Estimating the Lifetime of Low-cost Sensors for Ambiance Monitoring based on Natural Degradation

Despite the increasing prevalence of prognosis and condition monitoring applications in recent decades, sensor condition monitoring remains an area that has not been extensively explored. The current focus revolves around diagnosing the health of machinery through sensors rather than directly assessing the sensors’ condition. This study introduces a hybrid approach combining physics-informed models and data-driven methods to assess sensor health and estimate remaining useful life. A health index is constructed by integrating fatigue and environmental factors, facilitating sensor health evaluation in the absence of failure data. The efficiency of the proposed approach is also tested for various sensing materials. The experimental findings suggest that low-cost sensors may need recalibration or replacement before reaching the manufacturer’s specified lifetime limit, as indicated in the sensor datasheet. The model’s versatility has been confirmed through tests with diverse sensor components, consistently providing accurate estimates of sensor lifetimes.