Flame, Combustion and Explosion Thermometry
Journal Archive
doi: 10.1595/205651323X16643556587827
Flame, Combustion and Explosion Thermometry
From optimising waste incineration to internal combustion engines and defence applications
Article Synopsis
Flame is a natural phenomenon and is a basic element of any combustion process. The majority of flames consist of a gas; there is, however, a small amount of ionisation occurring in the flame. Despite the increased focus on combustion-free energy production such as wind, air and water power, and the refocus on nuclear energy now considered to be carbon-free, nonetheless combustion will remain, for the next few decades, the major energy and heat production route worldwide. Apart from carbon dioxide, which is commonly considered to be the major pollutant, there are other gases like nitric oxide and nitrogen dioxide which, although found in significantly lower amounts in the exhaust gases from combustion units, still present a large environmental impact and are a concern. There are however well-established technologies for removing combustion products from the exhaust gas, and the combustion process can in general be made CO2 and environmentally neutral. Combustion optimisation is a route for further reduction of undesirable byproducts, fuel consumption minimisation and finally an overall energy and heat production enhancement. The key parameter in any combustion process is reliable flame and (post-) combustion gas temperature measurement and control. Various combustion environments such as waste incineration, internal combustion engines or solids explosions cause the appearance of various optical emission features in different spectral ranges not accessible to the human eye. A combination of modern and newly developed fast spectral optical techniques with extensive theoretical developments in spectral and heat radiative transfer modelling allows us to obtain detailed snapshots of what is happening in the combustion process. That also gives a possibility to establish a direct link to the industrial process control and pollutant emission reduction. In this article some examples of in situ flame and gas temperature measurements in various combustion environments and advanced spectral modelling are given and perspectives for further commercial instrumentation developments are discussed.
1. Introduction
Temperature is both a thermodynamic property and a fundamental quantity. Accurate gas temperature measurements are essential in fundamental and applied science and technology. There is a vast amount of publications, handbooks and monographs with a focus on temperature measurements from fundamentals to applications, for example, (1–5). The most common way to measure temperature is to use a device commonly called a thermometer. One of the most well-known thermometers is a thermocouple which is used for contact measurements (1).
Radiation thermometry is one of the other temperature measurement methods, which is used for non-contact thermometry. It utilises thermal emission from an object which can be a surface or a gas volume. Basics of radiation thermometry are given by Nicholas and White (1), while many practical thermometry aspects are discussed by Michalski et al. and Childs (4, 5). Radiation thermometry is widely used for surface temperature measurements in various industrial processes. Zhang et al. (3) give an excellent overview of thermometry used for various very common industrial applications such as steel, glass, plastics and semiconductor production. In the majority of industrial applications of radiation thermometry, the temperature of a (solid) object is much higher than the temperature of the surrounding gas (a gas between the object and a measurement instrument). Therefore the thermal radiation emitted and registered by the measurement instrument can be described with the use of Planck’s law. The instruments which are used for measurements of thermal emission described by Planck’s law are commonly called pyrometers. There are various types of pyrometers based on their operation principle: for example, spectral band, total radiation, ratio, multi-waveband and thermal imagers (5). Childs (5) also gives a brief description of spectroscopy-based methods for temperature measurements. Most pyrometers essentially select one (or several) wavelength(s) or spectral interval(s) where contribution from possible thermal surrounding gas emission (like CO2 and water) will be negligible. If the temperature of the surrounding gas is not negligible compared to the object’s temperature and the distance to the object is substantial (a few metres), then the classical pyrometer (based on Planck’s law) is likely to give wrong measurement results.
For many high-temperature applications such as combustion and nitrogen oxides (NOx) reduction, in situ gas temperature measurements are very important. In both cases the gas is much hotter than the wall temperature of the combustion unit. Planck-based pyrometers can still be useful, for example, (soot) particles and (refractory) wall temperature measurements at relatively short distances from the instrument to the measurement object.
A hot gas can simultaneously absorb and emit thermal radiation. Radiation penetration through a hot medium is governed by a radiative heat transfer equation. In a simple case of a temperature-homogeneous gas column of length l the incoming radiation to the detector is described by Equation (i) (6):
where 
At long gas columns or high temperatures and CO2:H2O concentrations the exponential factor in Equation (i) becomes large and therefore e–kλl«1. At those conditions 
Despite the above-mentioned limitations, the pyrometers are frequently used on an industrial scale for measurements in combustion. There are even permanent installations with feedback to the process control. The temperature obtained from those systems is the temperature associated with the particles such as soot and ash. An example of the use of pyrometers for two-dimensional (2D) temperature measurements can be found in CombTec GmbH, Zittau, Germany. That approach is based on ‘passive’ diagnostics when there are only several receivers measuring emission from the gas along limited number and predefined lines of sight. Several access ports are required for the system and it makes the installation and running costs relatively expensive. Moreover, pyrometers measure near infrared (NIR) emission in the 0.7–1.15 μm spectral range that can only be used for soot or fine ash temperature calculations. However, as it is well known, equality of the particle and gas temperatures is not always fulfilled.
Moving towards the use of green and clean fuels in combustion means that combustion and post-combustion gases become very clean, with a minor particle (soot) fraction. However, in applications such as the waste incinerator sector, where any kind of waste is combusted, the particle load can be quite high. Here obviously the pyrometers can be used for particle temperature measurements.
Recent progress in industrial applications of laser-based diagnostics led to the development of various laser-based systems (9). Tunable diode laser absorption spectroscopy (TDLAS) is one of the most known techniques used in many industrial environments. In the TDLAS the laser scans one or few preselected isolated absorption lines of for example, CO2, water or oxygen molecules and then the temperature is deduced from a shape fit of the lines (the widths of the lines are temperature dependent). A TDLAS system requires a pair (emitter and receiver) installation, well-aligned along a LOS. Making such installation (with an access possibility inside of the combustion unit) at an industrial site might be difficult and therefore costly. Another disadvantage is in the steering of the laser beam when there are changes in the gas temperature.
An example of a larger scale TDLAS installation is the ZoloBOSS® system (10) where several TDLAS (emitter and receiver) pairs formed a grid used for tomographic temperature reconstruction. The ZoloBOSS® tomography algorithm is based on finite-domain direct inversion method (11), which was developed for use with sparse data sets, whereas the more conventional approach of CombTec GmbH is based on a soft computing approach (12) which has become popular in the last decade. Both algorithms assume smooth distributions and use a priori information in a form of initial guess. Their main disadvantage is due to limited information available from the measurement grid.
In opposition to laser-based approaches, various methods of broadband emission spectroscopy from ultraviolet (UV) to visible and infrared (IR) spectral ranges can be used for gas and particle temperature measurements. The main advantages of using the broadband spectroscopy tools are: (a) a possibility to cover a broad spectral range with an inexpensive (compared to laser systems) spectrometer; (b) there is no need for high-spectral resolution measurements (to resolve single molecular emission lines in the spectra); and (c) for practical applications, there is a need for only one access port.
At elevated temperatures (above 1100 K) the tail of the black-body emission continuum (described by Planck’s function) extends towards visible-UV spectral range. Therefore a very small for example, visible-optimised low-resolution spectrometer with a one-dimensional (1D) array detector can be used for particles (soot) temperature measurements in for example, 300–800 nm in the same way as an IR pyrometer at 3.9 μm. Moreover, because the tail is highly temperature dependent the cut-off of the emission can be used for very fast temperature measurements. An advantage of using the spectrometer is that the spectrum is recorded at the same time in the whole spectral range of the spectrometer and it is easy to distinguish between a continuum radiation and any molecular, atomic or radical emission or absorption those can either appear on the top of the continuum or under that, respectively.
Depending on the nature of the process several molecular emission features of molecules (for example, CO2, water in IR) or radicals (for example, CH, OH in UV-visible) can be used for gas temperature measurements. Because the spectral shapes are temperature dependent, the temperature can be deduced in the same way as for laser-based systems, namely as that which enables the best fit to the observed emission band. Based on the fit, the gas temperature is calculated.
While Planck-based pyrometers require periodic calibration using a traceable calibrated thermal light source such as a black-body, spectroscopy-based thermometers use a set of molecular parameters that are specific for a particular molecule used for gas temperature measurements. The set is used to calculate a modelled spectrum which is then used to fit the measured spectrum. This approach is virtually calibration-free because it relies only on the molecular parameters and spectra modelling. However, a quality of the molecular parameters is essential and there is no explicit way to validate the parameters, especially at elevated temperatures, except through a comparison of obtained highly-accurate experimental spectral data with spectra modelling results.
Molecular parameters for a vast amount of molecules can be obtained from spectroscopic databases such as HITRAN (13) and its high-temperature implementation HITEMP (14). The molecular parameters can be determined from complex, frequently laser-based, measurements or (in some cases) can be calculated with the use of advanced ab initio quantum mechanical methods. The latter, however, are quite demanding in computing resources and computation time. Progress in computational capabilities and methods nowadays allow for accurate molecular electronic spectra calculations which were impossible 10–15 years ago.
In our recent paper (15) we have validated two Fourier transform IR (FTIR)-based emission spectroscopy systems (single LOS and hyperspectral imaging) with a portable gas flame traceable to the International Temperature Scale of 1990 (ITS-90). It was shown that in the central post-flame region dominated by CO2 and water IR band emissions, the agreement between the Rayleigh and FTIR temperatures is within the combined measurement uncertainties (from the systems used and the HITEMP-based spectra modelling) and amounts of 1% (coverage factor k = 1 i.e. 67% coverage probability) of temperature. The portable flame itself (calibrated via Rayleigh scattering thermometry technique traceable to ITS-90) has an uncertainty of 0.5% (k = 1) of the temperature. This is a quite remarkable result that shows current achievements in the instrumentation and high-temperature spectral database developments.
Recent developments in artificial intelligence (AI) technologies allow us to develop efficient computational tools those can be used in deep analysis and predictions in various laboratory and industrial processes with dynamically changing and complex operation conditions. With the use of AI approaches such as machine learning several process scenarios can be modelled and an algorithm for a quick data analysis from unknown process conditions can be developed through the training data sets. In our recent article (16) we have developed an inverse radiation model for temperature and species concentrations determination from CO2 and water transmissivity measurements. More recently we applied a neutral network approach to retrieve temperature and species concentrations from IR emission measurements for combustion gases (17). The approach has been demonstrated on the portable flame standard mentioned above (15) and we have shown that uncertainties for gas temperature retrievals in the centre of the post-flame are below 0.5% for most practical operation conditions such as stoichiometric and lean combustion. This means that the AI approach used does not introduce an additional significant error in the temperature measurements at around the 2250 K level. A combination of a carefully-designed single LOS instrumentation and an AI data analysis approach opens new possibilities for very detailed information retrievals from emission measurements such as temperature and species concentrations.
The single LOS approach described above can further be developed towards tomographic temperature and species reconstructions. As is well known, tomographic reconstruction techniques require as many LOS as possible and the field of view of each LOS should be as narrow as possible. This requirement is often satisfied in medical tomographic systems where number of LOS reaches 100–1000 and reconstruction can be considered to take place on a semi-continuum LOS scale. However in tomography of gases, flames or plasmas this number is usually limited by significantly lower values, typically 5–12, and therefore reconstruction takes place on a discrete scale. At each LOS, temperature and species concentration profiles can be retrieved with use of AI tools (for example, machine learning). Then 2D temperature and species concentration tomographic images can be obtained by gluing together all retrieved profiles. Multiple LOS measurements can be performed using a sweeping approach when emission is collected under different LOSs accessible from only one access port by the same instrument.
There is no ‘one-size-fits-all’ solution for how the temperature measurements can best be done in practice for a particular system. It depends on the particular process and practical access possibilities. The following section gives examples of various experimental set ups used for measurements from UV to visible and IR spectral ranges.
2. Experimental
Depending on the process, practical implementations of spectroscopy-based measurement systems are different. There are two generic approaches for in situ measurements: active and passive.
In the active approach an external light source is used to probe the medium (flame or hot gas). A light attenuation is then measured. The benefit of the active approach is a possibility of measurements of the species in their ground states that can both be used for (effective) temperature and concentration measurements. The drawback of the active approach is necessarily of the light source and possible light beam steering if temperature of the medium (gas or flame) is changing over time. Moreover two well-aligned access ports are required that can be a costly solution for many industrial installations.
In the passive approach there is no external light source and the emission from the (hot) medium serves as the light source itself. The passive approach requires only one access port and gives possibilities for temperature profiles retrievals from, for example, a single LOS measurement. The drawback of the passive approach is that it requires a relative or absolute system calibration with use of a reference light source. The reference light source can be, for example, a black-body one with known reference temperature. The black-body source produces an IR emission which obeys the known thermodynamic Planck law.
Because both approaches utilise light they are prone to possible optics contamination that can affect the systems’ short or long-term performances. Optics contamination of the passive system can cause a change in the system calibration that will influence the quality of the measurement data (and therefore temperature), while for the active system the contamination can lead to a bias in the measurement data that can be accounted for. In the case of significant contamination both systems will need optics cleaning and recalibration. Therefore, in any practical installations, special attention to keep the optics clean is strictly necessarily.
A passive spectroscopic system design includes three basic elements: (a) focusing optics (lens or mirror); (b) optical fibre; and (c) spectrometer. The focusing optic is typically mounted on the access port with use of a support or a special (cooled) probe. The optical fibre is used to transmit light from the optics to the focusing optic. The optical fibre can be omitted when the spectrometer is directly coupled to the focusing optic. The benefit of using a fibre is that the spectrometer can be placed at a distance from the access port.
The choice of all three elements depends on the light which is going to be detected. UV quartz/sapphire-based materials and fibres can be used for measurements in UV-visible spectral range while ‘CaF2’-based or other IR materials and for example, chalcogenide (CIR) or hollow core fibres can be used for measurements in IR spectral range. The spectrometer can be as a grating-based one or FTIR one. Traditionally for the measurements in UV-visible grating spectrometers are used while for the IR spectral range it is common to use FTIR-based ones. Surprisingly, a spectral resolution of the spectrometer is not critical rather a wider spectral coverage (for grating units) and fast measurement time and fast scanning (for FTIR units) are essential.
In the following section three examples of passive approach for gas and particle temperature measurements are given.
3. Results and Discussion
3.1 Particle and Gas Temperature in Explosion Events
Gas or solids explosions are of interest for various defence applications because a high amount of energy is released in a very short time slot. Gas and particle temperatures are not necessarily equal in the explosion’s active phase. This is why spectroscopic measurements in a wide spectral range are preferential over other single (or few) wavelength based approaches. Figure 1 shows two consecutive visible emission spectra in 530–710 nm in large charge explosion with the use of pyrotechnical materials.
Fig. 1.
Two flame visible emission spectra obtained in a large charge pyrotechnic explosion (grey, blue) and respective grey-body fit (olive (1880 K), red (1950 K)). Spectra show various free metal emission lines such as potassium and sodium. Some of those show self-inversion and molecular species such as CaOH. Modelled CaOH (A–X) emission spectrum at 3000 K is shown by the wine coloured line
Each spectrum was recorded with about 4 ms acquisition time. The full explosion event takes about 30 ms. Because the explosion event generates a very luminous flame in the visible spectral range it is possible to make further optimisation of the measurement system in terms of compactness and speed (spectrometer) to perform measurements at a few microseconds acquisition time without compromising signal-to-noise ratio in the measured spectra.
The spectra show various free metals emission lines (for example, potassium, sodium, lithium) and CaOH (A–X) and CaOH (B–X) band systems on the top of grey-body continuum caused by emission from hot (soot) particles. Appearance of the emission lines and bands on the top of grey-body continuum indicates that gas is hotter than the particles. Spectral regions free from the emission features such as around 540 nm and 680 nm can be used for a grey-body fit that will give us particle temperature. A fit of CaOH (A–X, Δv = −1) group around 640 nm shape of which is temperature sensitive with use of recently obtained CaOH (A–X) spectroscopic data (18) shows that the gas temperature is 3000 K. One can see that in the 580–660 nm range many emission structures appear that can make detailed spectra analysis complicated. However the CaOH (B–X) group located at about 555 nm is free from any interference. The CaOH (B–X) spectroscopic data are presently under development. CaOH emission appears in the active explosion phase when the temperature reaches its highest values.
The double sodium resonance emission lines at about 589.6 nm are frequently observed with self-inversion that indicates temperature non-uniformity in the gas slab; a colder gas layer causes self-absorption of the sodium emission from the hot gas layer. This is a known phenomenon frequently observed in sodium high-pressure lamps (19). Modelling of the sodium resonance lines with self-inversion gives highest and lowest temperatures in the dense non-homogeneous gas.
3.2 In situ Gas Temperature Measurements in Very Hot Industrial Environment
Temperature measurements in very hot industrial environments (>1900 K) are always a challenge. Therefore in situ non-contact and fast measurements are preferable. The typical timescale when the measurements have to be done, however, differs from that of the explosion described in Section 3.1. Here the acquisition time for one data set (spectrum) of about 1 s can be acceptable. Figure 2 shows a part of UV emission spectrum in the OH (A–X) band vicinity measured in a hot stone melter (stone wool production). The stone melting environment is a combination of various (heterophase) processes in the hot melt (like lava) and in the very hot gas. To melt stones a very high temperature and consequently a high energy input are required. Therefore reliable in situ gas temperature measurements are important in the production process.
Fig. 2.
A UV emission spectrum at 1 s acquisition time (brown) in vicinity of OH (A–X) emission group measured in a stone melter and its fit with black-body (2598K) (magenta) and grey-body (2568 K, ems = 0.91) (olive). Calculated from the measured spectrum OH (A–X) emissivity (blue) and calculated OH (8 vol%) emissivity over 2 m at constant temperature profile (gas temperature 2596 K) (red). Copper lines (324–328 nm) are observed in emission with self-inversion
If the gas temperature becomes quite high, the black-body continuum and its maximum which is described by Planck’s law will tend to extend to shorter wavelengths and can reach the UV spectral range (8). Although the black-body emission in the UV range (200–400 nm) is not as strong as it is in the visible one, the UV emission measurements below 400 nm are possible if the spectroscopic system is tuned to target a specific spectral range of interest such as OH radical emission around 310 nm.
OH (A–X) emission (or absorption) at 310 nm is widely used in combustion diagnostics for dynamical gas temperature in situ measurements. Various classical absorption and laser-based methods (such as for example, laser induced fluorescence) have been developed and used for OH (A–X) measurements. A fit of the OH (A–X) spectrum gives actually a rotational temperature. Once an OH radical is formed, its intermediate rotational temperature may differ from the overall (thermodynamic) gas temperature. However at pressures around 1 bar or even higher the rotational temperature will be equal to the gas temperature because of the very short relaxation time of the OH radical in collisions with surrounding gas molecules. Because of the simple structure of OH its spectral modelling can be done very accurately.
The OH emission is relatively weak compared to the grey-body continuum and reaches the black-body continuum at the OH band head at 307 nm, Figure 2. Moreover, as we have seen, in a long measurement series the OH emission can disappear, appear again or becomes as OH absorption. Those three situations essentially correspond to cases when the gas and particle temperatures are equal, gas temperature is higher than the particle temperature and local gas temperature is lower than the particle temperature, respectively. As one can see from Figure 2 the difference in black-body (2598 K) and grey-body (2568 K) fits is very small (30 K) which means the gas and particles have about the same temperature. In addition to OH emission two copper-lines with self-inversion are observed. The self-inversion indicates that the gas has a non-uniform temperature profile along the line-of-sight that is caused by copper-lines inversion (see also Section 3.1).
The OH emissivity spectrum calculated from the measured emission (brown) is shown in Figure 2 by the blue line together with OH (A–X) emissivity fit at OH (8 vol%) and a constant gas temperature profile (2596 K) over 2 m (melter cross-section). One can see a very good OH (A–X) emissivity fit (red) of the measured OH emissivity (blue). The tail of the OH (A–X) band in 314–326 nm is temperature dependent (higher temperatures cause tail extension) and serves as a good temperature indicator. Calculated gas temperature from the OH (A–X) band fit is very close to the black-body temperature fit and proximity of all three temperatures (black-body, grey-body and gas temperature from OH (A–X) fit) show a high temperature uniformity of the heterophase medium in the stone melter which indicates high process efficiency.
In regard to experimental set up (with a high-resolution spectrometer) used it should be noted that a simplification of the system with use of a low-resolution (compact) spectrometer is possible. This is because the OH (A–X) emission is quite broad (304–328 nm) at those temperatures and although the low-resolution spectrometer will not allow us to resolve fine structures in the OH (A–X) bands the overall OH band shape will show the same temperature dependence as the one for the high-resolution.
3.3 Gas Temperature Retrievals for Selective Non Catalytic Reduction
Selective non catalytic reduction (SNCR) is an essential process in NOx removal from the flue gas for small and medium scale combustion units (such as domestic heating plants, waste incinerators) used for electricity and heat production. The SNCR is performed with an injection of ammonia or urea solution in the hot flue gas. The process has best efficiency at gas temperature around 1300 K with about ±100 K temperature window depending on the composition of flue gas. If the temperature falls outside of the temperature window there is a risk that either not all NOx will be removed or an ammonia slip into the atmosphere can occur. Therefore it is important to monitor and control the gas temperature at the point where the ammonia or urea injection takes place.
As we have shown in (15), single LOS IR emission measurements can be used for gas temperature profiles retrievals in post-flame region. Similar spectroscopic set up have been adopted for 2D maps temperature measurements on a waste incinerator plant. The system utilises a fast scanning FTIR spectrometer which runs at low spectral resolution (8 cm–1) and a sweeping short probe arrangement that allows multiple line-of-sight measurements in the horizontal plane. The probe was set into the hot gas just few centimetres deep from the inner wall surface.
Figure 3 shows an example of two FTIR emission spectra measured at two times (two hours’ separation) along the 0 degree line of sight (i.e. perpendicular the boiler wall). The spectra have been measured at low spectral resolution and 1 s acquisition time. The distance to the opposite wall was 10 m.
Fig. 3.
Two representative emission spectra measured at 8 cm–1 spectral resolution (blue, olive) and zero degree line-of-sight at different times. Black-body and grey-body fits are shown by red/black and orange/wine lines
The spectra are dominated by water emission bands with a minor CO2 bands contribution on the top of a continuum (caused by hot particles and opposite wall emissions). The grey-body fit in the CO2:H2O-free spectral ranges (2560 cm–1, 4500 cm–1 and 6000 cm–1) gives an effective grey-body temperature and the grey-body continuum can be subtracted from the measured spectra to get grey-body-free emission spectrum, Figure 4. It should be noted that the overall black-body fit for both spectra is unsuccessful that indicates about a non-uniform gas temperature profile (for uniform temperature profile and such long distances and given water concentration, the water band tops will approach the black-body continuum as one can see from Equation (i)).
Fig. 4.
Spectrum 2 from Figure 3 (after subtraction of grey-body fit) (blue) and its fit with use temperature profile (from Figure 5) and emission from opposite wall (930 K/ɛ = 0.84) using only water emission bands (red). The water emission spectrum using a constant temperature (1300 K) is in olive. Spectral regions with CO2 contribution are marked by arrows
The emission spectrum (corrected for the grey-body emission) in Figure 4 can be fitted with the use of a (constant over distance) water concentration and a gas temperature profile calculated emission spectrum. The best fit of the Spectrum 2 (blue) in Figure 4 can be obtained at water (35 vol%), opposite wall temperature (930 K, emissivity 0.84) and temperature profile given in Figure 5. The water concentration and opposite wall temperature and emissivity are obtained or estimated from other measurements and calculations. The calculated water emission spectrum (red) is shown in Figure 4. The calculations have been performed using the HITEMP-2010 database (14). One can see that there is a very good agreement between the measurements and calculations. For a comparison water emission spectrum at a constant temperature profile (1300 K) is also presented (olive). While in the wings of the water 1500 cm–1 and 3500 cm–1 bands the agreement between both calculations and the measured spectrum is very good, a significant difference is observed in the bands tops in 1200–1800 cm–1 and 3300–4000 cm–1. This is caused by water self-absorption in the colder gas layers similar to the metal lines self-inversion described in the previous sections. This self-absorption will only appear for resonance bands (or lines), i.e. for transitions between excited and ground states of molecules or atoms those can appear in both directions. The retrieved gas temperature profile agrees well with point suction pyrometer measurements performed up to the middle of the boiler.
Fig. 5.
Retrieved temperature profile at 0 degree line-of-sight (blue) across 10 m for the Spectrum 2 from Figure 3. Constant temperature profile (1300 K) is plotted for comparison
The CO2 emission bands have a minor influence on the overall spectrum modelling because they are located in narrow spectral regions and temperature effects on the width of those spectral regions are quite small compared to that for the water. Some CO2 bands (marked by arrows in the Figure 4) also appear with self-inversion because of local flue gas cooling by the purge air in front of the sweeping probe (purged air is used to keep optics cleaned).
Fine structure between 2500–2800 cm–1 is due to the HCl emission band. The HCl emission appears on the top of grey-body continuum at 2564 cm–1 (3.9 μm) used by many IR pyrometers for surface temperature measurements. The HCl is a common byproduct in waste incinerator plants.
If necessary, a full emission spectrum modelling with use of H2O, CO2, CO (and HCl) spectroscopic data can be performed if information about respective concentrations along the LOS is required. This however can best be done with the use of an AI approach for example, (16, 17) when both gas and concentration profiles can be retrieved. It should be noted that for the temperature retrievals from IR emission measurements there is no need for data sets obtained at high spectral resolution. In contrast, a low spectral resolution and fast scanning FTIR spectrometer are the most important.
4. Conclusions
We have presented three examples about how spectroscopy-based systems can be used from UV to visible and IR spectral ranges for simultaneous gas and particle temperature measurements. The measurements have been done in situ in various difficult industrial environments. Apart from the temperature other valuable information about the process can be extracted from the measurements. The data analysis can be done with use of AI tools.
There is no ‘one-size-fits-all’ solution for how the temperature measurements can best be done. It depends on the particular process and access possibilities. Most combustion processes are highly dynamic because of the combustion event and flame and gas turbulent behaviour and therefore it is essential that the measurements are done: (a) in situ (i.e. without direct interaction with the environment); and (b) very fast in order to capture any changes that may occur in the process.
At the same time, recent developments in spectroscopic measurement techniques allow us to freely choose spectral- and time-measurement domains (for example, in UV–visible-IR) spectral ranges from a few microseconds to seconds or minutes measurement time per one dataset) and to adopt those to a particular application where temperature measurements are in question.
The key element of a spectroscopy-based system is a spectrometer. Currently there are several affordable commercial compact grating spectrometers available on the market which can be adapted to any spectral range in UV-visible. The word ‘compact’ automatically means ‘low-resolution’ (compare to high-end systems). However, being low-resolution should not be considered as an ‘issue’; instead it is a ‘benefit’ in a compact and robust system design for a selected application. High grade optical fibres and high quality and robust optical components are available too. Therefore an affordable and high quality spectroscopy system for a demanding application can be built.
In the IR spectral range, the most costly elements are the FTIR spectrometer and IR detectors. Although there are some small FTIRs on the market, further steps to their adaptation to a specific application are needed. Recent progress in 1D IR array developments opens possibilities for compact IR spectrometer developments. Although the price level of the IR arrays at the moment is relatively high (it is a market-driven process), still the costs for an IR spectrometer are expected to be lower than that for a small FTIR spectrometer. In one of our projects we have demonstrated that at certain conditions the performance of the custom built IR spectrometer can be at the same level as for a commercial FTIR one.
The most critical element in spectroscopy-based system development these days is reference spectroscopic data or databases and data treatment approaches, and those, fortunately, are in continuous development. Here the new and promising AI methods will play an important role. In addition, the current work is remarkable in that the flame thermometry technique has been validated using an independently verified portable ‘standard’ flame.
References
- 1.
J. V. Nicholas and D. R. White, “Traceable Temperatures: An Introduction to Temperature Measurement and Calibration”, 2nd Edn., John Wiley and Sons Ltd, Chichester, UK, 2001, 421 pp
- 2.
“Radiometric Temperature Measurements: I. Fundamentals”, eds. Z. M. Zhang, B. K. Tsai and G. Machin, Experimental Methods in the Physical Sciences, Vol. 42, Elsevier Inc, San Diego, USA, 2010
- 3.
“Radiometric Temperature Measurements: II. Applications”, eds. Z. M. Zhang, B. K. Tsai and G. Machin, Experimental Methods in the Physical Sciences, Vol. 43, Elsevier Inc, San Diego, USA, 2010
- 4.
L. Michalski, K. Eckersdorf, J. Kucharski and J. McGhee, “Temperature Measurement”, 2nd Edn., John Wiley and Sons Ltd, Chichester, UK, 2001
- 5.
P. R. N. Childs, “Practical Temperature Measurement”, Butterworth-Heinemann, Oxford, UK, 2001
- 6.
A. P. Thorne, “Spectrophysics”, 2nd Edn., Chapman and Hall Ltd, London, UK, 1988
- 7.
A. Book, “Optical Temperature Measurement in Combustion Plants”, Technical Report TR_017_202007_en, KELLER HCW GmbH, Ibbenbüren, Germany, 2020 LINK https://www.keller.de/dl.php?f=optical-temperature-measurement-in-combustion-plants&h=8a3bf569ce2a0ffL3Zhci93d3cvZnRwL2l0cy9JVFMvw7ZmZmVudGxpY2hlciBCZXJlaWNoIGZyZWllciBadWdhbmcvZW4vVGVjaG5pY2FsIFJlcG9ydHMvMDIgQXBwbGljYXRpb25zL1RSIE9wdGljYWwgdGVtcGVyYXR1cmUgbWVhc3VyZW1lbnQgaW4gY29tYnVzdGlvbiBwbGFudHNfSUQyNTEwXzIwMjAwN19lbi5wZGY
- 8.
M. F. Modest and S. Mazumder, “Radiative Heat Transfer”, 4th Edn., Elsevier Inc, San Diego, USA, 2022, 987 pp
- 9.
Y. Deguchi, “Industrial Applications of Laser Diagnostics”, Taylor and Francis Group LLC, Boca Raton, USA, 2012, 302 pp LINK https://doi.org/10.1201/b11497
- 10.
R. Himes, “Demonstration of the ZoloBOSS System at TVA’s Gallatin Station”, Document No. 1018182, Electric Power Research Institute, Palo Alto, USA, 2008, 90 pp LINK https://www.epri.com/research/products/000000000001018182
- 11.
K. B. Chung, F. C. Gouldin and G. J. Wolga, Appl. Opt., 1995, 34, (24), 5492 LINK https://doi.org/10.1364/AO.34.005492
- 12.
U.-S. Altmann and T. Förster, “Development of a Firing Diagnosis System for Utility Steam Generator using Soft-Computing Methods”, Final Report aFuE-Project, FMER Grant No. 1744X04, University of Applied Science Zittau/Goerlitz, Zittau, Germany, 2007
- 13.
I. E. Gordon, L. S. Rothman, R. J. Hargreaves, R. Hashemi, E. V. Karlovets, F. M. Skinner, E. K. Conway, C. Hill, R. V. Kochanov, Y. Tan, P. Wcisło, A. A. Finenko, K. Nelson, P. F. Bernath, M. Birk, V. Boudon, A. Campargue, K. V. Chance, A. Coustenis, B. J. Drouin, J.-M. Flaud, R. R. Gamache, J. T. Hodges, D. Jacquemart, E. J. Mlawer, A. V. Nikitin, V. I. Perevalov, M. Rotger, J. Tennyson, G. C. Toon, H. Tran, V. G. Tyuterev, E. M. Adkins, A. Baker, A. Barbe, E. Canè, A. G. Császár, A. Dudaryonok, O. Egorov, A. J. Fleisher, H. Fleurbaey, A. Foltynowicz, T. Furtenbacher, J. J. Harrison, J.-M. Hartmann, V.-M. Horneman, X. Huang, T. Karman, J. Karns, S. Kassi, I. Kleiner, V. Kofman, F. Kwabia-Tchana, N. N. Lavrentieva, T. J. Lee, D. A. Long, A. A. Lukashevskaya, O. M. Lyulin, V. Yu. Makhnev, W. Matt, S. T. Massie, M. Melosso, S. N. Mikhailenko, D. Mondelain, H. S. P. Müller, O. V. Naumenko, A. Perrin, O. L. Polyansky, E. Raddaoui, P. L. Raston, Z. D. Reed, M. Rey, C. Richard, R. Tóbiás, I. Sadiek, D. W. Schwenke, E. Starikova, K. Sung, F. Tamassia, S. A. Tashkun, J. Vander Auwera, I. A. Vasilenko, A. A. Vigasin, G. L. Villanueva, B. Vispoel, G. Wagner, A. Yachmenev and S. N. Yurchenko, J. Quant. Spectrosc. Radiat. Trans., 2022, 277, 107949 LINK https://doi.org/10.1016/j.jqsrt.2021.107949
- 14.
L. S. Rothman, I. E. Gordon, R. J. Barber, H. Dothe, R. R. Gamache, A. Goldman, V. Perevalov, S. A. Tashkun and J. Tennyson, J. Quant. Spectrosc. Radiat. Trans., 2010, 111, (15), 2139 LINK https://doi.org/10.1016/j.jqsrt.2010.05.001
- 15.
G. Sutton, A. Fateev, M. A. Rodríguez-Conejo, J. Meléndez and G. Guarnizo, Int. J. Thermophys., 2019, 40, (11), 99 LINK https://doi.org/10.1007/s10765-019-2557-6
- 16.
T. Ren, M. F. Modest, A. Fateev and S. Clausen, J. Quant. Spectrosc. Radiat. Trans., 2015, 151, 198 LINK https://doi.org/10.1016/j.jqsrt.2014.10.005
- 17.
T. Ren, M. F. Modest, A. Fateev, G. Sutton, W. Zhao and F. Rusu, Appl. Energy, 2019, 252, 113448 LINK https://doi.org/10.1016/j.apenergy.2019.113448
- 18.
A. Owens, A. Mitrushchenkov, S. N. Yurchenko and J. Tennyson, Month. Not. Roy. Astron. Soc., 2022, 516, (3), 3995 LINK https://doi.org/10.1093/mnras/stac2462
- 19.
J. J. de Groot and J. A. J. M. van Vliet, ‘Spectrum and Related Discharge Properties: Calculated and Measured Spectra’, in “The High-Pressure Sodium Lamp”, Ch. 3, Kluwer Technische Boeken BV, Deventer, The Netherlands, 1986, pp. 71–76
CombTec GmbH, Zittau, Germany: http://www.combtec.de/ (Accessed on 15th November 2022)
Acknowledgements
This project has partly received funding from the European Metrology Programme for Innovation and Research (EMPIR) co-financed by the Participating States and from the European Union’s Horizon 2020 research and innovation programme (Grant numbers 17IND07 Development of Measurement and Calibration Techniques for Dynamic pressures and Temperatures (DynPT), and 17IND04 Enhancing Process Efficiency Through Improved Temperature Measurement 2 (EMPRESS2)). This study was partly funded by the ProBu project that is funded by Innovation Fund Denmark (IFD project no. 8055-00014A), ROCKWOOL A/S, Denmark; FLSmidth A/S, Denmark; and DTU. The author is grateful to Jonathan Pearce, National Physical Laboratory, UK, for his valuable comments about the manuscript.
The Author
Related articles
Find an article
