IMPLEMENTING INTERNET OF THINGS FOR REAL TIME CORROSION MONITORING AND RESPONSE

Traditional corrosion monitoring techniques rely on manual inspections and periodic assessments, often time-consuming, labor-intensive, and costly. These methods are limited in detecting early-stage corrosion, resulting in delayed maintenance interventions and an increased risk of catastrophic failures. Additionally, manual inspections are inherently subjective and may overlook critical areas, particularly in complex or hazardous environments. The limitations of these conventional approaches underscore the urgent need for more effective corrosion monitoring solutions. This project presents an IoT-based corrosion monitoring system, designed to assess and predict corrosion risk in real time through the integration of custom-built environmental sensors and a machine learning classifier. The system utilizes a combination of temperature and humidity sensors, developed specifically for this project, alongside electrical resistance measurement to evaluate corrosion potential. Data is processed on a microcontroller through a trained decision tree classifier, optimized with TinyML and TensorFlow Lite, to classify conditions as either favorable or unfavorable for corrosion, represented by binary outputs. The project’s results highlight the relationship between environmental factors and corrosion likelihood, presenting valuable trends in temperature, humidity, and resistance data collected over a period of time. Through this system, the potential of IoT-enabled predictive maintenance is demonstrated, showing both cost-effectiveness and a scalable application across industries where corrosion monitoring is essential for safety and operational efficiency.