A near infrared H2S leakage detection system based on WDO-ELM using a digital lock-in amplifier combined with discrete wavelet transform filter

A gas leakage detection system was developed based on tunable diode laser absorption spectroscopy (TDLAS) technique with the combination frequency absorption band of hydrogen sulfide (H2S) gas molecules at 1578 nm for telemetry. For further telemetry distance and signal-to-noise ratio (SNR), a Fresnel lens with a diameter of 250 mm is utilized to focus the laser beam that diverges gradually with the increase of telemetry distance. A digital lock-in amplifier (DLIA) combined with discrete wavelet transform (DWT) filter was devised to process faint photoelectric signal received by InGaAs photodiode. Meanwhile, a global optimization algorithm, combined wind driven optimization (WDO) with extreme learning machine (ELM), named WDO-ELM algorithm, is applied for H2S concentration inversion to make up the limitation of ELM algorithm and seek for the optimal weight and threshold in a more extensive range. The reliability and sensitivity of the H2S leakage detection system have been verified by calibration and stability experiments. The limit of detection (LoD) reaches a minimum of 0.983 parts-per-million·meter (ppm·m) with a 96 s integration time by Allan-Werle deviation analysis. This self-developed H2S leakage detection system with high integration and low cost is more convenient for commercial application in tank farm to realize remote detection of gas leakage.