WO2003100393A1

WO2003100393A1 - Photoacoustic detection method for measuring concentration of a non-hydrocarbon component of a methane-containing gas mixture

Info

Publication number: WO2003100393A1
Application number: PCT/HU2003/000038
Authority: WO
Inventors: Zsolt Bor; Zoltán Bozóki; Árpád MOHÁCSI; Sándor Puskás; Gábor SZABÓ; Miklós SZAKÁLL
Original assignee: MOL Magyar Olaj- és Gázipari Részvénytársaság; Szegedi Tudományegyetem
Priority date: 2002-05-24
Filing date: 2003-05-26
Publication date: 2003-12-04
Also published as: AU2003232375A1; EG23784A; HUP0201751A2; DE10392663T5; EP1511987A1; RU2004137093A; RU2336518C2; MA27505A1; HU225660B1; HU0201751D0; TNSN04232A1; AU2003232375A8

Abstract

The present invention relates to a photoacoustic detection method for measuring concentration of a non-hydrocarbon component of a methane-containing gas mixture. The essence of the method is that the photoacoustic absorption spectrum for the gas mixture is recorded over a suitably chosen wavelength range while the gas mixture is continuously flowing through the measuring apparatus, and then for determining the concentrationof the non-hydrocarbon component the thus obtained spectrum is used in combination with a photoacoustic signal generated by a reference cell (3) filled with a gas having predefined properties. The methane-containing gas mixtureand the non-hydrocarbon component are preferably chosen to be natural gas to be sent out to gas lines and water vapour, respectively. The invention hence allows high-precision detection, even under industrial conditions, of water vapour content present in concentrations as low as about 0.5 ppm in natural gas.

Description

PHOTOACOUSTIC DETECTION METHOD FOR MEASURING CONCENTRATION OF A NON-HYDROCARBON COMPONENT OF A METHANE-CONTAINING GAS MIXTURE

The present invention relates to a photoacoustic detection method applicable for concentration measurements in the field of natural gas industry. The quality and the impurity (eg. humidity or water vapour) content of natural gas to be fed into gas lines from various gas preparation technologies and to be used later as fuel, as well as safety and efficacy criteria for transportation via gas lines and conditions for trouble-free incineration by appropriate furnace installations of natural gas are under strict control prescribed by interna- tional standards (eg. EN ISO 11541 :2002, EN ISO 10101-1 :1998, 10101-2:1998 and 10101-3:1998). To comply with regulations, the natural gas to be sent out should be subjected to continuous monitoring or measuring. One of these measurements aims at determining the vapour content, or the water dew point T_dew of the dried natural gas ready for transportation; in particular, for carrying out such a measurement a number of different direct or indirect measuring techniques are known in natural gas industry. Direct methods are based on the separation of vapour and dry gas, followed by a water content determination. This class of methods involves total absorption methods (eg. of gravimetric or chemical type) and condensation methods based on freezeing out water vapour. Indi- rect methods include processes that aim at measuring one of the physical parameters functionally related to water content of the natural gas. Such measuring methods are, for example, electrolytic (coulombmet c) or dielectrometric in nature. These types of detection methods and the measuring apparatuses available therefor, however, are well known in the art, and hence are not discussed here in more detail.

The measuring methods and apparatuses widely used nowadays in gas industry allow detection of water vapour content or of water dew point T_dew of a gas to be transported with a relatively large error. This might lead to the appearance of malfunctions and various difficulties (eg. corrosion) on the user side, therefore inaccuracy is not acceptable. Hydrocarbons heavier than methane (C₂₊) present in a gas sample taken from natural gas in addition to methane, hydrogen sulphide (HS), and furthermore alcohols, mercaptanes and glycol, which are inevitable residues of the preparation techniques applied, can be considered mainly responsible for the inaccuracies appearing in the data measured. In order to eliminate deficiencies of vapour content detection devices used recently in gas industry, a study and a development of spectroscopic techniques were initiated. In the framework of spectroscopy (rotational and vibra- tional) spectral lines, ie. the spectra of gas-phase water molecules present in the natural gas are recorded. This means that there is no need to have the water content of the natural gas condensated or absorbed, ie. transformed into liquid or solid phase. According to the literature, two such techniques seem to be promising: the microwave gas spectroscopy and the infrared optical spectroscopy. However, so far these techniques have proven to be applicable only for detections under laboratory conditions; they have not been adapted to real con- ditions emerging in the field, ie. they have not been installed onto continuously operating gas lines yet. Some details of these techniques can be found eg. in the communication of V. A. Istomin published by IRC GAZPROM (No. 69, 1999) or in the paper by A. M. Ferber et al. (published in Measurement+Control, Vol. 34, 2001 , March issue). The main point of photoacoustic-based optical spectroscopic detection methods is that due to the absorption of illuminating light molecules of gases or vapours are excited from their ground states into higher energy states according to the rules of Quantum Mechanics. Relaxation from the excited state causes a heating-up in the spatial vicinity of the excitation. If the illumination takes place periodically, this heating-up will also show a periodic behaviour due to the periodic nature of the relaxation. This means that a periodic pressure difference builds up in gases and vapours at that volume where relaxation takes place. The periodic pressure difference results in a longitudinal wave, ie. a sound wave emerges which can be detected by means of a suitable apparatus. The core of measuring methods based on photoacoustic phenomenon is that a gas or a gas mixture containing different components is fed into a special volume known as the photoacoustic chamber in literature. The photoacoustic chamber is an acoustically optimalized chamber, which means that the sound of a certain acoustic frequency generated therein is amplified to a huge extent by the chamber. In particular, the illumination is provided by a laser light led through the photoacoustic chamber. The power of the laser light is continuously modulated at a frequency equal to one of the resonant frequencies of the chamber. In most cases the modulation is provided by switching on and off the source of the laser light. The wavelength of the illuminating laser light is fine-tuned to the absorption line(s) of a gas component to be studied. If the gas or the gas mixture fed into the chamber contains the gas component to be measured, a sound is generated by the absorption of the modulated illumination's light. This sound is then detected by means of a pressure change sensor, in particular by a microphone, arranged within the photoacoustic chamber. The amplitude (or strength) of the detected photoacoustic signal is proportional on the one hand to the lumi- nous flux of the illumination and on the other hand to the absorption coefficient of the gas component under study; the proportionality factor is set by the geometry of the chamber. The absorption coefficient, at the same time, depends on the number of excited molecules, ie. on the concentration of the gas component to be measured. The photoacoustic method, in comparison with other optical techniques, is extremely simple and does not require the application of precision-type or complicated optical systems. It can be easily automated and for carrying out a measurement thereby only small sample volumes (ie. a few cubic centimetres) are required. Nevertheless, the method is of large sensitivity despite of its sim- plicity; depending on the illuminating light source the photoacoustic method is suitable for revealing concentrations as low as ppm or ppb in magnitude. It is preferred for measuring a gas or gas mixture having the pressure of between 0.02 and 0.4 MPa; within this range of pressure, however, those measurements are the most preferred which take place at atmospheric pressure (ie. about 0.1 MPa). A further advantage of the photoacoustic methods is their selectivity and specificity, characteristic of optical methods in general. The photoacoustic method exhibits a uniquely broad dynamic range; the photoacoustic signal shows a linear dependency on the concentration of the gas component to be measured through 5-6 orders of magnitude. As a consequence, rapid and large changes in concentrations (eg. ranging throughout several orders of magnitude) can be traced, and additionally small changes of a huge concentration can also be observed and measured.

To analyse composition of gases, photoacoustic measuring methods and apparatuses to implement these methods are already known in literature.

Hungarian Patent No. 203.153B discloses a photoacoustic measuring method and a related apparatus for measuring composition of gases, wherein the apparatus is optically open but acoustically closed. In accordance with the measuring method, the acoustically closed photoacoustic chamber is opened optically, ie. for the incoming and outgoing illuminating laser light, openings are formed on opposite ends of the chamber. Hence the acoustical closure of the photoacoustic chamber becoming in this way also acoustically open, ie. sensitive to the ambient noise entering into the apparatus from the outside, is assured with noise suppression realized by means of low-pass filters.

US Patent No. 5,159,41 1 teaches a method and an apparatus associated thereto for measuring wavelength dependence of the strength and phase of a photoacoustic signal. When certain conditions are fulfilled, one can conclude on the presence of a given gas component from the fact that the phase recorded as a function of the continuously changing wavelength of the illuminating light source shows a sudden decrease followed by a rapid increase at a wavelength characteristic of the gas component; this indicates the absorption of the gas component under study at the wavelength at issue. The apparatus operates with a carbon dioxide gas laser (λ = 10μm), and the smallest measurable value of the gas component's concentration is relatively high, approximately about 500 ppm. Further disadvantage of the method and the apparatus is that they can be used only in combination with a special gas background. International Patent Application No. WO96/31765 describes a photoacoustic measuring device for selective measurement of gases and/or gas mix- tures carried out at a certain absorption line. The apparatus comprises, among other things, a photoacoustic chamber with a reference cell arranged behind it. The role of the reference cell is to provide a reference signal, on basis of which a shift in wavelength of the illuminating light source, or the incidental masking ef- fects due to the impurities of the gas under study could be compensated. A disadvantage of this apparatus is that the photoacoustic chamber should be closed in an airtight manner throughout the measurement, and hence investigation of a continuous gas flow cannot be effected by means of the present device.

By significant part of the cases dealt in detail in literature the (principally artificial) gas samples used in photoacoustic-based measurements comprise only two components: a so-called carrier gas that does not absorb at the optical wavelength of the light source, and the gas component itself to be measured that has a low concentration (tipically ranging from ppm to ppb) in the sample, and due to absorption of the exciting light it generates an acoustic signal and hence becomes detectable by means of the photoacoustic-based technique at issue. If the aim is to determine photoacoustically the concentration of a certain component of a more complex gas sample, eg. of a methane-containing gas mixture found in nature (such as eg. natural gas), it always represents a major problem that, in general, more than one component of the gas sample might ab- sorb at the same wavelength, and hence a resultant photoacoustical signal associated to the mixture of components is obtained. In most of such cases the contributions of individual components cannot be separated by means of simple methods (eg. by carrying out measurements only at a single wavelength). In those cases the so-called multicomponent analysis known from spectroscopy and also used in photoacoustic measuring techniques is applied for determining the concentration(s) of gas component(s) to be measured. Here, (photoacoustic) spectra are taken by measurements carried out at various wavelengths. In a first step, separate spectra are taken for each component of the gas sample as a function of the concentration of individual components. This step is the so-called calibration step, in the framework of which the calibration constants neccessary for evaluating subsequent measurements are calculated. Following this step, recording of the spectrum of the gas sample of unknown composition takes place at certain calibration wavelengths. Finally, the concentration of each gas component is determined from the spectrum thus obtained by means of algebraic calculations based on the calibration constants. Completion of the multi- component analysis for a natural gas mixture (ie. that contains a relatively large number of different components) is a time-consuming and complicated task. Moreover, for photoacoustic measurements it gives an accurate result only in those cases, wherein the concentration of components absorbing the illuminating light is jointly and separately relatively low (tipically at most about 0.1 per- centage by mass of the whole gas sample). A reason for this is that the components being present in large concentrations in the gas under study influence to a great extent the properties of the photoacoustic chamber that have essential role in generating the photoacoustic signal. As a consequence, the result obtained for the artificial sample with no components of great concentration used in the calibration step cannot be used for the gas mixture (also containing such components) under study at all, or only with a significant error. Since the overall concentration of light absorbing components in natural gas might tipically range from about several percentage up to almost one hundred percentage, in view of the above mentioned facts, it would be necessary to elaborate such a photoacoustic detection method by means of which the concentration of a given component (eg. water vapour) of a gas mixture with varying composition in time and, in general, containing a great portion of methane (eg. natural gas) could be easily and accurately determined in the presence of other components (preferably methane) also absorbing at the measuring wavelength and of further components (eg. hydrocarbons heavier than methane, carbon dioxide, various alcoholic vapours and the like) even for relatively low concentrations (eg. about 0.5 ppm) of the component to be measured.

The object of the present invention is therefore to work out a photoacoustic detection method that on the one hand satisfies the above mentioned re- quirements and on the other hand is suitable for and can be carried out under industrial circumstances, too. The aimed object of the present invention is reached by developing a photoacoustic detection method for measuring concentration of a non- hydrocarbon component of a methane-containing gas mixture, wherein the gas mixture is led through an acoustically optimalized photoacoustic measuring cell while it is illuminated by a periodically modulated light source operating at a wavelength within a certain wavelength range of the known absorption spectrum of the non-hydrocarbon component; by means of a pressure change sensor built into the photoacoustic measuring cell periodic changes in pressure due to the illumination are detected and converted into photoacoustic signals, the strength of which are also measured; by varying the wavelength of the illuminating light source a photoacoustic absorption spectrum of the gas mixture is recorded, wherein furthermore

(a) a measuring range is set in such a way that it contain at least two separate characteristic methane absorption wavelengths of different absorption strength and at least one characteristic absorption wavelength of the non- hydrocarbon component, wherein this latter wavelength differs from the methane absorption wavelengths concerned and is located therebetween;

(b) exact position/positions of the characteristic absorption wavelength/wavelengths of the non-hydrocarbon component is/are settled within the measur- ing range on basis of a reference photoacoustic absorption spectrum recorded within a reference cell arranged behind the measuring cell and illuminated by the light source simultaneously with the measuring cell, wherein the reference cell is filled up with a gas rich in the non-hydrocarbon component and being able to generate no photoacoustic signal alone; (c) photoacoustic signal's dependence on the concentration of the non- hydrocarbon component is established via the steps of leading at first a calibrating gas with a composition similar to that of the gas mixture through the photoacoustic measuring cell and in the meantime recording its photoacoustic spectrum in the measuring range; defining a quantity X_c depending purely on methane-concentration for two characteristic methane absorption wavelengths embedding a single charac- teristic absorption wavelength of the non-hydrocarbon component within the measuring range on basis of the obtained spectrum, and furthermore determining photoacoustic signal strength Y_c at one of the characteristic methane absorption wavelengths concerned; then varying the concentration of the non-hydrocarbon component of the calibrating gas; and determining the photoacoustic signal strength for each concentration of the non-hydrocarbon component;

(d) a quantity X_m depending purely on methane-concentration is defined for the chosen characteristic methane absorption wavelengths on basis of the photoacoustic absorption spectrum recorded for the gas mixture, and furthermore photoacoustic signal strength Y_m is determined at one of the characteristic methane absorption wavelengths concerned;

(e) the spectrum recorded for the gas mixture is subjected to a transformation achieved by a combination of the quantities X_c and X_m and the photoacoustic signal strengths Y_c and Y_m; and

(f) the sought concentration of the non-hydrocarbon component of the gas mixture is determined by using the spectrum transformed in step (e) as a start and by exploiting the relation between the concentration of the non- hydrocarbon component and the photoacoustic signal established in step (c).

The method according to the invention and its further advantages will be explained in detail with reference to the accompanied drawings, wherein

Figure 1 is a diagrammatic representation of a preferred embodiment of the measuring apparatus suitable for executing the method according to the in- vention;

Figure 2 shows photoacoustic spectra taken in the measuring cell (solid line) and in the reference cell (dashed line) of the apparatus depicted in Fig. 1 as the function of tuning of the illuminating light source for natural gas with a given water vapour content; and Figure 3 depicts the change in water vapour content of natural gas, detected in accordance with the invention, as a function of time within a 12-day measuring period extracted from a natural gas well, prepared and finally sent-out to a gas line.

Referring to Fig. 1 , a possible embodiment of the apparatus suitable for carrying out the photoacoustic-based detection method of the present invention contains an illuminating light source 1 tunable within a given wavelength range, a photoacoustic measuring cell 2 and a reference cell 3 both acoustically optimalized and equipped with pressure change sensors, embodied preferably in the form of microphones 9, an electronic unit 4 connected electrically to the microphones 9 and furthermore a controlling and signal processing unit, preferably in the form of a personal computer 5 connected electrically to the electronic unit 4 in order to govern the measurement and to acquire and to analyse data.

The light source 1 is prepared as a single-mode distributed feedback diode laser or an external cavity diode laser, both being rapidly, reliably and repro- ducibly tunable in a relatively narrow wavelength range (that has a tipical width of about 1 nm). In addition, the light source 1 is formed in such a way that its tuning range contains one or more significant (ie. well identifiable in wavelength) characteristic absorption lines of a component (eg. water vapour) having unknown concentration within the gas mixture (eg. natural gas) under study and that of methane, wherein any pairs of the absorption lines concerned at most partially overlap one another. If water vapour content of the natural gas is to be detected, the tuning range is defined by eg. a narrow wavelength range containing the absorption line at 1370.96 nm for water vapour. Further details of a possible embodiment of the light source 1 practically applicable to the method in accordance with the invention can be found eg. in the paper by A. Mohacsi et al. (published in Laser Physics, Vol. 10, 2000, No. 1 , pp. 1 to 4). The paper at issue discloses an external cavity diode laser exhibiting a beam power of about 2 mW and being tunable with high precision within the wavelength range of about 1365 to 1375 nm. Furthermore, the light source 1 can also be used under industrial circumstances, ie. it keeps its properties mentioned above under conditions characteristic of an industrial application, essentially in the presence of vibrations and temperature fluctuations. The measuring cell 2 and the reference cell 3 are arranged one after the other behind the light source 1 , preferably aligned therewith, in the path of illuminating light beam 6 emerging from the light source 1. The measuring cell 2 and the reference cell 3 are provided with optical windows 10 in the direction of propagation of the light beam 6. The optical windows 10 transmit the light beam 6 without significant loss of intensity, ie. they are almost non-absorbing. The measuring cell 2 has also equipped with a gas inlet 7 and a gas outlet 8; the gas mixture under study enters into and leaves from the measuring cell 2 through these means. The reference cell 3 contains the component to be detected at a high concentration, ie. in case of water vapour detection, it is filled with a gas of atmospheric pressure being fully or almost fully saturated by water vapour. The gas used (which, for the sake of simplicity, is ambient air in the present case) do not absorb itself within the tuning range of the light source 1 alone, ie. without water vapour content, and hence generate no acoustic signal of detectable magnitude. The measuring cell 2 and the reference cell 3 are acoustically optimalized, ie. they are characterized by the suppression of external noises disturbing the measurement accidentally, or, in case of the measuring cell 2, of noises generated by the gas mixture flowing through the measuring cell 2 in a continuous flow-through operation mode of the apparatus. Moreover, the meas- uring cell 2 and the reference cell 3 are formed geometrically in such a way that pressure waves (ie. sound) emerging therein due to the absorption of the periodically modulated (by a frequency in the kHz range) illuminating light beam 6 are amplified by the cells in a resonant manner, which facilitates the detection of the generated sound by the microphones 9 of high sensitivity. The electronic unit 4 is adapted to be suitable for controlling the current and the temperature of the illuminating light source 1 and for modulating the optical power thereof, for tuning the wavelength of the light beam 6 over a given range, for amplifying the electronic signals generated by the microphones 9 within the cells, for suppressing noise, averaging (in case of need), converting into digital signals and processing of the amplified microphone signals. This means that in order to execute the above operations, the electronic unit 4 contains various electronic subunits known perse.

The personal computer 5 is equipped with a software that carries out the photoacoustic detection method (to be discussed later in detail) according to the invention in an automated manner and communicates with the electronic unit 4. Further details and the construction of individual components of the photoacoustic apparatus used to execute the present measuring method can be found eg. in the paper by Z. Bozόki et al. (published in Measurement Scientific Technology, No. 10, 1999, pp. 999 to 1003). This paper discloses an automated detection technique for determining a non-stationary water vapour content of a gas (in particular synthetic air) having a preset composition.

The photoacoustic-based detection method for measuring water vapour content of a methane-containing gas mixture having varying composition comprises basically three steps. In a first step a wavelength range suitable for the measurement is chosen, this step is followed by a simplified calibration step, and finally the water vapour content of the given gas mixture is determined.

In choosing the proper wavelength range, the tuning range of the available light source 1 is taken into account. The distributed feedback laser diode or the external cavity laser diode used as the light source 1 can be tuned relatively rapidly, reliably and reproducibly only in a narrow wavelength range (that has a tipical width of at most 1 nm). Accordingly, the full wavelength range of the measurement should be chosen to be preferably less than 1 nm. It is a well- known fact that only molecules built up of few (tipically at most five) atoms exhibit separate rotational bands in their spectra at atmospheric pressures. Hence, only these molecules show a structured absorption spectrum in wavelength ranges narrower than 1 nm in width. If the wavelength-tuning takes place within a range of 1 nm in width, the absorption of larger molecules appears as a broad, non-characteristic, essentially constant background.

In light of the above, to execute a measurement aiming at quantitative determination of water vapour content of natural gas in particular, a (wavelength) range of at most 1 nm in width is chosen, wherein 1) among the components present in a gas sample under study only methane and water exhibit structured absorption, ie. all the other molecules also containing only few atoms (such as eg. carbon dioxide) show no measurable absorption in the given wavelength range; 2) there are at least two characteristic (ie. well-identifiable in wavelength) methane absorption lines, which can equally be absorption maxima or minima, that stand apart to an extent beyond the noise of the measurement by both their wavelengths and the extents of their absorption, and furthermore, water absorbs as low as possible at the wavelengths associated to these methane lines. Or, to put it in another way, this latter feature means that the absorption lines of water overlap the absorption lines of methane centred to the wavelengths at issue to the least possible extent. Of course, the gas components absorbing in broad bands also have got an absorption contribution at these wavelengths. The mentioned absorption lines of methane are preferably chosen in such manner that the difference in absorption values corresponding thereto be as large as possible. It is noted that reliability of the method is increased to a large extent if the photoacoustic signals obtained at the methane lines concerned are determined not for a single wavelength but are averaged around this wavelength; 3) there is at least one characteristic water absorption line, at the wavelength of which water absorbs significantly, and furthermore this water absorption line occupies a position between a pair of characteristic methane absorption lines, wherein the methane absorption at the wavelength associated to this position or in its close vicinity changes only slightly or is almost constant.

Our investigations led us to the conclusion that a wavelength range ap- proximately centred to the water absorption line of 1370.96 nm and having a total width of at most about 1 nm fully satisfies the above-detailed criteria. Moreover, positions of characteristic lines within this range can be unambiguously defined, which allows the automation/computerization of the process.

Position (ie. wavelength) of water absorption line suitable for the pur- poses of the present method is defined as follows: a light beam 6 emerging from the light source 1 is led through the reference cell 3 filled with the inert gas of about atmospheric pressure being saturated or almost saturated by water vapour. In the meanwhile the wavelength of the light beam 6 is swept over the tuning range in fine steps, and a photoacoustic signal is recorded at each wavelength. The thus obtained photoacoustic absorption reference spectrum for wa- ter vapour plotted against the change in the wavelength (expressed in relative units) of the light beam 6 is shown by the dashed line of Fig. 2. The suitable water vapour absorption line is just the characteristic peak 14 of the thus obtained water vapour spectrum. The difficulty of water vapour content detection of natural gas by means of a photoacoustic-based measurement is clearly illus- trated by the solid line of Fig. 2 which was generated by recording the photoacoustic signals emerging in a similar measurement, however, carried out for a natural gas sample with a stationary water vapour content. This curve of Fig. 2 clearly shows that in a spectrum taken for the natural gas sample, the characteristic methane lines 12, 13 suppress remarkably the characteristic water va- pour peak 14; in this spectrum the characteristic water vapour peak 14 shows up only as a tiny peak 11. This means that by taking merely the spectrum of the natural gas sample into account the position of the water absorption line could not be defined at all, or only extremely inaccurately. Hence, in the measuring method according to the invention the role of the reference cell 3 is always to provide accurate positions of water vapour lines in the photoacoustic spectra taken within the measuring cell 2 filled with the gas mixture under study, ie. with the natural gas containing water vapour in the present case.

During the calibration, contrary to the methods used nowadays, the ratio of the various non-water components of the gas mixture fed through the meas- uring cell 2 is not modified, only water vapour content of the gas used for calibration is altered. It is important that a knowledge of the calibrating gas's composition is not a requisite for the application of the present method. For completing the calibration, however, such a gas sample is used preferably, the composition of which (considering the methane content, hydrocarbon components heavier than methane, etc.) resembles that of a gas mixture to be detected subse- quently. The only essential requirement to be met by the calibrating gas is that its methane content should be detectable.

In a first step of the calibration, a methane-containing gas sample (which preferably has a similar composition to that of the gas mixture, and preferably contains water vapour in low concentrations) is lead through the measuring cell 2 and, in the meanwhile, the photoacoustic spectrum for this gas sample is recorded by tuning the wavelength of the light source 1 over the chosen wavelength range. Next, characteristic methane absorption points are defined that can be equally positions (in wavelength) of absorption minima and maxima in this photoacoustic spectrum. Thereafter, a quantity X_c dependent merely on methane concentration, but independent of other components' concentrations of the calibrating gas is calculated at the chosen points. If, particularly, two characteristic methane absorption points are chosen, this quantity can be defined eg. as a difference between the photoacoustic signal strengths measured at the two points (ie. at the two characteristic methane absorption wavelengths). The thus obtained quantity X_c is directly proportional to the methane content of the gas sample. In addition to this quantity, the photoacoustic signal strength Y_c is also determined at one of the methane absorption points concerned.

In a second step of the calibration, gas samples of given compositions with various amounts of water vapour are fed through the measuring cell 2 successively, and by means of the light source 1 adjusted to the characteristic water line settled unambiguously by the reference cell 3, the photoacoustic spectra are recorded. In this way water vapour dependency of the photoacoustic signal is determined. Gas samples of totally identical composition but with different water content are prepared eg. via the steps of dividing a starting dry gas sample into two streams by means of appropriate members, watering one of the streams eg. by leading it through water and/or water vapour, and then mixing the thus watered stream with the dry stream in various ratios. Of course, the gas samples with different water vapour content can be prepared in other ways too; never- theless, the method described here is particularly suitable for being carried out in a fully automated manner. It is noted that even for a totally dry gas sample, a photoacoustic signal of non-zero strength can be detected at the chosen characteristic water vapour absorption wavelength. This is on the one hand due to the methane, and on the other hand to other components also present in the gas sample. After the calibration of the measuring apparatus has been completed, ie. having revealed the photoacoustic signal's dependence on water vapour concentration for a fixed hydrocarbon composition, water vapour content of a methane-containing gas mixture varying in composition can also be determined with high precision. For this purpose, the gas mixture under study is led through the measuring cell 2 and by tuning the light source 1 in wavelength and by modulating the light beam 6 in intensity in accordance with photoacoustic measuring techniques, its photoacoustic spectrum is recorded in the chosen measuring range. As a next step a quantity X_m as the difference of the photoacoustic signal strengths obtained for the gas mixture at the characteristic methane absorption points chosen previously in the calibration step is calculated. The thus obtained quantity depends purely on the methane concentration. Now, taking the quantity X_c derived in a similar manner during the calibration, a ratio X_c X_m is created, and by multiplying the spectrum obtained in the present step by this ratio a further spectrum is derived that is independent of changes in methane concentra- tion appearing in the measurement as a function of time. This latter spectrum does not depend on changes in methane concentration with respect to time, since in case of a change in the concentration, the absorption lines of the methane spectrum vary proportionally to each other. The multiplication of the spectrum by the ratio X_c Xm is realized via multiplications by the ratio Xc/Xm of the in- dividual photoacoustic signals obtained for each wavelength within the measuring range. The spectrum is then "shifted" along the photoacoustic signal axis (ie. along the vertical axis) until the photoacoustic signal strength Y_m measured at one of the methane absorption points chosen previously becomes equal to the photoacoustic signal strength Y_c obtained during the calibration (this step is ac- tually equivalent with a shift of the spectrum by the value of I Y_c-Y_m I )■ The application of this transformation eliminates the measuring error due to the concen- tration change in time of gas components generating the broad absorption background in the measuring range. Finally, applying the calibration's result at the wavelength of the characteristic water absorption line settled accurately by the reference cell 3, the water vapour concentration of a gas sample with a non- stationary composition is derived from the photoacoustic signal obtained therefor. Here (ie. for a gas mixture with a composition differing from that of the gas sample used for the calibration) the results obtained previously in the calibration process fully hold, as the photoacoustic spectrum recorded for the gas mixture under study was transformed by the multiplication and shifting operations into a spectrum generated in the calibration step.

If only few measurements to be done (eg. in order to check the water content of the natural gas sent-out to a gas line), the above-detailed detection method can be carried out manually. If, however, lots of successive measurements are required (eg. in case of the continuous monitoring of water vapour content), the present detection method can be executed by having a suitable computer program run in the personal computer 5, ie. in an automated manner. Figure 3 illustrates the change of water vapour content as a function of time measured continuously within a 12-day period by the method in accordance with the invention to monitor natural gas extracted from a gas well, prepared and then sent-out to a gas line; the measurements were carried out in a fully automated manner.

Briefly summarized: the photoacoustic-based method according to the invention provides a high-precision and extremely sensitive detection process (with a limit of as low as about 0.5 ppm) to measure water vapour content of a methane-containing gas mixture with varying composition, preferentially of natural gas to be sent out to a gas line after preparation. The present method can be easily automated, is also applicable under industrial circumstances in the field, and can be carried out continuously without abruption. A major characteristic of the method is that it gives accurate and reliable results even for significant methane contents. Another advantageous feature of the present method is that it can be used successfully even in such cases where the gas mixture to be measured contains inevitable alcoholic vapour residuals coming from the antecedent preparation processes. The reason for this is that alcoholic vapours show a broad, wavelength-independent absorption within the measuring range chosen in accordance with the invention at atmospheric pressures, and hence their in- fluence can be separated from the signals obtained. This latter feature represents a basic difference and an important advantage in comparison with other water vapour detection methods (eg. dew point measurements based on condensation), wherein the presence of alcoholic vapours distorts the result of the measurement up to an inacceptable extent. A further advantage of the method according to the invention is that a net concentration of hydrocarbon components heavier than methane present in the gas mixture under study can also be determined in a simple and accurate manner by using it.

Claims

1. Photoacoustic detection method for measuring concentration of a non- hydrocarbon component of a methane-containing gas mixture, wherein the gas mixture is led through an acoustically optimalized photoacoustic measuring cell (2) while it is illuminated by a periodically modulated light source (1) operating at a wavelength within a certain wavelength range of the known absorption spectrum of the non-hydrocarbon component; by means of a pressure change sensor built into the photoacoustic measuring cell (2) periodic changes in pressure due to the illumination are detected and converted into photoacoustic signals, the strength of which are also measured; by varying the wavelength of the illuminating light source (1) a photoacoustic absorption spectrum of the gas mixture is recorded, characterized in that

(b) exact position/positions of the characteristic absorption wavelength/wavelengths of the non-hydrocarbon component is/are settled within the measur- ing range on basis of a reference photoacoustic absorption spectrum recorded within a reference cell (3) arranged behind the measuring cell (2) and illuminated by the light source (1) simultaneously with the measuring cell (2), wherein the reference cell (3) is filled up with a gas rich in the non- hydrocarbon component and being able to generate no photoacoustic signal alone;

(c) photoacoustic signal's dependence on the concentration of the non- hydrocarbon component is established via the steps of leading at first a calibrating gas with a composition similar to that of the gas mixture through the photoacoustic measuring cell (2) and in the mean- time recording its photoacoustic spectrum in the measuring range; defining a quantity X_c depending purely on methane-concentration for two characteristic methane absorption wavelengths embedding a single characteristic absorption wavelength of the non-hydrocarbon component within the measuring range on basis of the obtained spectrum, and furthermore deter- mining photoacoustic signal strength Y_c at one of the characteristic methane absorption wavelengths concerned; then varying the concentration of the non-hydrocarbon component of the calibrating gas; and determining the photoacoustic signal strength for each concentration of the non-hydrocarbon component;

(d) a quantity X_m depending purely on methane-concentration is defined for the chosen characteristic methane absorption wavelengths on basis of the photoacoustic absorption spectrum recorded for the gas mixture, and furthermore photoacoustic signal strength Y_m is determined at one of the charac- teristic methane absorption wavelengths concerned;

(f) the sought concentration of the non-hydrocarbon component of the gas mix- ture is determined by using the spectrum transformed in step (e) as a start and by exploiting the relation between the concentration of the non- hydrocarbon component and the photoacoustic signal established in step (c).

2. The photoacoustic detection method according to Claim 1 , characterized in that the gas mixture is natural gas and the non-hydrocarbon component is water vapour.

3. The photoacoustic detection method according to Claim 2, characterized in that a wavelength range of at most 1 nm in width centred approximately on the absorption line at 1370.96 nm for water vapour at ambient temperature and atmospheric pressure is chosen as the measuring range.

4. The photoacoustic detection method according to any of the preceding

Claims, characterized in that a distributed feedback diode laser or an external cavity diode laser both tunable within a wavelength range ranging from 1365 to 1375 nm is used as the light source (1).

5. The photoacoustic detection method according to any of the preceding Claims, characterized in that quantities X_c and X_m are defined as the difference in the photoacoustic signal strengths obtained at the chosen characteristic methane absorption wavelengths for the calibrating gas and for the gas mixture, respectively.

6. The photoacoustic detection method according to any of the preceding Claims, characterized in that transformation of the photoacoustic spectrum re- corded for the gas mixture is achieved by the steps of defining the ratio Xc/X_m> multiplying the spectrum by this ratio, and then equalizing photoacoustic absorption strengths Y_c and Y_m of the spectrum obtained by the multiplication via a shift by the value of | Y_c-Y_m I ■

7. The photoacoustic detection method according to any of the preceding Claims, characterized in that it is carried out by means of a computer program run by a personal computer (5).