[Objective] Precise temperature monitoring within the optimal reaction window(850～1 100 ℃) is critical for the efficiency and low-emission operation of selective non-catalytic reduction(SNCR) systems in waste incineration furnaces. Conventional contact-based measurement methods often suffer from time lag

mechanical fragility

and limited adaptability to harsh

high-temperature environments. This study aims to develop a robust acoustic thermometry technique to provide highly reliable time-delay estimation

thereby enabling accurate and real-time temperature monitoring for SNCR system control. [Methods] To address the challenge of strong background noise in industrial settings

an enhanced time-delay estimation method named VMD-WT-GCC is proposed. The core innovation lies in the synergistic combination of variational mode decomposition(VMD) and wavelet threshold(WT) denoising as a pre-processing step for the generalized cross-correlation(GCC) algorithm. The VMD adaptively decomposes the acquired acoustic signal into intrinsic mode functions

effectively separating the useful signal components from noise. Subsequently

the WT technique is applied for further refinement and thresholding of these components. This two-stage denoising strategy significantly enhances the signal quality before performing GCC for precise time-of-flight calculation

which is then converted to temperature via the known sound speed-temperature relationship. The method's performance was rigorously evaluated through simulation

laboratory experiment

and industrial field trials. [Results] Simulation analysis

utilizing actual on-site noise data

was conducted across a challenging signal-to-noise ratio(SNR) ranging from 5 dB to –10 dB. The proposed VMD-WT denoising method demonstrated superior performance compared to the conventional wavelet threshold method alone

showing marked improvements in key metrics: simulated SNR gain

the structural similarity index(SSIM)

and the correlation coefficient(CC). Laboratory experiments were performed in a controlled acoustic environment with a stable temperature of(–3.2 ± 0.5) ℃. Results confirmed the method's high accuracy

with the average absolute error of temperature measurement remaining below 0.20 ℃ across various propagation paths ranging from 1 m to 5 m. A field trial was implemented on a 750 t/day waste incinerator

specifically at the SNCR ammonia injection plane located 30.1 m above the furnace grate. The single-path acoustic temperature measurements showed consistent stability

with the average deviation from benchmark reference data maintained within 50 ℃

validating the method's robustness under real industrial operating conditions. [Conclusion] The developed VMD-WT-GCC method exhibits exceptional accuracy and stability for time-delay estimation in strong noise backgrounds

overcoming a major limitation of acoustic thermometry in practical applications. It provides a viable and effective technical solution for high-precision

non-contact temperature monitoring in complex industrial environments like waste incineration furnaces. The successful validation from simulation to field deployment underscores its strong engineering applicability and promising potential for optimizing SNCR system performance

contributing to improved combustion control and emission reduction.