Transient Modeling for Faster Gas Sensor Response: An Efficient Mathematical Approach for MOX Sensors in Ethanol Detection

Abstract

Gas sensor modeling remains a persistent challenge in sensor research, playing a vital role in enhancing performance, calibrating outputs, and developing robust detection algorithms. Among various sensor types, metal oxide gas sensors, often embedded in electronic nose systems, face significant issues such as sensitivity to humidity and temperature fluctuations, complex calibration needs, cross-sensitivity to other gases, and slow response times. Addressing these limitations is crucial for improving sensor accuracy and reliability. This paper introduces a novel modeling approach for the behavior of the metal oxide gas sensor TGS 2620 in the transient domain, with a focus on capturing the effects of temperature and humidity. The proposed model is based on the Gaussian function, chosen for its mathematical simplicity, smoothness, statistical relevance, and broad applicability. By applying this function within a new framework, we aim to improve calibration accuracy and enhance the interpretation of sensor responses under dynamic conditions. Experimental data were obtained through rigorous laboratory testing, using an ENose system equipped with TGS 2620 sensors exposed to varying temperature and humidity levels. The dataset includes 800 samples across gas concentrations ranging from 60 to 400 ppm. Complementary numerical simulations were performed to further analyze sensor dynamics. Our results demonstrate the effectiveness of modeling the sensor's transient response, enabling more accurate estimation of gas concentrations under fluctuating environmental conditions. This work has broader implications in environmental monitoring, industrial safety, and healthcare by enhancing sensor performance for early detection of hazardous gases, ultimately contributing to public safety and quality of life.