Ultrasonic Sensor-Based Flood Prediction and Early Warning System

Floods are undoubtedly some of the most devastating and common natural hazards on Earth, resulting in numerous casualties and displacements, not to mention infrastructural destruction. Flash floods, in particular, require an efficient and resilient early warning system capable of operating even in less developed areas. In this study, an affordable embedded flood forecasting and warning system is designed, built, and simulated using an HC-SR04 ultrasonic sensor and the Arduino Uno development board. The system works by constantly gauging the instantaneous water height in a water body, such as a river channel or reservoir. The raw data from the HC-SR04 sensor undergoes several software stages including filtering out noise, comparing with predetermined threshold levels, and calculating the rate of rise. There are three alert modes, namely NORMAL, WARNING, and DANGER, to cater to varying levels of flood threat. The alert modes are conveyed visually via LED indicators and audibly via a piezoelectric buzzer. Furthermore, besides simple threshold detection, the system features a prediction component, which relies on the rate of rise of the water level between two consecutive measurement phases. Using such rate-of-rise detection helps to provide early flood warning alerts to the user even before the level of water reaches any predetermined static danger level. In order to simulate the system, the complete solution was designed within the online embedded simulation environment provided by Wokwi, in which the different flood cases were replicated by changing the distance between the sensors and the surface of water level. The experimental simulation results indicate that the solution responds quickly (below 100 ms per each measurement), achieves a high degree of detection accuracy, and works reliably when simulating a series of flood situations. The proposed solution proves highly energy efficient, needs minimal maintenance and, finally, can be easily built using off-the-shelf hardware at considerably low prices. Therefore, the solution can be successfully deployed at scale across flood prone regions in many developing countries. The future enhancements of the project include incorporation of IoT data transmission, GSM SMS alerting capability, solar energy harvesting, and AI/ML based flood prediction among other improvements.