An AI-Based Multi-Depth Water Quality Monitoring and Early Warning System for Storage Tanks

Water stored in tanks may change in quality because of sediment accumulation, mixing, temperature variations, and other local disturbances. However, many existing monitoring approaches rely on periodic sampling or on measurements taken at a single location. This paper proposes AquaWatch, an IoT- and AI-based system that monitors water quality at three depths within a tank. The proposed AquaWatch system uses three sensing stations positioned at different depths of the tank: 0.40 m, 1.30 m, and 2.10 m. Each station is designed to monitor important water-quality parameters, including pH, turbidity, electrical conductivity (EC), estimated total dissolved solids (TDS), temperature, dissolved oxygen (DO), and oxidation-reduction potential (ORP). The collected readings are processed by an ESP32 and transmitted to a gateway through LoRa communication. From there, the data are forwarded to a cloud database for further analysis. To identify unusual changes in water quality, the cloud platform uses Isolation Forest, Z-score, exponentially weighted moving average (EWMA), and cumulative sum (CUSUM) techniques. These methods work together with cross-depth analysis to determine whether a change affects the entire tank or occurs only at a particular depth. The monitoring information is then displayed on a dashboard, while mobile notifications inform users. The system is also designed to support solar-assisted operation in remote locations. This paper describes the proposed system design, methodology, implementation plan, and evaluation approach. Experimental measurements are not included at this stage and will be considered during prototype testing. The generated alerts indicate abnormal patterns only and require further inspection and confirmatory laboratory testing.