Data analysis and uncertainty estimation in supercontinuum laser absorption spectroscopy

Towards a spatially resolved, single-ended TDLAS system for characterizing the distribution of gaseous species

Many applications require diagnostics that can quantify the distribution of chemical gas species and gas temperature along a single line-of-sight, which is challenging in process environments with limited optical access. To this end, we present an approach that combines time-of-flight Light Detection and Ranging (LiDAR) with Tunable Diode Laser Absorption Spectroscopy (TDLAS) to scan individual gas molecular transition lines. This method is applicable in situations where scattering objects are distributed along the beam path, such as solid fuel combustion, or when dealing with multiple gas volumes separated by weakly reflecting windows. The approach is demonstrated through simulation studies and an initial experimental proof of concept for separated gas volumes.

Comparison of Frequency-Domain and Time-Domain Baseline Correction Approaches for Infrared Absorption Spectroscopy of Mixtures Containing Up to 464 Components

Many baseline correction approaches have been developed to address baseline artifacts observed in measured infrared (IR) absorption spectra during post-processing. These approaches offer distinct advantages and disadvantages, and the choice of which one to employ depends on the complexity of baseline artifacts present in a particular application. In this paper, we compare the performance of two baseline correction approaches: a frequency-domain polynomial fitting approach and a time-domain modified free induction decay approach, under various baseline scenarios, spectral resolutions, and noise levels for mixtures containing up to 464 species. Our results showed that the frequency-domain approach outperformed the time-domain approach by a factor of up to 16 when the baseline was represented by a sine wave with fewer than two cycles over the full spectral range. On the other hand, the time-domain approach performed up to 12 times better when the baseline featured two cycles of a sine wave. Additionally, we observed that the time-domain approach exhibited higher sensitivity to spectral resolution and underperformed when the noise level was high. The findings of this study emphasize the importance of numerically testing a few candidate approaches for a given application, taking into consideration baseline characteristics, as well as the spectral resolution and noise constraints of the application.

Performance improvement in a supercontinuum fiber-coupled system for near infrared absorption spectroscopy

Accurate, in-field-compatible, sensing based on near infrared spectroscopy (NIRS) requires development of instrumentation with low noise and long-term stability. Here, we present a fully fiber-optic spectroscopy setup using a supercontinuum source in the long-pulse regime (2 ns) and a balanced detector scheme to demonstrate high-accuracy NIRS-based sensing. The noise sources of the system are studied theoretically and experimentally. The relative intensity noise was reduced from typical values up to 6% to less than 0.1% by deploying a balanced detector and averaging. At well-balanced wavelengths, the system without transmission cells achieved a signal to noise ratio (SNR) above 70 dB, approaching the shot noise limit. With transmission cells and long-term measurements, the overall SNR was 55 dB. Glucose in physiological concentrations was measured as a model system, yielding a root mean square error of 4.8 mM, approaching the needed accuracy for physiological glucose monitoring.

A new approach of H∞ filtering for combustion systems using optical instrumentation

This paper proposes a new filtering scheme applied to a linearized model of a nonlinear representation for combustion systems, whose parameters are obtained by means of optical sensors. To ensure a robust representation regarding the chosen operation point and external disturbances variations, a linear parameter-varying (LPV) state-space representation is proposed in terms of noise disturbances and time-varying parameters affecting the plant (like the instrumentation noise and non-laminar air flow). Concerning the proposed filtering scheme, a new observer structure, which includes the incorporation of the control signal as an additional input of the filter, is proposed to assure improved stability margins and performance given in terms of the H_∞ norm. The filter design method is based on a convex optimization technique and is capable to deal with unstable dynamics. A numerical experiment, whose data were obtained from an actual combustion plant, illustrates the flexibility and advantages of the method when compared with the maximum correntropy criterion based Kalman filter, the full-order filter and the standard Luenberger observer.

Ultrafast-laser-absorption spectroscopy in the mid-infrared for single-shot, calibration-free temperature and species measurements in low- and high-pressure combustion gases

This manuscript presents an ultrafast-laser-absorption-spectroscopy (ULAS) diagnostic capable of providing calibration-free, single-shot measurements of temperature and CO at 5 kHz in combustion gases at low and high pressures. Additionally, this diagnostic was extended to provide 1D, single-shot measurements of temperature and CO in a propellant flame. A detailed description of the spectral-fitting routine, data-processing procedures, and determination of the instrument response function are also presented. The accuracy of the diagnostic was validated at 1000 K and pressures up to 40 bar in a heated-gas cell before being applied to characterize the spatiotemporal evolution of temperature and CO in AP-HTPB and AP-HTPB-aluminum propellant flames at pressures between 1 and 40 bar. The results presented here demonstrate that ULAS in the mid-IR can provide high-fidelity, calibration-free measurements of gas properties with sub-nanosecond time resolution in harsh, high-pressure combustion environments representative of rocket motors.

Cepstral analysis for baseline-insensitive absorption spectroscopy using light sources with pronounced intensity variations

This paper presents a data-processing technique that improves the accuracy and precision of absorption-spectroscopy measurements by isolating the molecular absorbance signal from errors in the baseline light intensity (I_o) using cepstral analysis. Recently, cepstral analysis has been used with traditional absorption spectrometers to create a modified form of the time-domain molecular free-induction decay (m-FID) signal, which can be analyzed independently from I_o. However, independent analysis of the molecular signature is not possible when the baseline intensity and molecular response do not separate well in the time domain, which is typical when using injection-current-tuned lasers [e.g., tunable diode and quantum cascade lasers (QCLs)] and other light sources with pronounced intensity tuning. In contrast, the method presented here is applicable to virtually all light sources since it determines gas properties by least-squares fitting a simulated m-FID signal (comprising an estimated I_o and simulated absorbance spectrum) to the measured m-FID signal in the time domain. This method is insensitive to errors in the estimated I_o, which vary slowly with optical frequency and, therefore, decay rapidly in the time domain. The benefits provided by this method are demonstrated via scanned-wavelength direct-absorption-spectroscopy measurements acquired with a distributed-feedback (DFB) QCL. The wavelength of a DFB QCL was scanned across the CO P(0,20) and P(1,14) absorption transitions at 1 kHz to measure the gas temperature and concentration of CO. Measurements were acquired in a gas cell and in a laminar ethylene-air diffusion flame at 1 atm. The measured spectra were processed using the new m-FID-based method and two traditional methods, which rely on inferring (instead of rejecting) the baseline error within the spectral-fitting routine. The m-FID-based method demonstrated superior accuracy in all cases and a measurement precision that was ≈1.5 to 10 times smaller than that provided using traditional methods.

All Single-Mode-Fiber Supercontinuum Source Setup for Monitoring of Multiple Gases Applications

In this paper, a gas sensing system based on a conventional absorption technique using a single-mode-fiber supercontinuum source (SMF-SC) is presented. The SC source was implemented by channeling pulses from a microchip laser into a one kilometer long single-mode fiber (SMF), obtaining a flat high-spectrum with a bandwidth of up to 350 nm in the region from 1350 to 1700 nm, and high stability in power and wavelength. The supercontinuum radiation was used for simultaneously sensing water vapor and acetylene gas in the regions from 1350 to 1420 nm and 1510 to 1540 nm, respectively. The experimental results show that the absorption peaks of acetylene have a maximum depth of approximately 30 dB and contain about 60 strong lines in the R and P branches, demonstrating a high sensitivity of the sensing setup to acetylene. Finally, to verify the experimental results, the experimental spectra are compared to simulations obtained from the Hitran database. This shows that the implemented system can be used to develop sensors for applications in broadband absorption spectroscopy and as a low-cost absorption spectrophotometer of multiple gases.

Baseline-free quantitative absorption spectroscopy based on cepstral analysis

The accuracy of quantitative absorption spectroscopy depends on correctly distinguishing molecular absorption signatures in a measured transmission spectrum from the varying intensity or 'baseline' of the light source. Baseline correction becomes particularly difficult when the measurement involves complex, broadly absorbing molecules or non-ideal transmission effects such as etalons. We demonstrate a technique that eliminates the need to account for the laser intensity in absorption spectroscopy by converting the measured transmission spectrum of a gas sample to a modified form of the time-domain molecular free induction decay (m-FID) using a cepstral analysis approach developed for audio signal processing. Much of the m-FID signal is temporally separated from and independent of the source intensity, and this portion can be fit directly with a model to determine sample gas properties without correcting for the light source intensity. We validate the new approach in several complex absorption spectroscopy scenarios and discuss its limitations. The technique is applicable to spectra obtained with any absorption spectrometer and provides a fast and accurate approach for analyzing complex spectra.

Highly Sensitive and Precise Analysis of SF6 Decomposition Component CO by Multi-comb Optical-feedback Cavity Enhanced Absorption Spectroscopy with a 2.3 μm Diode Laser

A ppb-level CO sensor based on multi-comb optical-feedback cavity enhanced absorption spectroscopy with a 2.3 μm diode laser was developed for SF₆ decomposition analysis in electric power system. The effective optical path reached to 4.5 km within 35 cm length cavity. Besides, through modulating the cavity length five times automatically, the spectral resolution was improved to 0.0015 cm⁻¹ from 0.0071 cm⁻¹. Targeting the R(6) line of CO first overtone band at 4285.01 cm⁻¹, which is interference free from absorption spectra of SF₆ mixtures (SF₆, SO₂, H₂S, SO₂F₂, HF, CF₄, CO₂, COS, O₂ and H₂O), the minimum detection limit and detection precision under different gas pressures were performed. At optimum integration time of 30 s determined by Allan deviation analysis and gas pressure of 40 torr, the minimum detection limit and detection precision of CO were better than 18 ppb and 150 ppt, respectively.

Efficient Bayesian inference of absorbance spectra from transmitted intensity spectra

High-resolution absorption spectroscopy is a promising method for non-invasive process monitoring, but the computational effort required to evaluate the data can be prohibitive in high-speed, real-time applications. This study presents a fast method to estimate absorbance spectra from transmitted intensity signals. We employ Bayesian statistics to combine a measurement model with prior information about the shape of the baseline intensity and absorbance spectrum. The resulting linear least-squares problem shifts most of the computational effort to a preparation step, thereby facilitating quick processing and low latency for any number of measurements. The method is demonstrated on simulated tunable diode laser absorption spectroscopy data with additive noise and a fluctuating fringe. Results were highly accurate and the method was computationally efficient, having a processing time of only 2 ms per spectrum.

Axisymmetric Linear Hyperspectral Absorption Spectroscopy and Residuum-Based Parameter Selection on a Counter Flow Burner

Chemical species tomography enables non-invasive measurements of temperatures andconcentrations in gas phase processes. In this work, we demonstrate the recently introducedlinear hyperspectral absorption tomography (LHAT) on an axisymmetric counterflow burner usedfor speciation studies of Oxyfuel combustion. As LHAT reconstructs spectrally resolved localabsorption coefficient spectra, the physical plausibility of these reconstructed spectra degradeswith an over-regularization of the tomographic problem. As presented in this work, this behaviorcan be employed in a novel regularization parameter choice method based on the residuals of localspectroscopic fits to the reconstructed spectra. After determining the regularization parameter,the reconstructions of the temperature and water mole fraction profiles of different flames arecompared to numerical simulations, showing a good agreement.

PolySpec: polynomial spectrum models for fast and light-weight spectroscopic evaluation

For the analysis of spectroscopic data, model fitting approaches are commonly applied. The spectrum model applied in these fitting processes significantly influences the performance of the spectroscopic evaluation, which can be critical in real-time process diagnostics and control. In this work a spectrum model is introduced that uses a polynomial description of absorbances, transmittances, or similar in dependence on parameters such as temperature, pressure, and mole fraction. Using this approach, either experimental spectra or spectrum databases can be compressed into a matrix of polynomial coefficients. The evaluation of this model consists of a single matrix multiplication and, with a slight modification, derivatives with regard to specific parameters can be calculated in the same way. Both these points are important to model fitting methods for spectroscopic data, as the simple evaluation method allows for a fast analysis and the direct calculation of derivatives simplifies the application of gradient-based fitting methods. Additionally, the easy parallelizability of the matrix multiplication promotes the application of this method in real-time evaluations on programmable logic devices.