Method for accurate gas concentration retrieval in an in situ environment with particulate interference

Area of Science:

• Spectroscopy and Analytical Chemistry
• Environmental Monitoring and Sensing
• Signal Processing and Data Analysis

Background:

• Tunable diode laser absorption spectroscopy (TDLAS) is a sensitive technique for trace gas detection.
• Particulate matter in complex environments attenuates spectral signals, reduces signal-to-noise ratio (SNR), and compromises accuracy.
• Existing denoising methods struggle with combined particle extinction and noise, especially under high particulate concentrations.

Purpose of the Study:

• To develop a novel spectral processing method for TDLAS systems affected by particulate matter.
• To improve the accuracy and reliability of gas concentration measurements in challenging in situ conditions.
• To address limitations of current denoising techniques in handling particle extinction and baseline fitting.

Main Methods:

• Proposed a transmitted spectral processing method using variational Bayesian adaptive extended Kalman filtering (VB-AEKF).
• Integrated Mie scattering theory with the Beer-Lambert law to formulate a state vector including concentration and extinction coefficient.
• Employed a dual-layer iterative structure for state estimation under unknown baseline conditions.

Main Results:

• VB-AEKF demonstrated enhanced denoising for simulated signals with concentration retrieval errors less than 1%.
• Achieved superior accuracy and stability in extinction coefficient calculation compared to robust linear regression.
• In situ methane measurements showed a mean relative error of 0.46% (vs. 4.69% for traditional methods) and a 61.7% reduction in concentration standard deviation.

Conclusions:

• The VB-AEKF method effectively overcomes limitations posed by particulate matter in TDLAS measurements.
• The algorithm significantly improves baseline estimation, signal reconstruction, and gas concentration retrieval accuracy.
• VB-AEKF offers a robust solution for reliable trace gas detection in environments with severe particle interference.