Research on Malodor Component Identification Based on Sensor Array.

Area of Science:

• Environmental Science
• Sensor Technology
• Data Analysis

Background:

• Malodor pollution is a growing concern impacting public health and environmental quality.
• Existing online malodor detection systems face challenges with sensor drift, affecting accuracy and long-term stability.
• Accurate prediction of malodor gas composition is crucial for effective environmental monitoring and safety management.

Purpose of the Study:

• To develop an advanced electronic nose (e-nose) detection framework for accurate malodor gas composition prediction.
• To address sensor drift and enhance the long-term stability of malodor detection systems.
• To improve the precision and reliability of detecting malodorous compounds in real-time.

Main Methods:

• Utilized a multi-channel sensor array for gas sampling to establish sensor response patterns and gas concentration distributions.
• Proposed an innovative Encoder-Decoder architecture, IED-CNN-LSTM, incorporating external compensation mechanisms.
• Applied time series data analysis for enhanced feature extraction from sensor data.

Main Results:

• The IED-CNN-LSTM model demonstrated significantly superior prediction accuracy compared to conventional methods.
• The proposed framework achieved enhanced long-term stability, overcoming sensor drift issues.
• Experimental results confirmed more precise and reliable detection of malodorous compounds.

Conclusions:

• The developed IED-CNN-LSTM framework offers an effective solution for real-time environmental monitoring of malodors.
• The study provides substantial improvements in performance metrics and practical implementation for industrial and regulatory applications.
• This research enhances the capability for accurate and stable malodor detection, contributing to better environmental management.