Research on Temperature Compensation of TDLAS Gas Detection System Based on WOA-SSA-BP Modeling

Abstract

This paper presents a novel WOA-SSA-BP model to improve the measurement accuracy of tunable diode laser absorption spectroscopy (TDLAS) gas detection systems in variable temperature environments. The model integrates the efficient search capability of the Whale Optimization Algorithm (WOA), the local exploration capability of the Sparrow Search Algorithm (SSA), and the powerful data fitting capability of the Backpropagation Neural Network (BPNN), forming a collaborative optimization algorithm architecture. The model aims to achieve precise correction of ethane (C₂H₆) concentration by accounting for the impacts of three aspects: the variation in gas characteristic spectral line intensity due to temperature fluctuations, the performance instability of infrared light sources, and the electrical property instability of electronic components. The experimental results demonstrate that the WOA-SSA-BP model outperforms the conventional BPNN, WOA-BP, and SSA-BP models. The model exhibits a maximum prediction error of merely 0.29 ppm, alongside an exceptional linear regression coefficient of 0.99998, evidencing its high precision and reliability. Such results suggest that the WOA-SSA-BP model adeptly compensates for the effects of varying temperatures on the TDLAS gas detection system.