unnoticed until streets and homes flood. This project builds an IoT flood detection and early-warning system that watches the water level continuously at three points — a river gauge, a reservoir outflow and a low-lying residential drain — using ESP32 controllers with ultrasonic and float sensors. Each node compares its reading against three staged thresholds (watch 2.50 m, warning 3.00 m, evacuation 3.60 m, all configurable design targets) and pushes the level to a public dashboard every 10 minutes. Crossing a threshold triggers the matching response: a dashboard flag, SMS and app alerts, or the community siren. The system gives residents warning time measured in hours rather than minutes.

Water-level data travels from the river to a resident's phone in five stages. Each sensor node hangs its ultrasonic sensor above the water and measures the distance to the surface, converting it to a water level; a stilling-well pipe around the sensor damps wave ripples so the reading stays steady. The ESP32 on the node takes 5 readings, keeps the median, and checks the result against the float switch and the three staged thresholds. The node then sends the level over Wi-Fi to the cloud dashboard every 10 minutes, or immediately when a threshold is crossed, and the GSM module takes over for SMS alerts where Wi-Fi is unavailable. The dashboard plots the level against the threshold lines, and the alert logic fires the matching response — flag, SMS, or siren — while writing every crossing to the event log.

1. The ultrasonic sensor mounted above the maximum flood line measures the distance to the water surface, and the ESP32 converts the distance into a water level.
2. A float switch at the warning level independently confirms the ultrasonic reading, and the node keeps the median of 5 samples to reject wave noise.
3. The ESP32 compares the level with the three staged thresholds (Watch 2.50 m, Warning 3.00 m, Evacuation 3.60 m) and shortens the sampling interval when the level is rising.
4. Readings are sent over Wi-Fi to the cloud dashboard every 10 minutes, or immediately when a threshold is crossed.
5. The dashboard plots the live level and the 24-hour trend against the threshold lines and records every crossing in the event log.
6. On a Warning crossing, the SIM800L GSM module sends SMS alerts to the registered numbers; on an Evacuation crossing, the community siren also sounds and app push alerts go out.
7. The river and street nodes keep running on solar power with battery backup, so monitoring continues through storm-related power cuts.

Tech stack:

• ESP32 DevKit with Arduino IDE firmware
• HC-SR04 ultrasonic sensor · JSN-SR04T waterproof variant
• Float switch · water-level probe
• SIM800L GSM module for SMS alerts
• 10 W solar panel · 18650 battery backup
• Wi-Fi 802.11 b/g/n · Firebase Realtime Database
• HTML · CSS · JavaScript dashboard
• 12 V community siren · relay driver

The system is built around four capabilities that work together: continuous water-level sensing at three separate points, staged alert thresholds that escalate the response as the water rises, two independent communication paths (Wi-Fi to the dashboard and GSM for SMS), and a public dashboard that shows every station's live level, its 24-hour trend and the full event log. The sensing uses an ultrasonic sensor as the primary measurement with a float switch as an independent cross-check, so a single faulty reading cannot trigger a false alarm. The siren at the low-lying station gives people without phones a warning they can hear, and the solar panel with battery backup keeps the river and street nodes running when mains power is unavailable. Every threshold, sampling interval and alert recipient is configurable in the firmware, so the same build adapts to a campus tank, a village stream or a reservoir outflow without hardware changes.

• [Three-point water-level sensing] Three ESP32 nodes watch a river gauge, a reservoir outflow and a low-lying street drain, each with an ultrasonic sensor and a float switch, so the system sees a flood developing upstream before it reaches homes.
• [Staged alert thresholds] Four levels — Normal, Watch, Warning and Evacuation — at 2.50 m, 3.00 m and 3.60 m (configurable design targets), so the response escalates from a dashboard flag to SMS alerts to the community siren.
• [Dual-sensor cross-check] Every level reading pairs the ultrasonic measurement with a float-switch confirmation, and each sample is the median of 5 readings, which suppresses wave noise and single bad echoes.
• [Siren, SMS and app alerts] Crossing the warning line sends SMS messages through the SIM800L GSM module; crossing the evacuation line also sounds the community siren and triggers app push alerts.
• [Live public dashboard] A web dashboard shows all three stations' current levels, rates of rise, a 24-hour level chart with threshold lines, and an event log of every threshold crossing.
• [Solar power with battery backup] The river and street nodes run from a 10 W solar panel with an 18650 battery (design target), so monitoring continues through power cuts during storms.
• [Rainfall-aware monitoring] A rain gauge at the low-lying station records rainfall, so a sharp rise in rain is visible on the dashboard before the water level itself starts climbing.

• Three assembled sensor nodes (ESP32, ultrasonic sensor, float switch, siren at the street node, solar panel and battery where specified)
• Complete ESP32 firmware with threshold logic, median filtering and GSM fallback
• Live web dashboard with station cards, 24-hour level chart, alert table and event log
• Circuit diagrams, wiring documentation and component list
• Threshold configuration guide (how to set Watch/Warning/Evacuation levels for a local river or tank)
• Project report PDF (background, methodology, architecture, results from the demo build)
• PPT presentation for final review
• Viva Q&A preparation document (ultrasonic ranging, threshold logic, GSM vs Wi-Fi, calibration)