Raman spectrometry quantification method

Abstract

Method for quantification by Raman spectrometry This method comprises: - the measurement (TON), in the presence of an excitation beam, of values ​​S3(λ) and S4(λ), at a wavelength λ, respectively, of optical signals S3 and S4 measured with polarization directions, respectively, parallel and perpendicular to a predetermined direction, - the measurement (TOFF), in the absence of the excitation beam, of values ​​S1(λ) and S2(λ), at the wavelength λ, respectively, of optical signals S1 and S2 measured with polarization directions, respectively, parallel and perpendicular to the predetermined direction, and - the determination (230) of a quantity of a gaseous component present inside a containment vessel of a nuclear reactor, from a difference [(S3(λ) / t3 – S1(λ) / t1] and a difference [S4(λ) / t4 – S2(λ) / t2], where t1, t2, t3 and t4 are time window durations during which the signals S1, S2, S3 and S4 are measured. Fig. 5

Description

Title of the invention: Method of quantification by Raman spectrometry

[0001] The invention relates to a method and a device for quantifying, by Raman spectrometry, at least one gaseous component inside a containment enclosure of a nuclear reactor.

[0002] For example, such a quantification device is used to measure the concentrations of gaseous components in a gaseous medium located inside the containment of a nuclear power plant. In this case, the containment is known as the "reactor building" (RB). Indeed, nuclear power plants used for the production of electrical energy comprise, inside a reactor building, a vessel containing the core of the nuclear reactor which can be cooled, inside the vessel, by pressurized or boiling light water or by heavy water. In the case of a nuclear accident resulting, for example, in a loss of the core cooling function, there is an increase in the temperature of the core components and in particular of the nuclear fuel. The increase in the core temperature can lead to a meltdown of the internal components of the nuclear reactor.

[0003] At high temperatures, metals become very reactive and oxidize on contact with the water vapor present in the reactor vessel containing the core. The cladding of the fuel assemblies, which are made of a zirconium alloy, are elements which are first subjected to these oxidation reactions. The zirconium is oxidized by the water vapor following the reaction Zr + 2H20 -> ZrO2 + 2H2. This results in a very rapid release of a significant quantity of hydrogen.

[0004] In the event of a rupture of the reactor vessel, the interaction between the corium (molten core) and the concrete base of the containment vessel leads to the production of hydrogen (H2), carbon monoxide (CO) and carbon dioxide (CO2).

[0005] Thus, during a nuclear accident, the containment enclosure may contain flammable gases (H2 and CO) and oxygen which can lead to the formation of an explosive atmosphere (ATEX) and the risk of loss of the integrity of the containment enclosure. This risk is known as the "hydrogen risk" or "H2 risk".

[0006] Therefore, it is important to be able to measure the concentration of gaseous components present in the reactor building during a nuclear accident.

[0007] In particular, the presence of gaseous components such as hydrogen, oxygen and Carbon monoxide can lead to explosive mixtures. It is therefore necessary to implement control measures to determine the nature and proportion of the main gaseous components present in the reactor building. This makes it possible to assess the risk of explosion and to take the necessary preventive actions to mitigate this risk.

[0008] Application FR2733050A1 describes a device for quantifying, by Raman spectrometry, gaseous components present inside the reactor building of a nuclear power plant. This device comprises one or more probes located inside the reactor building and a processing unit located outside. The processing unit is connected to the probes by fibers and electrical cables. Thus, its implementation requires providing passages through the reactor building. This very little degrades the sealing of the reactor building with respect to radiation and radioactive materials. In this respect, this device is very advantageous compared to other known quantification devices which require extracting a sample of the gaseous medium to the outside of the reactor building for analysis. Indeed, in the latter case, it is necessary to use tubes which pass through the reactor building to connect the inside of the reactor building to the outside.Such tubes are problematic because they call into question the tightness of the reactor building.

[0009] The following article also describes in detail a Raman spectrometry quantification device that can be used in the nuclear field: S. Magne et al: “In situ gas monitoring by fiber-coupled Raman spectrometry for H2-risk management in nuclear containment during a severe nuclear accident”, IEEE Trans. Nucl. Sci. 67 (4), 2020, pp. 617-624. Hereinafter, this article is referred to as “Magne2020”. In particular, this article proposes a solution for eliminating background noise in the measured optical signal. In particular, for this, a scientific CCD (“Charge Coupled Device”) imager is used, which remains a relatively fragile and expensive laboratory tool.

[0010] The invention aims to propose a quantification method at least as precise as the method of the article Magne2020 while allowing simplified implementation.

[0011] The subject of the invention is therefore a method for quantifying, by Raman spectrometry, at least one gaseous component inside a confinement enclosure of a nuclear reactor, this method comprising a first measurement period during which:

[0012] - a monochromatic excitation beam excites, inside the enclosure of confinement, a sample of the gaseous medium located inside this confinement enclosure,

[0013] - a reading unit measures the optical signal scattered, by Raman effect, by the sample excited by the excitation beam, and

[0014] - an electronic calculator determines, from the optical signal measured by the unit reading, a quantity of the gaseous component present in the sample,

[0015] in which:

[0016] - during the first measurement period: the excitation beam is polarized li nearly in a predetermined direction, and the reading unit measures values ​​S3(X) and S4(X), respectively, of an optical signal S3 and an optical signal S4 scattered by the sample, where:

[0017] - the optical signals S3 and S4 are the optical signals measured with directions of polarization, respectively, parallel and perpendicular to the predetermined direction of polarization of the excitation beam, when the sample is excited by the excitation beam,

[0018] - X is a wavelength of interest, i.e. a wavelength at which appears, due to the Raman effect, a line in the spectrum of the scattered optical signal when the gaseous component to be quantified is present in the sample, and

[0019] - the method also comprises a second measurement period in the absence of the excitation beam during which the reading unit measures values ​​S1(X) and S2(X), respectively, of an optical signal SI and an optical signal S2 scattered by the sample, where the optical signals SI and S2 are the optical signals measured with polarization directions, respectively, parallel and perpendicular to the predetermined direction of polarization of the excitation beam used during the first measurement period, when the excitation beam is absent, and

[0020] - the calculator determines the quantity of the gaseous component present in the sample from the difference [(S3(X) / t3 - S1(X) / t1] and the difference [S4(X) / t4 - S2(X) / t2], where t1, t2, t3 and t4 are the durations of the time windows during which the signals SI, S2, S3 and S4 are measured to obtain, respectively, the values ​​S1(X), S2(X), S3(X) and S4(X) in number of counts.

[0021] Embodiments of this method may include one or more of the following features:

[0022] 1) The calculator determines the quantity of the gaseous component present in the sample using the following relation: r(X) = [S3(X) / t3 - Sl(X) / tl] -k(X)*[S4(X) / t4 - S2(X) / t2], where:

[0023] - r(X) is a value representative of the quantity of the gaseous component present in the sample expressed in counts per second,

[0024] - k(X) is the value, for the wavelength X, of a scale coefficient k pre determined.

[0025] 2) The ratios tl / t2 and t3 / t4 are equal.

[0026] 3) The method comprises, prior to the execution of the first and second measurement periods, a calibration phase during which:

[0027] - the monochromatic excitation beam excites a reference sample, and

[0028] - a sensor measures a Spara optical signal and a Sperp optical signal diffused by the reference sample excited by the excitation beam, the optical signals Spara and Sperp being the optical signals scattered by the measured reference sample with polarization directions, respectively, parallel and perpendicular to the predetermined direction of polarization of the excitation beam, and the Spara and Sperp signals being measured at wavelengths X; different from the wavelengths at which, due to the Raman effect, lines appear in the spectrum of the optical signal scattered by this reference sample, then

[0029] - the value k(X) of the coefficient k for each wavelength of interest / . is calculated from the measured Spara and Sperp optical signals, then recorded in the calculator.

[0030] 4) During the calibration phase:

[0031] - several values ​​k(X;) of the coefficient k are calculated using the following relation: k(Xi) = [Spara(Xi) - DN] / [Sperp(X;) - DN], where:

[0032] - Spara(Xi) and Sperp(X;) are the values, respectively, of the optical signals Spara and Sperp at wavelength Xi5 and

[0033] - DN is the self-noise of the reading unit, i.e. the optical signal measured by the reading unit when no optical signal is received by the reading unit, then

[0034] - the value k(X) is calculated by interpolation from the values ​​k(X;).

[0035] 5) The first and second measurement periods follow each other immediately and the duration of each of these first and second measurement periods is less than 10 s.

[0036] 6)

[0037] - the first and second periods and the determination of the quantity of the component gaseous present in the sample from the values ​​S1(X), S2(X), S3(X) and S4(X) measured during these first and second periods, form a cycle (Ci) of measurements, and

[0038] - the method comprises repeating this measurement cycle several times during a time window of less than fifteen minutes, and

[0039] - the calculation, by the calculator, of a more precise value of the quantity of the component gaseous by averaging the quantities determined during each of the measurement cycles contained within this time window.

[0040] 7) During the first and second periods, the optical signals S1, S2, S3 and S4 are measured for several different X-ray wavelengths of interest, each of these wavelengths of interest corresponds to a particular gaseous component to be quantified in the excited sample, the measurements at these different wavelengths of interest being carried out simultaneously.

[0041] 8) The wavelengths of interest comprise wavelengths of interest cor- responding to water vapor, oxygen and nitrogen.

[0042] 9) The wavelengths of interest include wavelengths of interest cor responding to carbon monoxide and carbon dioxide.

[0043] 10) A wavelength of interest corresponds to hydrogen.

[0044] 11) The excitation beam is a continuous laser beam.

[0045] 12) The method comprises:

[0046] - the measurement of the total pressure inside the enclosure,

[0047] - the calculation of partial pressures of hydrogen, carbon dioxide, oxygen, carbon monoxide and nitrogen from quantities quantified by Raman spectrometry of these gaseous components, at least one of these quantities being determined from the difference [(S3(X) / t3 - S1(X) / t1] and the difference [S4(X) / t4 -S2(X) / t2], and

[0048] - the calculation of the partial pressure of water vapor from the difference between the total pressure PT minus the sum of the calculated partial pressures of hydrogen, carbon dioxide, oxygen, carbon monoxide and nitrogen.

[0049] The invention also relates to a device for quantifying, by Raman spectrometry, at least one gaseous component inside a confinement enclosure of a nuclear reactor, this device comprising:

[0050] - a laser source capable of emitting a monochromatic excitation beam which excites, inside the containment enclosure, a sample of the gaseous medium located inside this containment enclosure,

[0051] - a reading unit capable of measuring the optical signal scattered, by Raman effect, by the sample excited by the excitation beam, and

[0052] - an electronic calculator configured to determine, from the optical signal measured by the reading unit, a quantity of the gaseous component present in the sample,

[0053] in which the calculator (36) is configured to automatically perform the following steps:

[0054] - during a first measurement period: the excitation beam is polarized li nearly in a predetermined direction, and the reading unit measures values ​​S3(X) and S4(X), respectively, of an optical signal S3 and an optical signal S4 scattered by the sample, where:

[0055] - the optical signals S3 and S4 are the optical signals measured with directions of polarization, respectively, parallel and perpendicular to the predetermined direction of polarization of the excitation beam, when the sample is excited by the excitation beam,

[0056] - X is a wavelength of interest, i.e. a wavelength at which appears, due to the Raman effect, a line in the spectrum of the scattered optical signal when the gaseous component to be quantified is present in the sample, and

[0057] - during a second measurement period in the absence of the excitation beam, the unit of reading measurement of the values ​​S1(X) and S2(X), respectively, of an optical signal SI and an optical signal S2 scattered by the sample, where the optical signals SI and S2 are the optical signals measured with polarization directions, respectively, parallel and perpendicular to the predetermined direction of polarization of the excitation beam used during the first measurement period, when the excitation beam is absent, and

[0058] - the calculator determines the quantity of the gaseous component present in the sample from the difference [(S3(X) / t3 - S1 (7.) / t 1 ] and the difference [S4(X) / t4 - S2(X) / t2], where tl, t2, t3 and t4 are the durations of the time windows during which the signals SI, S2, S3 and S4 are measured to obtain, respectively, the values ​​S1(X), S2(X), S3(X) and S4(X) in number of counts.

[0059] Embodiments of this device may include the following characteristic: the reading unit includes only one photodetector for each wavelength of interest.

[0060] The invention will be better understood on reading the description which follows, given solely by way of non-limiting example and made with reference to the drawings in which:

[0061] - [Fig.l] is a schematic illustration of the architecture of a quantization device fication, by Raman spectrometry, of gaseous components inside a containment vessel of a nuclear reactor,

[0062] - [Fig.2] is an illustration of a Raman spectrum of the gaseous components at quantify,

[0063] - [Fig.3] is a schematic illustration of the architecture of a probe of the device of [Fig.l],

[0064] - [Fig.4] is an illustration of Raman spectra in the presence of radiation Cherenkov,

[0065] - [Fig.5] is a flowchart of a quantification method using the device of [Fig.l],

[0066] - [Fig.6] is a timing diagram of different operating periods of the device of [Fig.l].

[0067] In this description, the terminology, conventions and definitions of the terms used in this text are introduced in a chapter I. Then, a detailed example of an embodiment is described in a chapter II with reference to the figures. In a chapter III, variants of this embodiment are presented. Finally, the advantages of the different embodiments are specified in a chapter IV.

[0068] Chapter I: Definitions, terminologies and conventions:

[0069] In the figures, the same references are used to designate the same elements.

[0070] In the remainder of this description, the characteristics and functions well known to those skilled in the art are not described in detail.

[0071] The symbol “*” denotes scalar multiplication.

[0072] A gaseous component is typically a gaseous molecule.

[0073] The terms “quantification” and “quantify” designate the operation of measuring the quantity of a gaseous component and, in particular, a concentration expressed in % or a partial pressure expressed in Pa.

[0074] The term "quantity", when used without further specification, refers to both an absolute quantity, such as a partial pressure, and a relative quantity, such as a concentration in %.

[0075] In this text, a relative quantity is a proportion, for example expressed in percent, of the gaseous component within the gaseous medium.

[0076] An absolute quantity is typically the partial pressure of the gas component.

[0077] A Raman spectrum is the power spectrum of the optical signal scattered by the Raman effect.

[0078] In this text, the optical signal scattered by the Raman effect is also called “Raman scattering optical signal”.

[0079] The term "nuclear accident" of loss of core cooling refers to any accident that leads to the partial or complete meltdown of the core of the nuclear reactor. In such an accident, hydrogen is produced. Eventually, other gaseous components are produced such as carbon monoxide and carbon dioxide. Such an accident often originates from a total or partial loss of cooling of the core of the nuclear reactor.

[0080] A beam is considered “monochromatic” when the half-width of its single line, in its power spectrum, is less than 1 nm and, preferably, less than 0.5 nm.

[0081] Chapter II: Example of embodiment

[0082] [Fig.l] represents a device 2 for quantifying, by Raman spectrometry, at least one gaseous component inside a containment enclosure 4 of a nuclear reactor supported by a concrete base 7. Subsequently, the description of the device 2 is made in the particular case where the enclosure 4 is a reactor building of a nuclear power plant which produces electricity. The structure of a reactor building is well known. To simplify [Fig.l], only the following elements contained inside the reactor building have been schematically represented:

[0083] - core 6 of the nuclear reactor,

[0084] - the tank 8 in which the heart 6 is baked, and

[0085] - the gaseous medium 10 in which the different elements located inside are bathed of enclosure 4.

[0086] The enclosure 4 is sealed, as far as possible, against radiation and radioactive materials in order to prevent them from escaping into the external environment, including, if possible, in the event of a nuclear accident. Typically, the enclosure 4 and the base 7 are essentially made of concrete.

[0087] The gaseous medium 10 is typically air, that is to say a gaseous mixture composed, at more than 98% by volume, of oxygen (O2) and nitrogen (N2) in known proportions.

[0088] In this example, the device 2 is designed to quantify the following gaseous components: hydrogen (H2), water vapor (H2O), oxygen (O2), nitrogen (N2), carbon monoxide (CO) and carbon dioxide (CO2). Indeed, the H2 risk is estimated on the basis of the quantities of hydrogen, air (O2, N2) and water vapor (H2O) represented on a Shapiro-Moffette ternary diagram. The measurement of the quantities of carbon monoxide and carbon dioxide makes it possible to monitor the corium-concrete interaction (Molten Core Concrete Interaction - MCCI). In this particular context, the term "gaseous components" therefore designates the gaseous components belonging to the group consisting of hydrogen (H2), water vapor (H2O), oxygen (O2), nitrogen (N2), carbon monoxide (CO) and carbon dioxide (CO2).

[0089] The device 2 comprises, fixed inside the enclosure 4:

[0090] - one or more probes 20 located at different locations to measure the concentrations of gaseous components at these different locations,

[0091] - one or more pressure sensors 22, and

[0092] - one or more temperature sensors 24.

[0093] Subsequently, the device 4 is described in the simplified case where the device 2 comprises a single probe 20, a single pressure sensor 22 and a single temperature sensor 24. However, everything described in this simplified case also applies to the case where there are several probes 20 and / or several pressure and / or temperature sensors 22 or 24. Typically, the temperature sensor 24 is placed in the probe 20. Conversely, a single pressure sensor 22 common to all the probes 20 may suffice.

[0094] The device 2 also comprises, outside the enclosure 4, a processing unit 30 which processes the data transmitted by the probe 20 and the sensors 22 and 24 to quantify the presence of the gaseous components in the medium 10.

[0095] The processing unit 30 comprises for this purpose: a laser source 32, a reading unit 34, an electronic computer 36, a man-machine interface 38, and a power source 40.

[0096] The probe 20 is connected to the processing unit 30 via two optical fibers 42 and 44. The first fiber 42 brings an excitation beam, generated by the laser source 32, inside the probe 20. For this purpose, its proximal end is connected to the laser source 32 and its distal end opens inside the probe 20. The second fiber 44 brings the optical signal scattered by the Raman effect from the probe 20 to the reading unit 34. For this purpose, its proximal end is connected to the reading unit 34 and its distal end opens inside the probe 20. These fibers 42 and 44 pass through the enclosure 4, each using its own sealed optical passage better known by the acronym OPA (“Optical Penetration Assembly”). Typically, fibers 42 and 44 are protected by a flexible stainless steel sheath, coated with a black PVC sheath, which is light-tight and decontaminable.

[0097] By way of example, the first fiber 42 has a core diameter of 400 pm and a numerical aperture of between 0.22 and 0.26. Still by way of example, the second fiber 44 has a diameter of 600 pm and a numerical aperture of between 0.22 and 0.37.

[0098] The probe 20 is also connected to the unit 30 by means of several electrical cables collectively designated by the reference numeral 46. Each cable 46 passes through the enclosure 4 by using an electrical sealed crossing better known by the acronym EPA (“Electrical Penetration Assembly”). These cables 46 make it possible in particular to power the probe 20 and to convey control signals for the probe 20.

[0099] The sensors 22 and 24 are connected to the unit 30 by electrical cables, respectively, 48 and 49. These cables 48, 49 pass through the enclosure 4, each using a respective sealed electrical crossing.

[0100] To simplify [Fig.l], the oblique lines on each cable 46, 48 and 49 indicate that this cable is implemented, in practice, by several electrical cables. In particular, the probe 20 and the sensors 22, 24 are connected to ground via at least one electrical cable.

[0101] The laser source 32 emits an excitation beam. This beam is a monochromatic beam at a wavelength Xb. The wavelength Xi is preferably between 730 nm and 750 nm because it is in this range that the radiation-induced attenuation or RIA (Radiation-Indue Attenuation) of the fibers is the lowest. For example, here, the wavelength Xi is equal to 750 nm.

[0102] The power of the optical signal generated by the Raman effect is in particular proportional to the power of the excitation beam. Here, to improve the sensitivity and precision of the device 2, the power of the excitation beam is greater than 500 mW and, preferably, greater than or equal to 1 W. In this example, the power of the excitation beam is chosen so that its power inside the gas mixture is greater than or equal to 180 mW. To generate such an excitation beam with a small footprint, the laser source 32 is a continuous laser source with power equal to or greater than one Watt.

[0103] The reading unit 34 makes it possible to record the power of the optical signal at the different wavelengths of interest. Here, the wavelengths of interest are equal to the wavelengths at which lines, characteristic of the presence of the gaseous components to be quantified, appear in the power spectrum of the optical signal scattered by the Raman effect by the medium 10. Subsequently, the wavelengths of interest of hydrogen, carbon dioxide, oxygen, carbon monoxide, nitrogen and water vapor are noted, respectively, XH2, / <02- ^02, XCo, XN2 and XH20. These wavelengths of interest are known and represented on the Raman spectrum of [Fig.2],

[0104] The spectrum of [Fig.2] is reconstructed from experimental Raman shift values. The amplitudes of the lines are normalized to that of nitrogen (N2) at the unit value for a partial pressure of 1 atm and room temperature. The amplitudes of the lines are proportional to the quantity of each gaseous component in the medium 10.

[0105] In this embodiment, the full width at half maximum or FWHM of each line is typically 6 nm. Under these conditions, it is the surface, or the integral, of the line which is proportional to the quantity of the corresponding gaseous component.

[0106] In the particular case of hydrogen, several rotational lines So to S3 correspond to the presence of hydrogen. Among the four lines So to S3 of hydrogen, the line Si is the most intense of all. Moreover, in practice, the line So is not observed because it is contained in the rejection band of the filter 96 ([Fig.3]). Here, the wavelength X h2 is taken equal to the wavelength at which the line Si of hydrogen appears. In this case, the value of the optical signal scattered by the Raman effect by the hydrogen is taken equal to the surface of the line Sp

[0107] The two carbon dioxide lines overlap but can still be distinguished. The value of the optical signal for carbon dioxide is equal to the area of ​​these two lines.

[0108] In addition to the wavelengths of interest, between the So and Si lines of hydrogen, a first reference wavelength Xrefi is located. Between the Si and S2 lines of hydrogen, a second reference wavelength Xref2 is located. As explained later, the wavelengths Xrefi and Xref2 make it possible to extract the value of the optical signal scattered by the Raman effect by hydrogen using the baseline subtraction (BLS) method.

[0109] To obtain measurements for each of the wavelengths of interest and for the Xrefi and Xref2 lengths, unit 34 includes a circuit 51 for demultiplexing these different wavelengths and for each of the demultiplexed wavelengths, only one acquisition chain.

[0110] The proximal end of the second fiber 44 is connected to an input of the circuit 51. The circuit 51 has as many outputs as there are wavelengths to be processed. On each of these outputs, the circuit 51 delivers the optical signal corresponding to the wavelength to be processed. A respective acquisition chain is connected to each of these outputs. Here, these acquisition chains are structurally identical to each other and only one acquisition chain 50 for a wavelength X is described and shown in [Fig. 1]. The demultiplexing circuit 51 is for example produced as described in application US7385692.

[0111] The chain 50 successively comprises a band-pass filter 52, a photodetector 54, an amplification-discrimination module 56 and a counter 58.

[0112] The transfer function of the bandpass filter 52 has the form of a Gaussian centered on the wavelength X. The width of this transfer function at mid-height is comprised:

[0113] - between 10 nm and 12 nm for the wavelengths X02, XCo, XN2 and XH2o,

[0114] - between 5 nm and 6 nm for the wavelengths XH2, Xrefi and Xref2, and

[0115] - between 15 nm and 18 nm for the XCo2 wavelength in order to take into account the fact that carbon dioxide corresponds to two overlapping lines.

[0116] The photodetector 54 generates a voltage pulse for each photon received in the optical signal filtered by the bandpass filter 52. Preferably, the photodetector 54 is a photomultiplier. For example, the photomultiplier is equipped with a photocathode extended in the visible red and near infrared range made of GaAs-Cs or InGaAs-Cs. This photomultiplier operates at a temperature of -20°C, by cooling by the Peltier effect (thermoelectric effect) for example.

[0117] The amplification-discrimination module 56 first amplifies the voltage pulses at the output of the photodetector 54. Then, the amplification-discrimination module 56 eliminates most of the pulses coming from the background noise. For this, typically, the amplification-discrimination module 56 eliminates the pulses whose amplitude is lower than a predetermined threshold.

[0118] The counter 58 is incremented by each pulse delivered at the output of the amplification-discrimination module 56. The counter 58 therefore counts the number of pulses received during an interval of a predetermined duration Ata. At the end of the interval, the value of the counter 58 is reset to zero and the counter starts counting the pulses received during the following interval again. For example, the duration Ata is typically of the order of 10 ms, i.e. an acquisition rate of 100 Hz.

[0119] At the end of each interval, a measurement of the number of counts received during this interval is therefore delivered. This measurement is expressed in cps (counts per second).

[0120] The computer 36 typically comprises a programmable microprocessor 60 and a memory 62 containing the data and instructions necessary for executing the method of [Fig.5]. The computer 36 is in particular capable of controlling the laser source 32, the probe 20 and the human-machine interface 38. The computer 36 also acquires the measurements delivered by the unit 34 and the sensors 22 and 24. The computer 36 integrates all the measurements delivered during a predetermined acquisition window to obtain the value S(X) measured during this acquisition time window and expressed in number of counts.

[0121] The human-machine interface 38 typically comprises a screen 66.

[0122] Here, the power source 40 is a backed-up power source capable of powering all of the components of the processing unit 30, the probe 20 and the sensors 22 and 24 even in the event of a power outage. For this purpose, the source 40 comprises a battery capable of storing sufficient energy to power the processing unit 30, the probe 20 and the sensors 22 and 24 for at least one day and, preferably, for at least three consecutive days.

[0123] [Fig. 3] shows the probe 20 in more detail. The probe 20 comprises a housing 80 inside which are housed the various optical components necessary to produce and collect the optical signal diffused, by the Raman effect, by the gaseous components to be quantified, located in the chamber 94. The distal ends of the optical fibers 42 and 44 are located inside the housing 80.

[0124] This housing 80 is waterproof and dustproof to protect the optical components. For example, the housing 80 has a protection rating equal to IP69. Inside the housing 80, the optical components are immersed in a neutral gas introduced inside this housing via a purge valve 81. The neutral gas is a gas different from the gaseous components to be quantified. In addition, the neutral gas does not exhibit Raman scattering when excited by the excitation beam or exhibits Raman scattering at a wavelength different from the wavelengths of interest. For example, the neutral gas is argon. The pressure of the neutral gas inside the housing 80 is typically close to atmospheric pressure. For example, the pressure of the neutral gas is here equal to 1 atm (101.32 kPa).

[0125] Internal walls of the housing 80 are covered with a textured material that traps light and minimizes reflections of the laser beam on these internal walls.

[0126] Inside the housing 80, the probe 20 comprises the following components arranged one after the other along the path of the excitation beam and the optical signal from the first fiber 42 to the second fiber 44: a collimator 82, a polarizer ... rizer 84, a filter 86, a beam splitter 88, a parabolic mirror 90, a window 92, an analysis chamber 94, a filter 96, a polarizer 98, a rotator 100 and a parabolic mirror 102. In [Fig.3], the path of the excitation beam and the optical signal is represented by oriented arrows.

[0127] Here, inside the housing 80 the excitation beam and the optical signal move parallel to a predetermined PI plane fixed relative to the housing. For example, the PI plane is parallel to a base on which the various components of the probe 20 are fixed.

[0128] The collimator 82 makes it possible to obtain a collimated beam, directed towards the polarizer 84, from the excitation beam emitted by the distal end of the first fiber 42. For this purpose, for example, the collimator 82 is a parabolic mirror.

[0129] The polarizer 84 eliminates the polarization component perpendicular to the PI plane, to retain only the polarization component parallel to this PI plane. Thus, at the output of the polarizer 84, the polarization of the excitation beam is linear and parallel to the PI plane. Preferably, the polarizer 84 is a polarizer cube or PBS (“P ola-rization BeamSplitter cube”) which simultaneously exhibits good resistance to the laser beam but also low reported fluorescence.

[0130] The filter 86 eliminates or reduces the fluorescence generated by the polarizer 84 and the first fiber 42 as well as the components of the polarized beam greater than the wavelength Xp. More precisely, the filter 86 eliminates or greatly reduces all the components of the polarized beam likely to interfere with the optical signal scattered by the Raman effect. For this, the filter 86 allows the polarized excitation beam to pass while eliminating the wavelengths greater than the wavelength Xi of the excitation beam. For this purpose, the filter 86 is a low-pass or band-pass filter whose cut-off wavelength at -70 dB is, preferably, between Xi + 15 nm and Xi + 25 nm at normal incidence. For example, the filter 86 is a SWP filter (“Short Wave Pass filter”) mounted on an adjustable mount which makes it possible to adjust the angle of incidence of the laser beam relative to the normal to the surface of the filter 86.This angle of incidence is adjusted to obtain the desired rejection performance on the Raman spectrum. For example, filter 86 is a bandpass filter centered at 766 nm. Its cutoff wavelength at -70 dB is equal to 787 nm at normal incidence. It is oriented so that the angle of incidence is approximately 22°, which allows a rejection of -70 dB at 773 nm to be obtained. Thus, filter 86 allows the laser beam to pass at 750 nm, ensuring rejection as close as possible to be able to observe the SI rotational line of hydrogen.

[0131] The splitter blade 88 folds the excitation beam onto the parabolic mirror 90. For the optical signal reflected by the mirror 90 and which passes through the splitter blade 88, the splitter blade 88 also behaves like a high-pass filter which eliminates or reduces the components of the optical signal passing through it whose wavelengths are equal to or less than the wavelength Xb. For this purpose, typically, the blade 88 has a cut-off wavelength at -70 dB, preferably between Xi + 5 nm and Xi + 22 nm. For this purpose, here, the blade 88 is also mounted on an orientable mount which makes it possible to adjust the angle of incidence of the incident optical signal relative to the normal to the surface of the blade 88. This angle of incidence is adjusted to obtain the desired rejection performance on the Raman spectrum. For example, the cut-off wavelength at -70 dB of the blade 88 is equal to 776 nm at normal incidence. The blade is oriented so that the angle of incidence is approximately 20° which makes it possible to obtain a rejection of -70 dB at 758 nm.

[0132] The parabolic mirror 90 focuses the polarized beam at a point R located inside the analysis chamber 94.

[0133] The chamber 94 contains the sample of the medium 10 to be analyzed. The chamber 94 isolates the sample from the light which may be present inside the enclosure 4. In addition, the chamber 94 is also designed to protect the sample from possible water splashes as well as pollution by aerosols.

[0134] Here, the chamber 94 is crossed by a gas flow of the medium 10 so that the sample of the medium 10 is constantly renewed. This gas flow of the medium 10 occurs, for example, by natural convection. For this, the chamber 94 has vents for the entry and exit of the medium 10. The internal volume of the chamber 94 is for example 500 cm3.

[0135] The chamber 94 is fluidically isolated from the interior of the housing 80. For this purpose, the housing 80 comprises the window 92 which is transparent to the excitation beam and to the optical signal scattered, by Raman effect, by the gaseous components to be quantified. Here, the window 92 is configured to transmit the polarized excitation beam in full without reflecting a part of it. For this purpose, the window is here a Brewster window, that is to say a window inclined at the Brewster angle relative to the optical axis of the incident excitation beam so as not to reflect the incident beam.

[0136] At least the inner wall of the chamber 94, located opposite the window 92 and on the other side of the point R, is covered with a textured material which traps the light and minimizes the reflection and diffusion of the excitation beam on this inner wall. Preferably, this inner wall of the chamber 94 is placed far enough away to limit the scattered spectral background. For example, the distance between the window 92 and this inner wall of the chamber 94 is greater than or equal to 5 cm.

[0137] In this embodiment, the power of the polarized excitation beam at point R is approximately 180 mW for a power emitted by the laser source 32 of approximately 1 Watt.

[0138] Here, the temperature sensor 24 is placed inside the chamber 94 as close as possible from point R. In addition, preferably, the chamber 94 also comprises an anti-condensation module 95 which makes it possible to prevent water vapor from condensing on the window 92. This anti-condensation module 95 is supplied via the cables 46. However, to simplify [Fig.2], the connection of this anti-condensation module 95 to the cables 46 is not shown.

[0139] The optical signal diffused, by Raman effect, by the interaction, at point R, of the excitation beam with the gaseous components present, is returned to the interior of the housing 80 and collected by the mirror 90. The mirror 90 then returns the optical signal to the separating plate 88 which transmits it to the filter 96.

[0140] The filter 96 is a high-pass filter or LWP filter (“Long-Wave Pass”) which also has a cut-off wavelength at -70 dB preferably, between Xi + 5 nm and Xi + 22 nm. For this purpose, here, the filter 96 is also mounted on an orientable mount which makes it possible to adjust the angle of incidence of the incident optical signal relative to the normal to the surface of the filter 96. This angle of incidence is adjusted to obtain the desired rejection performance on the Raman spectrum. For example, the cut-off wavelength at -70 dB of the filter 96 is equal to 776 nm at normal incidence. The filter 96 is then oriented so that the angle of incidence is approximately 20° which makes it possible to obtain a rejection of -70 dB at 758 nm.

[0141] The optical signal filtered by filter 96 is then polarized by polarizer 98. For example, polarizer 98 is a PBS (“Polarizing Beam Splitter”) polarizer similar to polarizer 84.

[0142] The electromagnetic rotator 100 makes it possible to rotate the polarizer 98 by 90° in response to a control signal received via the cables 46. For example, the rotator 100 comprises a shaft on which a set 97 of gears is mounted. This set 97 of gears comprises, for example, a first toothed wheel which rotates a second toothed wheel on which the polarizer 98 is mounted. The rotator 100 is powered via the cables 46. When this rotator 100 is powered, this triggers a rotation of its shaft which rotates the polarizer 98 by 90° around its optical axis. When the power is cut off, the rotator 100 returns to its initial position. Here, in the absence of power, the passing axis of the polarizer 98 is parallel to the plane PI so that the power of the optical signal passing through it is maximum.Conversely, when the rotator 100 is powered, the passing axis of the polarizer 98 is perpendicular to the PI plane and the power of the optical signal passing through the polarizer 98 is minimal. Subsequently, when the passing axis of the polarizer 98 is parallel to the PI plane, the optical signal is said to be measured with “parallel polarization”. Conversely, when the passing axis of the polarizer 98 is perpendicular to the PI plane, the optical signal is said to be measured with “perpendicular polarization”. The electronics which make it possible to power the rotator 100 and, alternately, to . cutting off its power supply is located in the processing unit 30 and has not been shown to simplify [Fig.l].

[0143] To rotate its shaft, the rotator 100 comprises an electromagnetic actuator comprising coils and magnets. In order for it to be able to operate correctly even in the event of a nuclear accident, typically, the electrical wires of the coils are coated with silicone. Similarly, the permanent magnets are made of magnetic materials which have a Curie temperature greater than 300°C or 450°C. For this purpose, for example, magnets made of Sm-Co (samarium-cobalt) or Al-Ni-Co (Aluminium-Nickel-Cobalt) are recommended.

[0144] The parabolic mirror 102 focuses the optical signal which has passed through the polarizer 98 onto the distal end of the fiber 44. The optical signal scattered by the Raman effect is therefore guided by the fiber 44 to the reading unit 34.

[0145] [Fig.4] represents four Raman spectra 120, 122, 124 and 126 of air, measured using a device almost identical to device 2 in which the reading unit 34 is replaced by a spectrometer equipped with a CCD camera which makes it possible to measure the amplitude, in cps (counts per second), of the optical signal collected for a very large number of wavelengths distributed between 750 nm and 1050 nm. In [Fig.4], a solid line curve corresponds to the Raman spectrum of an optical signal measured with parallel polarization. A dotted line curve corresponds to the Raman spectrum of an optical signal measured with perpendicular polarization. Spectra 120 and 122 were recorded by exposing a 10 m length of fiber 44 to photon radiation with an energy equal to 1.25 MeV and a dose rate equal to 250 Gy / h. Both spectra 124 and 126 were recorded under the same conditions except that the dose rate was equal to 1016 Gy / h.

[0146] In this [Fig.4], the lines, produced by the Raman effect, corresponding to oxygen and nitrogen are identified by the symbols, respectively, “02” and “N2”.

[0147] First of all, it can be observed that the "O2" and "N2" lines only exist in optical signals measured with parallel polarization. This is because the Raman effect that produces these lines has a very low depolarization ratio p for oxygen and nitrogen. The depolarization ratio p is the ratio between the amplitude of the signal measured with perpendicular polarization and the amplitude of the signal measured with parallel polarization. Thus, for oxygen and nitrogen, the Raman effect is only visible on the optical signal whose polarization direction is identical to the polarization direction of the excitation beam. This is also true for carbon dioxide, carbon monoxide, and water vapor. However, this is false for the Si line of hydrogen. In the case of hydrogen, the Si line exists in the Raman spectra of optical signals measured with parallel polarization and with perpendicular polarization.However, the amplitude of the Si line in the Raman spectrum of the . optical signal measured with parallel polarization is about 25% larger than the amplitude of this same Si line in the Raman spectrum of the optical signal measured with perpendicular polarization. In other words, the p depolarization ratio of the Si line of hydrogen is about 75%.

[0148] Cherenkov radiation is radiation that appears in optical fibers and optical components when they are exposed to high-energy photon or electron radiation. In silica, the threshold for the appearance of this radiation is located at 180 keV. Cherenkov radiation can increase substantially in the event of a nuclear accident. As illustrated by spectrum 120, for a fiber length of 10 meters exposed under a dose rate equal to 250 Gy / h, the amplitude of the Cherenkov radiation exceeds the amplitude of the "O2" line by a factor greater than 50 and the amplitude of the "N2" line by a factor greater than 20. In the case of spectrum 124, the amplitude of the Cherenkov radiation exceeds the amplitude of the "O2" line by a factor greater than 180 and the amplitude of the "N2" line by a factor greater than 60.This shows that during operation of the device 4, the optical Raman scattering signal is drowned in a masking signal corresponding, to a significant extent, to the Cherenkov radiation which appears mainly in the silica and therefore in the fiber 44 as well as, to a lesser extent, in the optical components of the probe 20.

[0149] Another part of this masking signal originates from:

[0150] - the electronic noise of the acquisition chain 50 known by the English term of "dark noise", and

[0151] - a fluorescence signal from optical components.

[0152] The fluorescence signal of the optical components is typically caused by the optical components of the device 2 when illuminated by the excitation beam.

[0153] The operation of the device 4 will now be described with reference to the method of [Fig. 5]. In this embodiment, the quantity of hydrogen present in the sample is determined according to the known method of baseline subtraction or SLB method. Thus, the quantity of hydrogen is determined from the values ​​S3(XH2), S3(Xrefi) and S3(Xref2) of a signal S3. The signal S3 is the optical signal measured according to the parallel polarization while the laser source 32 is switched on. More precisely, the SLB method consists of defining a spectral function f representing the spectral background on which the Raman peaks appear. This can be linear or polynomial, for example, of order two. Here, for illustration, this function f is defined by one of the following relations: f(X) = a + b*(^ - kcte)2 and f(X) = a + b*(^ - kcte ), where a, b and Xcte are calibration parameters.The last parameter Xcte is an optimization parameter, chosen a priori in the Raman spectral band. The parameters a and b are deduced from the measurements S3(Xrefi) and 83()½^) carried out outside the . each Raman line by solving the following system of equations: f(Xrefi) = S3(Xrefi) and f(Xref2) = S3(Xref2). The useful signal r(XH2) is then deduced by subtraction of the spectral function, i.e. using the following relation: r(XH2) = [S3(XH2) - f(XH2)] / t3, where t3 is the acquisition time of the signal S3.

[0154] Subsequently, only the quantification of the other gaseous components is described in detail.

[0155] To extract the optical signal scattered by Raman effect from the masking signal, here, the cal culator 36 determines the quantity of a gaseous component corresponding to the wavelength of interest X using the following relation (1): r(X) = [S3(X) / t3 - Sl(X) / tl] -k(X)*[S4(X) / t4 - S2(X) / t2], where:

[0156] - r(X) is a value representative of the quantity of the gaseous component, which corresponds to the Raman line centered on the wavelength X, present in the sample inside chamber 94, expressed in counts per second,

[0157] - S1(X) and S2(X) are the values, for wavelength X, respectively, of optical signals S1 and S2 measured when the laser source 32 is switched off,

[0158] - S3(X) and S4(X) are the values, for wavelength X, respectively, of optical signals S3 and S4 measured when the laser source 32 is switched on,

[0159] - k(X) is the value, for the wavelength X, of a scale coefficient k pre determined, and

[0160] - tl, t2, t3 and t4 are the durations of the time windows during which the signals SI, S2, S3 and S4 are measured.

[0161] The values ​​S1(X), S2(X), S3(X) and S4(X) are expressed in number of moves.

[0162] The signals SI and S3 are measured with parallel polarization during the durations, respectively, tl and t3. The signals S2 and S4 are measured with perpendicular polarization for the durations, respectively, t2 and t4. The ratios Sl(X) / tl, S2(X) / t2, S3(X) / t3 and S4(X) / t4 are therefore measurements, in counts per second, of the optical signal scattered by the Raman effect at the wavelength X.

[0163] Since the signals S1 and S2 are measured while the laser source 32 is off, these signals represent measurements of the masking signal, respectively, with parallel polarization and with perpendicular polarization. Under these conditions, the first difference S3(X) / t3 - S1(X) / t1 makes it possible to obtain an optical signal corrected for Cherenkov radiation. However, this first difference is not corrected for the fluorescence signal coming from the optical components. This fluorescence signal is proportional to the laser emission but essentially depolarized. Indeed, this fluorescence signal is absent from the signal S1 because the signal S1 is measured when the laser source 32 is off. Here, this fluorescence signal is estimated from the second difference S4(X) / t4 - S2(X) / t2. More precisely, this second difference estimates the amplitude of the fluorescence signal when the optical signals are measured with perpendicular polarization.In this embodiment, the fluorescence signal . present in the signal S3 measured with parallel polarization is assumed to be proportional to the fluorescence signal present in the signal S4. Thus, the fluorescence signal present in the signal S3 is estimated by the term k(X)*[S4(X) / t4 - S2(X) / t2] of the relation (1).

[0164] The value of the coefficient k for all wavelengths of interest must first be determined before relation (1) can be used. Thus, the method begins with a calibration phase 200. Here, phase 200 is executed while the probe 20 is fixed inside the enclosure 4. In this case, preferably, phase 200 is executed at a time when the nuclear reactor is shut down so that the Cherenkov radiation is of very low magnitude, or even unmeasurable. For example, phase 200 is executed on the occasion of a ten-year shutdown of the nuclear reactor.

[0165] During phase 200, a reference sample, the composition of which is known, is located inside the chamber 94. Here, the reference sample is air.

[0166] Finally, during phase 200, the processing unit 30 is replaced by a processing unit which makes it possible to measure the amplitude of the optical signal both at the wavelengths of interest and at other wavelengths. For example, for this purpose the reading unit 34 is replaced by a spectrometer identical to that used to measure the Raman spectra of [Fig.4].

[0167] Here, the variation of the coefficient k as a function of the wavelength X is assumed to be linear over the entire range of interest, i.e. here between 750 nm and 1050 nm. Under these conditions, the values ​​of the coefficients of the line which approximates the evolution of the value k(X) as a function of the wavelength X, are obtained from measurements carried out for wavelengths X; located outside the oxygen and nitrogen lines and therefore far from the wavelengths X02 and XN2. The index i is an identifier of the wavelength.

[0168] Therefore, during a step 202, for each wavelength X;:

[0169] - a Spara(X;) value of an optical signal Spara is measured with a polarization parallel, and

[0170] - a Sperp(X;) value of an optical signal Sperp is measured with a polarization perpendicular.

[0171] The Spara and Sperp signals are measured while the laser source 32 is switched on and therefore when the sample present inside the chamber 94 is excited by the excitation beam.

[0172] Furthermore, in this embodiment, during a step 204, the electronic noise of each acquisition chain 50 of the reading unit 34 is measured. Here, this electronic noise is assumed to be identical for each acquisition chain 50. Furthermore, it is assumed that this electronic noise does not vary as a function of the wavelength X within the range of interest. Finally, it is also assumed that this electronic noise is independent of the polarization of the measured optical signal. For example, the amplitude DN of this electronic noise is measured by occluding the distal end of the fiber 44 then averaging the amplitudes measured by each acquisition chain 50.

[0173] During a step 206, for each wavelength Xi5 a value k(X;) is calculated using the following relation: k(X;) = [Spara(X;) - DN] / [Sperp(X;) - DN].

[0174] Then, during a step 208, the values ​​Id / .cmk k(XO2), k(XCo), k(XN2) and k(XH20) are calculated, by interpolation, from the values ​​k(X;). For this purpose, the function which associates with each wavelength X, the corresponding value k(X) of the coefficient k, is obtained by mathematical adjustment. For example, here, the interpolation used is a linear interpolation. For this purpose, the coefficients a and b of the straight line which minimizes the deviations with the different values ​​k(X;), are calculated. Then, using the equation of this straight line, the values ​​IdXcmk k(X02), k(XCo), k(XN2) and k(XH20) are calculated.

[0175] During a step 210, the values ​​IdXcmk k(XO2), k(XCo), k(XN2) and k(XH20) are recorded in the memory 62 of the computer 36.

[0176] The calibration phase 200 is then completed and a phase 220 of operation of the device 2 begins. Phase 220 is typically executed during a nuclear accident. However, it can also be executed while the nuclear reactor is operating normally, i.e. in an operational situation.

[0177] Phase 220 consists of a measurement cycle repeated throughout the duration of phase 220. In [Fig.6], the different measurement cycles are identified by the symbol Q, where the index j is the order number of the cycle since the start of phase 220. To simplify [Fig.6], only cycles Ci and C2 are fully represented.

[0178] Each cycle Q comprises a period T0FF and a period T0N which, preferably, are immediately consecutive. During the period T0FF, the computer 36 controls the laser source 32 so that it is switched off for the entire duration A0FF of this period T0FF.

[0179] The period T0FF comprises two time windows designated in Figure 6, respectively, by the symbols C / / and C ± . Preferably, these two windows C / / and C ± follow each other immediately. Here, the sum of the durations t1 and t2 of the windows, respectively C / / and C ± , is equal to the duration A0FF.

[0180] Throughout the duration tl of the window C / / , the computer 36 maintains the passing axis of the polarizer 98 parallel to the plane PI. For this, the computer 36 does not supply the rotator 100. At the same time, the computer acquires the values ​​S l(XC02h S 1(XO2), Sl(Xco), S1(XN2) and Sl(XH2O).

[0181] Throughout the duration t2 of the window C ± I, the computer 36 maintains the passing axis of the polarizer 98 perpendicular to the plane PI. For this, the computer 36 supplies the rotator 100 via the cables 46. At the same time, the computer 36 acquires the values ​​S2(XCo2), S2(XO2), S2(XCo), S2(XN2) and S2(XH2o)-

[0182] The period T0N comprises two time windows designated in Figure 6, respectively, by the symbols R / / and R -L . Preferably, these two windows R / / and R -L follow each other immediately. Here, the sum of the durations t3 and t4 of the windows, respectively R / / and R -L , is equal to the duration A0N of the period T0N.

[0183] Preferably, the ratio t3 / t4 is equal to the ratio tl / t2. In addition, to simplify relation (1), here, the durations t3 and t4 are chosen to additionally verify the following condition: tl+t2 = t3+t4, i.e. the durations A0FF and A0N are equal. For this, here, all the durations tl, t2, t3 and t4 are equal. Under these conditions, relation (1) is written in the following simplified form: R(X) = S3(X) - S1(X) - k(X)*(S4(X) - S2(X)), where R(X) is the amplitude of the optical signal scattered by the Raman effect at wavelength X, expressed in number of counts.

[0184] Advantageously, the durations A0FF and A0N are less than 10 s or 5 s. Thus, over the duration of a cycle, even in the event of a nuclear accident, the radiation-induced attenuation does not substantially modify the attenuation of the optical signal in the fibers and optical components. In addition, to obtain good precision, the durations A0FF and A0N are chosen to be greater than 0.01 s and, generally, greater than 1 s.

[0185] Throughout the duration t3 of the window R / / , the computer 36 maintains the passing axis of the polarizer 98 parallel to the plane PI. For this, the computer 36 does not supply the rotator 100. At the same time, the computer 36 acquires the values ​​S3( / <02h S3(XO2), S3(Xco), S3(XN2) and S3(Xh2O).

[0186] Throughout the duration t4 of the R-L window, the computer 36 maintains the passing axis of the polarizer 98 perpendicular to the PI plane. For this, the computer 36 supplies the rotator 100 via the cables 46. At the same time, the computer 36 acquires the values ​​S4(XCO2), S4(XO2), S4(XC0), S4(XN2) and S4(XH2O).

[0187] Then, during a step 230, the computer 36 determines the quantity of each gaseous component. For this, the computer 36 calculates, during an operation 232, using the relation (1) and for each wavelength / <02- An- / <0- XN2 and XH20, the values ​​r(XC02), r(XO2), r(XC0), r(XN2) and r(XH20). The values ​​ri / <02 hr(XO2), r(XC0), r(XN2) and r(XH2o) are expressed in counts per second (cps). These values ​​ri / <02)- r(X02), r(XCo), r(X N2) and r(XH20) are proportional to the quantities, respectively, of carbon dioxide, oxygen, carbon monoxide, nitrogen and water vapor present in the sample.

[0188] Then, during an operation 234, from these values ​​r(XC02), r(X02), r(XCo), r(XN2) and r(XH2o) and from the value r(XH2) obtained, the calculator 36 establishes the proportions, in percent, of each of these gaseous components in the medium 10 taking into account, for example, the Raman interaction sections of each of the gaseous components. The Raman interaction section is defined in the article Magne2020 and corresponds to the term called “Raman differential cross section” in this article. Here, the sum of the partial pressures of all gaseous components to be quantified (02, N2, H20, H2, CO and C02) is close to the total pressure PT in containment enclosure 4. The temperature of each probe is not involved since only a relative indication is sought.

[0189] The proportion, in percent, of the component] is then obtained, for example, using the following self-normalization relationship:

[0190] [Math.l]

[0191] where:

[0192] - P, / PT is the proportion, in chamber 94, of the component corresponding to the line Raman at wavelength X,

[0193] - P, is the partial pressure of the component corresponding to the Raman line at the wavelength X,

[0194] - PT is the total pressure inside the enclosure 4 measured by the sensor of pressure 22,

[0195] - Ox is a constant, for example determined experimentally, defined by the following relation: r(X)*T = o^ * P^, where T is the temperature measured in probe 20,

[0196] - Ne is the total number of quantified components, i.e. six in this example,

[0197] - r(Xj) is the measurement of the quantity of the j-th component to be quantified carried out using of probe 20, and

[0198] - o^j is a constant, for example determined experimentally, defined by the following relation: r(Xj)*T = o^ * Py, where P^ is the partial pressure of the j-th component corresponding to the Raman line at wavelength Xj.-,

[0199] The relation r(X)*T = o^ * P, follows from relation (5) of the article Magne2020. The constant o^ is a function of the Raman interaction section of the component corresponding to the Raman line at wavelength X.

[0200] Furthermore, as demonstrated in particular by relation (5) of the article Magne2020, when these values ​​r(XC02), r(XO2), r(XCo), r(XN2) and r(XH2o) are multiplied by the temperature, each of these products is proportional to the partial pressure of the corresponding gaseous component. Thus, if the temperature T and the constant o^ are known, the partial pressure P, can be calculated using the relation Px=r(X)*T / Ov Otherwise, in an embodiment like the one described here where the sum of the partial pressures of the quantified gaseous components is close to or equal to the total pressure PT, then if in addition the total pressure inside the enclosure 4 is measured, the calculator 36 can then calculate the partial pressure of each of the gaseous components in mul folding the ratio P, / P| by the total pressure PT measured by the sensor 22. Preferably, during an operation 236, the computer 36 acquires the temperature measured by the sensor 24 and the total pressure PT measured by the sensor 22 then calculates the partial pressure of hydrogen, carbon dioxide, oxygen, carbon monoxide and nitrogen from the values ​​r(XH2), r(XCo2), r(X02), r(XCo) and r(XN2). Then, the partial pressure P(XH20) of the water vapor is taken equal to the difference between the total pressure PT minus the sum of the partial pressures of hydrogen, carbon dioxide, oxygen, carbon monoxide and nitrogen. In this case, the acquisition chain 50 used to process the optical signal at the wavelength XH20 can be omitted.

[0201] Thus, at the end of each cycle Q, a new quantification of the gaseous components is available. For example, during a step 238, the computer 36 controls the human-machine interface 38 to display these quantifications on the screen 66.

[0202] Furthermore, during a step 250, the computer 36 calculates a more precise value of the quantity of each gaseous component by averaging the quantities determined during several successive cycles Cj. Typically, for this, it uses only the quantities determined during the cycles contained within a sliding time window. The duration of this sliding time window is preferably less than 10 min and, generally, greater than 1 min or 5 min. The more precise quantities are also displayed on the screen 66.

[0203] Chapter III: Variants:

[0204] Probe variants:

[0205] Other solutions are possible for focusing the laser beam on the point R. For example, instead of using the parabolic mirror 90 for this, it is possible to use an achromatic doublet of converging lenses arranged so that the different wavelengths are focused on the same point R.

[0206] Alternatively, the filter 96 is omitted. In this case, the filtering function is provided solely by the blade 88.

[0207] Other embodiments of the polarizer 84 or 98 are possible. For example, the polarizer 84 or 98 is a glass polarizer (sodium-silicate) doped with ellipsoidal nanoparticles having a strong polarizing effect.

[0208] Alternatively, the polarizer 98 is replaced by a polarization splitter cube. This splitter cube separates the parallel and perpendicular polarization directions. Thus, the splitter cube simultaneously generates the signals S1 and S2 and then simultaneously the signals S3 and S4. In this case, the fiber 44 is replaced by a first optical fiber and a second optical fiber. The first and second fibers guide the measured signals with, respectively, the parallel and perpendicular polarizations, to the reading unit. The reading unit 34 is then modified to be able to measure, simultaneously, the signals SI and S2 and the signals S3 and S4.

[0209] In a simplified embodiment, the textured material is omitted.

[0210] Processing unit variants:

[0211] In another embodiment, the source 32 is optically connected to the probe 20 via an optical switch which directs the excitation beam towards the probe 20 and, alternately, towards another probe. In this case, the source 32 continuously emits the excitation beam and, during the period T0N, the switch directs the excitation beam towards the probe 20. During the period T0FF of the probe 20, the switch directs the excitation beam towards the other probe and no longer towards the probe 20.

[0212] Alternatively, a CCD (“Charge Coupled Device”) sensor is used instead of the reading unit 34. Typically, such a CCD sensor is capable of measuring the optical signal at more than 300 or 1000 or 2048 different wavelengths regularly distributed within the range of interest. Such a CCD sensor therefore measures both the optical signal received at each wavelength X of interest and also at other wavelengths of no interest.

[0213] The photomultiplier 54 can be replaced by a photodiode such as an avalanche photodiode or a SiPM (“Silicon-PhotoMultiplier”) detector.

[0214] Depending on the context of use, the processing unit may be configured to measure the quantity of one or more additional gaseous components other than those from the group consisting of water vapor, oxygen, nitrogen, carbon monoxide, carbon dioxide and hydrogen. In this case, the processing unit is configured to use an additional wavelength of interest corresponding to each additional gaseous component to be quantified. Conversely, the processing unit may also be configured to measure the quantity of only one or only part of the gaseous components from the group consisting of water vapor, oxygen, nitrogen, carbon monoxide, carbon dioxide and hydrogen. In this case, typically, the number of acquisition chains 50 is reduced.

[0215] Variants of the quantification method:

[0216] Other approximations of the evolution of the value of the coefficient k as a function of the wavelength X are possible. For example, as a variant, the evolution of the value of the coefficient k as a function of the wavelength X is approximated using a polynomial of order greater than one.

[0217] In another embodiment, the value of the coefficient k also varies as a function of the measured radiation dose. In this case, the quantification device is also capable of measuring the radiation dose received by the sample of the gaseous medium. For example, this radiation dose is estimated from the intensity of the Cherenkov radiation measured in the absence of the excitation beam.

[0218] In a simplified variant, the electronic noise of the acquisition chain is neglected. In this case, each value k(X;) is simply taken equal to the ratio Spara(X; ) / Sperp(Xi).

[0219] In a very simplified variant, the value of the coefficient k is assumed to be constant regardless of the wavelength X.

[0220] In a degraded embodiment, the calibration phase 200 is executed when the nuclear reactor is operating. In this case, the calibration phase is triggered during nominal operation of the nuclear reactor and the reference sample is identical to the sample of the gaseous medium 10.

[0221] In another variant, the calibration phase of the coefficient k is carried out during the manufacture of the device 4, that is to say at a time when the probe 20 is not installed inside the enclosure.

[0222] In another embodiment, the wavelength of the excitation beam is decreased so that the lowest vibrational line of hydrogen is also measurable using the reading unit 34. This lowest vibrational line is present in the Raman spectrum of the signal S3 and absent from the Raman spectrum of the signal S4. In this case, one of the wavelengths of interest used is the wavelength of this vibrational line of hydrogen and the amount of hydrogen is then measured using relation (1). In another variant, although the use of relation (1) leads to using only approximately 25% of the signal S3(XH2), the amount of hydrogen is still quantified using relation (1). In these last two cases, the reading unit 34 does not need to carry out measurements at the wavelengths Xref1 and ^ef2.

[0223] Other relationships than relationship (1) are possible. For example, as a variant, an additional correction term is added in this relationship (1). Typically, this additional correction term is constructed so as to eliminate or reduce another source of noise than Cherenkov radiation. For example, this additional term is a term which depends on the quantity of ambient light present in the chamber 94 if it is not totally eliminated by the walls of the chamber 94.

[0224] Alternatively, the ratios tl / t2 and t3 / t4 are different from one. The ratios tl / t2 and t3 / t4 are also not necessarily equal.

[0225] The ratio A0FF / A0N is also not necessarily equal to one. For example, the duration Aoff is greater than the duration A0N or vice versa.

[0226] What has been described so far also works if the durations A0FF and A0N are greater than 10 s or 1 min. However, preferably, particularly in the presence of a nuclear accident, these durations A0FF and A0N are less than 5 min.

[0227] The calculation of a more precise value of the quantity of the gaseous component by averaging quantities determined at the end of several measurement cycles, can to be omitted.

[0228] Alternatively, during operation 236, the partial pressure of water vapor is calculated from the value r(XH20).

[0229] In another embodiment, the ratio P(XH2o) / PsatH2O is also determined, where P(X h2o) is the partial pressure of water vapor determined during operation 236 and Psat H20 is the saturation pressure of water vapor determined from the temperature measured by the temperature sensor 24. To determine the saturation pressure PsatH20, an abacus associating with several temperatures, the corresponding saturation vapor pressure PsatH20 is, for example, pre-recorded in the memory 62.

[0230] The signals SI and S2 are representative of the amplitude of the Cherenkov radiation. The Cherenkov radiation is representative of the average dose rate on the fiber section 44. Thus, the signals SI and S2 can also be used to estimate this average dose rate and therefore estimate the average radiation dose at different locations, inside the containment enclosure.

[0231] The acquisition frequency of the optical signal may be greater than 100 Hz.

[0232] Alternatively, only a relative amount or only an absolute amount of the gaseous component is measured.

[0233] Other variants:

[0234] The continuous laser source can be replaced by a pulsed laser source.

[0235] If only a measurement of the relative amount of the gaseous components is required, then the temperature sensor 24 and the pressure sensor 22 may be omitted. In the latter case, the gaseous component taken as the reference gaseous component is often nitrogen.

[0236] In practice, the gaseous medium contained inside the containment vessel of a nuclear reactor does not always have a homogeneous composition. Thus, it is often necessary to quantify the gaseous components at several different locations inside the vessel 4. For this, a probe 20 is installed at each of these locations. In this case, the processing unit 30 is connected to all of these probes 20 in order to be able to quantify the gaseous components present at each of these locations.

[0237] The teaching given here applies to any containment enclosure inside which a nuclear reactor is housed and not only to the case where the enclosure is that formed by a reactor building of a nuclear power plant which produces electricity.

[0238] Several of the variants described above can be combined in the same embodiment.

[0239] Chapter IV: Advantages of the embodiments described:

[0240] Determining the quantity of a gaseous component present inside the enclosure 4 from the difference S3(X) / t3-S 1 (X) / t 1 makes it possible to reduce the noise caused by Cherenkov radiation. In addition, determining the amount of the additional gas component from the difference S4(X) / t4-S2(X) / t2 makes it possible to reduce the fluorescence signal caused by the optical components of the quantification device when illuminated by the excitation beam. Thus, by using these two differences, the accuracy of the quantification device is equal to or better than that of the device described in the Magne2020 article. Finally, to achieve this accuracy, the device 4 uses measurements of the optical signal carried out for a smaller number of wavelengths. In particular, for each gas component to be quantified, it is not necessary to measure the optical signal at the wavelength of interest and in addition at at least two reference wavelengths. Therefore, the reading unit can be considerably simplified and its cost greatly reduced.

[0241] Choosing the ratios tl / t2 and t3 / t4 equal increases the precision.

[0242] Adjusting the value of the coefficient k as a function of the wavelength X of interest helps improve the accuracy of the measurement process.

[0243] Constructing the values ​​k(X) of the coefficient k using the relation k(X) = [SPara(X)-DN] / [SPerp(X)-DN] makes it possible to obtain values ​​which are practically independent of the dose of radioactive radiation received by the probe. Thus, the operation of the quantification device remains precise even in the event of a nuclear accident.

[0244] Limiting the duration AOff and A0N to less than 10 s makes it possible to limit the variation, during each of the periods T0Ff and T0N, of the radiation-induced attenuation in the optical components and the fibers. Thus, the quantification device continues to be accurate even in the event of a nuclear accident.

[0245] Calculating a measurement of the quantity of the gaseous component by averaging the quantities determined during several measurement cycles close to each other makes it possible to limit the influence of the radiation-induced drift which occurs, in particular, in the event of a nuclear accident. Indeed, in the event of a nuclear accident, the ambient temperature and the radiation-induced attenuation or RIA (Radiation-Indue Attenuation) of the fibers increase rapidly.

[0246] Measuring the signals S1, S2, S3 and S4 at different wavelengths of interest makes it possible to simultaneously measure the quantity of several different gas components.

[0247] By simultaneously measuring the quantities of water vapor, oxygen and nitrogen, it is possible to assess the risk of an explosion occurring inside the enclosure.

[0248] Measuring the quantities of carbon monoxide and carbon dioxide makes it possible to trace the corium-concrete interaction (Molten Core Concrete Interaction -MCCI) which occurs after the vessel is breached in the event of a nuclear reactor core meltdown.

[0249] Measuring the amount of hydrogen using relation (1) makes it possible to carry out this measurement using only the two values ​​S3(XH2) and S4(XH2).

[0250] Using a continuous laser beam makes it possible, for the same size, to have a laser beam whose average power is much greater than if a pulsed laser beam were used. However, the greater the average power of the laser beam, the greater the precision of the measurement of the quantity of the gaseous component.

[0251] Calculating the partial pressure of water vapor from the difference between the total pressure PT and the partial pressures of hydrogen, carbon dioxide, oxygen, carbon monoxide and nitrogen makes it possible to improve the accuracy of the measurement of the partial pressure of water vapor.

[0252] Using a single photodetector per wavelength of interest simplifies the quantification device.

Claims

Claims

1. Method for quantifying, by Raman spectrometry, at least one gaseous component inside a containment enclosure of a nuclear reactor, this method comprising a first measurement period (TOn) during which: - a monochromatic excitation beam excites, inside the containment enclosure, a sample of the gaseous medium located inside this containment enclosure, - a reading unit measures the optical signal scattered, by Raman effect, by the sample excited by the excitation beam, and - an electronic calculator determines, from the optical signal measured by the reading unit, a quantity of the gaseous component present in the sample, characterized in that: - during the first measurement period (TON): - the excitation beam is linearly polarized in a predetermined direction, and - the reading unit measures (R / / , R -L ) the values ​​S3(X) and S4(X), respectively, of an optical signal S3 and an optical signal S4 scattered by the sample, where: - the optical signals S3 and S4 are the optical signals measured with polarization directions, respectively, parallel and perpendicular to the predetermined polarization direction of the excitation beam, when the sample is excited by the excitation beam, - X is a wavelength of interest, i.e. a wavelength at which, due to the Raman effect, a line appears in the spectrum of the scattered optical signal when the gaseous component to be quantified is present in the sample, and - the method also comprises a second measurement period (TOff) in the absence of the excitation beam during which the reading unit measures values ​​S1(X) and S2(X), respectively, of an optical signal SI and of an optical signal S2 scattered by the sample, where the optical signals SI and S2 are the optical signals measured with polarization directions, respectively, parallel and perpendicular to the predetermined polarization direction of the excitation beam used during the first measurement period, when the excitation beam is absent, and - the calculator determines (230) the quantity of the gaseous component present in the sample from the difference [(S3(X) / t3 - S1(X) / t1] and the difference [S4(X) / t4 - S2(X) / t2], where t1, t2, t3 and t4 are the durations of the time windows during which the signals SI, S2, S3 and S4 are measured to obtain, respectively, the values ​​S1(X), S2(X), S3(X) and S4(X) in number of counts.

2. Method according to claim 1, in which the calculator determines (230) the quantity of the gaseous component present in the sample using the following relation: r(X) = [S3(X) / t3 - Sl(X) / tl] -k(X)*[S4(X) / t4 - S2(X) / t2], where: - r(X) is a value representative of the quantity of the gaseous component present in the sample expressed in counts per second, - k(X) is the value, for the wavelength X, of a predetermined scale coefficient k.

3. The method of claim 2, wherein the ratios t1 / t2 and t3 / t4 are equal.

4. Method according to claim 2 or 3, in which the method comprises, prior to the execution of the first and second measurement periods, a calibration phase (200) during which: - the monochromatic excitation beam excites a reference sample, and - a sensor measures (202) an optical signal Spara and an optical signal Sperp scattered by the reference sample excited by the excitation beam, the optical signals Spara and Sperp being the optical signals scattered by the reference sample measured with directions of polarization, respectively, parallel and perpendicular to the predetermined direction of polarization of the excitation beam, and the signals Spara and Sperp being measured at wavelengths X;different from the wavelengths at which, due to the Raman effect, lines appear in the spectrum of the optical signal scattered by this reference sample, then - the value k(X) of the coefficient k for each wavelength of interest X is calculated (208) from the measured Spara and Sperp optical signals, then recorded in the calculator.;

5. Method according to claim 4, in which during the calibration phase: - several values ​​k(X;) of the coefficient k are calculated (206) using the following relation: k(X;) = [Spara(Xi) - DN] / [Sperp(X;) - DN], where: - Spara(X;) and Sperp(X;) are the values, respectively, of the optical signals Spara and Sperp at wavelength Xi5 and - DN is the self-noise of the reading unit, i.e. the optical signal measured by the reading unit when no optical signal is received by the reading unit, then - the value k(X) is calculated (208) by interpolation from the values ​​W.

6. Method according to any one of the preceding claims, in which the first and second measurement periods (TON, T0FF) follow each other immediately and the duration of each of these first and second measurement periods is less than 10 s.

7. Method according to claim 6, in which: - the first and second periods and the determination of the quantity of the gaseous component present in the sample from the values ​​S1(X), S2(X), S3(X) and S4(X) measured during these first and second periods, form a cycle (Ci) of measurements, and - the method comprises the repetition of this cycle of measurements several times during a time window of less than fifteen minutes, and - the calculation (250), by the computer, of a more precise value of the quantity of the gaseous component by averaging the quantities determined during each of the cycles of measurements contained within this time window.

8. Method according to any one of the preceding claims, in which, during the first and second periods, the optical signals S1, S2, S3 and S4 are measured for several different wavelengths X of interest, each of these wavelengths of interest corresponding to a particular gaseous component to be quantified in the excited sample, the measurements at these different wavelengths of interest being carried out simultaneously.

9. The method of claim 8, wherein the wavelengths of interest include wavelengths of interest corresponding to water vapor, oxygen, and nitrogen.

10. The method of claim 8, wherein the wavelengths of interest include wavelengths of interest corresponding to carbon monoxide and carbon dioxide.

11. A method according to any preceding claim, wherein a wavelength of interest corresponds to hydrogen.

12. A method according to any preceding claim, in in which the excitation beam is a continuous laser beam.

13. A method according to any preceding claim, wherein the method comprises: - the measurement (236) of the total pressure inside the enclosure, - the calculation (236) of partial pressures of hydrogen, carbon dioxide, oxygen, carbon monoxide and nitrogen from quantities quantified by Raman spectrometry of these gaseous components, at least one of these quantities being determined from the difference [(S3(X) / t3 - Sl(X) / tl] and the difference [S4(X) / t4 - S2(X) / t2], and - the calculation (236) of the partial pressure of water vapor from the difference between the total pressure PT minus the sum of the calculated partial pressures of hydrogen, carbon dioxide, oxygen, carbon monoxide and nitrogen.

14. Device for quantifying, by Raman spectrometry, at least one gaseous component inside a containment enclosure of a nuclear reactor, this device comprising: - a laser source (32) capable of emitting a monochromatic excitation beam which excites, inside the confinement enclosure, a sample of the gaseous medium located inside this confinement enclosure, - a reading unit (34) capable of measuring the optical signal scattered, by Raman effect, by the sample excited by the excitation beam, and - an electronic computer (36) configured to determine, from the optical signal measured by the reading unit, a quantity of the gaseous component present in the sample, characterized in that the calculator (36) is configured to automatically carry out the following steps: - during a first measurement period: - the excitation beam is linearly polarized in a predetermined direction, and - the reading unit measures values ​​S3(X) and S4(X), respectively, of an optical signal S3 and an optical signal S4 scattered by the sample, where: - optical signals S3 and S4 are the optical signals measured with polarization directions, respectively, parallel and perpendicular to the predetermined polarization direction of the excitation beam, when the sample is excited by the excitation beam, - X is a wavelength of interest, i.e. a wavelength at which, due to the Raman effect, a line appears in the spectrum of the scattered optical signal when the gaseous component to be quantified is present in the sample, and - during a second measurement period in the absence of the excitation beam, the reading unit measures values ​​S1(X) and S2(X), respectively, of an optical signal SI and of an optical signal S2 scattered by the sample, where the optical signals SI and S2 are the optical signals measured with polarization directions, respectively, parallel and perpendicular to the predetermined polarization direction of the excitation beam used during the first measurement period, when the excitation beam is absent, and - the calculator (36) determines the quantity of the gaseous component present in the sample from the difference [(S3(X) / t3 - S1(X) / t1] and the difference [S4(X) / t4 - S2(X) / t2], where t1, t2, t3 and t4 are the durations of the time windows during which the signals S1, S2, S3 and S4 are measured to obtain, respectively, the values ​​S1(X), S2(X), S3(X) and S4(X) in number of counts.

15. Device according to claim 14, in which the reading unit (34) comprises only one photodetector (54) for each wavelength of interest.

Images