A measurement system comprising a spectrometry device and a device for classifying a set of products into a plurality of classes.

Abstract

Measurement system comprising a spectrometry device and a device for classifying a set of products into a plurality of classes. This measurement system (10) comprises: - a spectrometry device (12) including an ionization source (14), a separation module (16), and a detector (18) capable of measuring an ionic flux of the sample, - an electronic device (20) for classifying a set of products into several classes, including a module (22) for acquiring a respective mass spectrum for each of the products in the set; a module (24) for determining, from each mass spectrum, a group of compound(s) included in the product associated with the respective mass spectrum; a module (26) for generating a table including, for each of the products in the set, a value of the intensity of the ionic flux for each compound present in said product;a module (28) for assigning, to each product of the set, a respective class from among the plurality of classes, by applying a multivariate statistical algorithm to the generated table. Figure for the abstract: Figure 1;

Description

Title of the invention: Measuring system comprising a spectrometry device and a device for classifying a set of products among a plurality of classes

[0001] The present invention relates to a measuring system comprising a spectrometry device and an electronic device for classifying a set of products among a plurality of classes.

[0002] The invention then relates to the field of product analysis by spectrometry. Spectrometry is used in many technical fields, such as petrochemistry, pharmacy, gas phase chemistry, organic chemistry, physics, astrophysics, biology.

[0003] Given the very close similarity in chemical composition between certain products in a set of products, it is quite difficult to differentiate them from each other. For example, in the field of lubricating oils, it is complex to differentiate recycled oils from non-recycled oils, also called conventional oils, or even original oils. Indeed, recycled oils undergo different stages of re-refining, or various chemical treatments to remove the various impurities (additives, oxidation products), which makes their composition very close to a conventional oil.

[0004] It is nevertheless necessary to be able to distinguish these products from each other. In the field of lubricating oils, it is generally preferable to add an additive to a recycled oil in order to improve its performance, and it is therefore useful to know whether the oil is recycled or not. Furthermore, from an ecological and reuse point of view, it is also interesting to know whether the oil is recycled or not.

[0005] It is known to analyze products, such as lubricating oils, by spectrometry to determine which are the major compounds in these products.

[0006] However, often, this determination of majority compounds does not allow the products to be differentiated from each other.

[0007] The aim of the invention is then to propose a measurement system making it possible to classify a set of products among a plurality of classes from a spectrometry device, that is to say to classify each of the products of the set in a respective class of the plurality of classes.

[0008] To this end, the invention relates to a measuring system comprising:

[0009] - a spectrometry device comprising:

[0010] + an ionization source inside which a sample of each product is capable of being ionized,

[0011] + a separation module connected to the output of the ionization source and capable of separate in an electrostatic field ions from the ionized sample according to their mass to charge ratio, and

[0012] + a detector connected to the output of the separation module and capable of measuring a flow ionic of the sample,

[0013] - an electronic device for classifying a set of products from a plurality of classes, the electronic classification device being connected to the spectrometry device and comprising:

[0014] + an acquisition module configured to acquire, from the spec device trometry, a respective mass spectrum for each of the products in the set, each mass spectrum representing an ion flux intensity as a function of a mass to charge ratio;

[0015] + a determination module configured to determine, from each spectrum mass, a group of compound(s) included in the product associated with the respective mass spectrum, each compound corresponding to a peak of the mass spectrum and being determined from the mass to charge ratio for said peak;

[0016] + a generation module configured to generate a table comprising, for each products of the set, an ion flux intensity value for each compound present in said product;

[0017] + an allocation module configured to allocate, to each product in the set, a respective class among the plurality of classes, by applying a multivariate statistical algorithm to the generated table.

[0018] With the measurement system according to the invention, the application of a multivariate statistical algorithm to the table comprising, for each of the products in the set, a value of the intensity of the ionic flux for each compound present in the product analyzed, each compound having been determined from its mass to charge ratio, makes it possible to assign to each product in the set a class from among the plurality of classes.

[0019] In other words, the application of the multivariate statistical algorithm, taking into account as variables on the one hand the intensity, and on the other hand the mass to charge ratio, makes it possible to carry out a precise classification of each of the products in the set.

[0020] Preferably, the multivariate statistical algorithm is a partial least squares regression multivariate statistical algorithm, such as an orthogonal partial least squares regression algorithm, also called O-PLS (from the English Orthogonal Partial Least Squares regression), or an orthogonal partial least squares regression algorithm with discriminant analysis, also called O-PLS-DA (from the English Orthogonal Partial Least Squares re- Discriminant Analysis).

[0021] According to other advantageous aspects of the invention, the measuring system comprises one or more of the following characteristics, taken individually or in all technically possible combinations:

[0022] - the ionization source is an electro-nebulizer ionization source;

[0023] - the spectrometry device is an Orbitrap type mass spectrometer;

[0024] - the spectrometry device is calibrated with at least one reference product, prior to obtaining mass spectra for the set of products;

[0025] - the multivariate statistical algorithm is a multivariate statistical algorithm by re partial least squares regression;

[0026] the multivariate statistical algorithm preferably being chosen from the group consisting of: an algorithm by orthogonal regression of partial least squares, and an algorithm by orthogonal regression of partial least squares with discriminant analysis;

[0027] - each mass spectrum has several peaks, each corresponding to a respective value of the mass to charge ratio m / z, and the acquisition module is configured to detect each peak of a corresponding mass spectrum by thresholding, a peak being detected when its intensity is greater than a threshold;

[0028] the threshold preferably corresponding to a predefined signal-to-noise ratio, such as a signal-to-noise ratio greater than or equal to 3;

[0029] the threshold being preferably still equal to a multiple of an average noise of the mass spectrum, such as a multiple greater than or equal to 3;

[0030] - the determination module is configured to determine the compound cor corresponding to each peak in the mass spectrum according to predefined conditions including a list of heteroelement types, a maximum number of heteroelements in the compound and a maximum error tolerance between a theoretical mass-to-charge ratio and an experimental mass-to-charge ratio of the compound;

[0031] the predefined conditions preferably further including a configuration of the ions selected from a radical configuration and a deprotonated configuration; and a range of permitted values ​​for a ratio of the number of hydrogen atoms divided by the number of carbon atoms within the compound;

[0032] - the generation module is configured to position a value in the table zero for each compound absent from a respective product;

[0033] - each product is a lubricating oil; and

[0034] - the plurality of classes comprises an original product class and a product class. recycled product.

[0035] These characteristics and advantages of the invention will appear more clearly on reading the description which follows, given solely by way of example, not limiting, and made with reference to the attached drawings, in which:

[0036] [Fig.l] [Fig.l] is a schematic representation of a measuring system according to the invention, comprising a spectrometry device and an electronic device for classifying, via the spectrometry device, a set of products from among a plurality of classes;

[0037] [Fig.2] [Fig.2] illustrates a mass spectrum obtained via the spectrometry device of [Fig.l];

[0038] [Fig.3] [Fig.3] is a flowchart of a method for classifying a set of products from among a plurality of classes using the spectrometry device of [Fig.l], the method being implemented by the electronic classification device of [Fig.l]; and

[0039] [Fig.4] [Fig.4] is a view of a diagram showing the classifications obtained for two respective sets of products.

[0040] In the remainder of the description, the expression “substantially equal to” designates a relationship of equality to plus or minus 10%, preferably to plus or minus 5%.

[0041] In [Fig.l], a measuring system 10 comprises a spectrometry device 12. The spectrometry device 12 comprises an ionization source 14 inside which a sample of each product is capable of being ionized; a separation module 16 connected to the output of the ionization source 14 and capable of separating ions from the ionized sample; and a detector 18 connected to the output of the separation module 16 and capable of measuring an ionic flux for said sample.

[0042] The measurement system 10 also comprises an electronic device 20 for classifying a set of products among a plurality of classes from the spectrometry device 12, the electronic classification device 20 being connected to the spectrometry device 12.

[0043] The spectrometry device 12 is more particularly a mass spectrometer making it possible to detect and characterize molecules of interest in each product, by measuring their mass, and then to characterize their chemical structure. The spectrometry is then based on the gas phase separation of charged molecules, namely ions, as a function of their mass to charge ratio (also noted mass / charge), also called mass / charge ratio, also noted m / z.

[0044] The ionization source 14 is capable of vaporizing the molecules of the sample and ionizing them. The ionization source 14 is capable of being used either in positive mode to study positive ions, or in negative mode to study negative ions.

[0045] The ionization source 14 is typically an electrospray ionization source, also called an ESI (ElectroSpray Ionization) type source. Electrospray ionization is the dispersion of a liquid in the form of electrically charged droplets.

[0046] The ionization source 14 is then for example chosen from the group consisting of: ESI source (from the English ElectroSpray lonization); DESI source (from the English De-sorption ElectroSpray lonization); SESI source (from the English Secondary ElectroSpray lonization); LAESI source (from the English Laser Ablation ElectroSpray lonization); APCI source (from the English Atmospheric Pressure Chemical lonization); and APPI source (from English Atmospheric Pressure Photo lonization).

[0047] The separation module 16 comprises, for example, a mass spectrometry cell, also known as MS (from the English Mass Spectrometry), capable of separating the ions according to their mass / charge ratio, also called mass to charge ratio or mass to charge ratio, and the spectrometry device 12 is then also called a mass spectrometer.

[0048] The mass spectrometry cell is an analyzer capable of separating ions in an electrostatic field, and the spectrometry device 12 is typically an Orbitrap type mass spectrometer.

[0049] A mass spectrum 40 is obtained via a Fourier transform which converts the acquired time signal into a frequency spectrum, the Orbitrap type mass spectrometer using the Fourier transform to analyze the revolution frequencies of the ions trapped in the electrostatic field.

[0050] The detector 18 is capable of transforming the ions into an electrical signal. The more ions there are, the greater the current. In addition, the detector 18 is capable of amplifying the signal obtained, in particular so that it can be processed more easily by the electronic classification device 20.

[0051] In addition, the spectrometry device 12 is calibrated with at least one reference product, prior to obtaining mass spectra 40 for the set of products. This calibration makes it possible to correct the mass deviations of the spectrometry device 12 based on the exact mass of compound(s) of the at least one reference product. The at least one reference product is for example a Pierce™ negative ion calibration solution, marketed by the company Thermo Scientific Chemicals, comprising a mixture of ionizable molecules. The at least one reference product preferably comprises a mixture of sodium dodecyl sulfate, also noted SDS, sodium taurocholate and Ultramark™ 1621, a compound marketed by the company Thermo Scientific Chemicals.

[0052] The calibration is carried out prior to a step of acquiring the respective mass spectra 40 for each of the products in the set (step described below), and this calibration is then also called pre-calibration, or pre-acquisition calibration. This calibration is also called external calibration, being carried out with a reference product external to the set of products, i.e. distinct from the products in the set.

[0053] The electronic classification device 20 is configured to classify, i.e. classify, the set of products among the predefined plurality of classes. In other words, the classification device 20 is configured to assign to each product of the set a respective class among said plurality of classes, i.e. classify each product of the set in a respective class. Those skilled in the art will of course understand that the same class can be assigned to several products of the set, and that said products are then classified in this same class.

[0054] The electronic classification device 20 comprises a module 22 for acquiring, from the spectrometry device 12, a respective mass spectrum 40 for each of the products in the set; a module 24 for determining, from each mass spectrum 40, a group of compound(s) included in the product associated with the respective mass spectrum 40; a module 26 for generating a table comprising, for each of the products in the set, an ion flux intensity for each compound present in said product; and a module 28 for assigning, to each product in the set, a respective class by applying a multivariate statistical algorithm to the table.

[0055] In the example of [Fig.l], the electronic classification device 20 comprises an information processing unit 30 formed for example of a memory 32 and a processor 34 associated with the memory 32.

[0056] Each product is for example a lubricating oil.

[0057] The plurality of classes includes, for example, an original product class, also called a conventional product class, or a non-recycled product class; and a recycled product class.

[0058] In the example of [Fig. 4], the plurality of classes consists of the original product class, also called first class C1, and the recycled product class, also called second class C2. In other words, in the example of [Fig. 4], the electronic classification device 20 is configured to determine, for each product in the set, whether said product is an original product or a recycled product. In the example of [Fig. 4], each product is a lubricating oil, and the electronic classification device 20 is then configured to determine, for each of the oils analyzed, whether the oil is an original oil, also called conventional oil; or a recycled oil.The electronic classification device 20 then makes it possible to determine more effectively which, among a set of oils, are the recycled oils to which an additive should be added, in order to obtain better performance, particularly lubrication.

[0059] In the example of [Fig.l], the acquisition module 22, the determination module 24, the generation module 26 and the allocation module 28 are each implemented under form of software, or a software brick, executable by the processor 34. The memory 32 of the electronic classification device 20 is then capable of storing software for acquisition, from the spectrometry device 12, of a respective mass spectrum 40 for each of the products in the set; software for determining, from each mass spectrum 40, the group of compound(s) included in the product associated with the respective mass spectrum 40; software for generating the table comprising, for each of the products in the set, an ion flux intensity for each compound present in said product; and software for assigning, to each product in the set, a respective class by applying the multivariate statistical algorithm to the table. The processor 34 is then capable of executing each of the software among the acquisition software, the determination software, the generation software and the assignment software.

[0060] In a variant not shown, the acquisition module 22, the determination module 24, the generation module 26 and the allocation module 28 are each produced in the form of a programmable logic component, such as an FPGA (Field Programmable Gate Away) or in the form of a dedicated integrated circuit, such as an ASIC (Application Specific Integrated Circuit).

[0061] When the electronic classification device 20 is produced in the form of one or more software programs, that is to say in the form of a computer program, it is also capable of being recorded on a medium, not shown, that is readable by a computer. The computer-readable medium is, for example, a medium capable of storing electronic instructions and of being coupled to a bus of a computer system. By way of example, the readable medium is an optical disk, a magneto-optical disk, a ROM memory, a RAM memory, any type of non-volatile memory (for example EPROM, EEPROM, FLASH, NVRAM), a magnetic card or an optical card. A computer program comprising software instructions is then stored on the readable medium.

[0062] The acquisition module 22 is configured to acquire, from the spectrometry device 12, a respective mass spectrum 40 for each of the products in the set.

[0063] Each mass spectrum 40 represents the intensity I of the ionic flux as a function of the mass to charge ratio m / z, as shown in [Fig.2]. Each mass spectrum 40 typically comprises several peaks 45, each corresponding to a respective value of the mass to charge ratio m / z, and having a respective intensity I greater than a threshold, not shown. Those skilled in the art will understand that the value of the mass to charge ratio m / z varies from one peak 45 to another.

[0064] The acquisition module 22 is for example configured to detect each peak 45 of a corresponding mass spectrum 40 by thresholding, a peak 45 being detected when its ionic flux intensity I is greater than the threshold defined above.

[0065] The threshold preferably corresponds to a predefined signal-to-noise ratio, for example a signal-to-noise ratio greater than or equal to a predefined minimum value, such as 3. The threshold is then typically equal to a multiple of an average noise of the corresponding mass spectrum 40, the multiple being greater than or equal to the predefined minimum value, such as 3.

[0066] In other words, for the detection of each peak 45, the acquisition module 22 is for example configured to determine the average noise of the mass spectrum 40, to calculate the threshold by multiplying said average noise by the desired multiple, such as 3, then to detect all the local extrema of the spectrum 40 having an ionic flux intensity value I greater than or equal to the threshold thus calculated.

[0067] Internal calibration is for example carried out during the acquisition of the mass spectrum, by choosing one or more known peaks of the sample to correct the differences in mass to charge ratio between the exact mass to charge ratio of these known compounds, and the experimental mass to charge ratio observed during the acquisition. This type of calibration is also known under the English term "lock mass".

[0068] Alternatively, or in addition, the internal calibration is carried out after the acquisition of the mass spectrum, by reading the contents of the mass spectrum, and calibrating it with a calibration list of mass to charge ratios of reference compounds. The mass spectrum is then not recalibrated with a known spectrum, but with a set of reference mass to charge ratios.

[0069] The internal calibration, whether carried out during the acquisition of the mass spectrum or after the acquisition of the mass spectrum, is preferably carried out from the 45 peaks previously detected by thresholding, in order to have a more reliable calibration. Indeed, the thresholding makes it possible to dissociate the ions and the background noise by the use of a suitable signal / noise threshold, and the internal calibration is then carried out for only the 45 peaks corresponding to ions, and not for background noise.

[0070] The determination module 24 is configured to determine, from each mass spectrum 40, a group of compound(s) included in the product associated with the respective mass spectrum 40, each compound corresponding to a peak 45 of the mass spectrum 40 and being determined from the mass to charge ratio m / z for said peak 45.

[0071] The determination module 24 is for example configured to determine the compound corresponding to each peak 45 of the mass spectrum 40 according to predefined conditions including a list of heteroelement types, a maximum number of heteroelements in the compound and a maximum error tolerance between a theoretical mass to charge ratio and an experimental mass to charge ratio of the compound.

[0072] The list of heteroelement types includes, for example, nitrogen, oxygen and sulfur. In addition, the list of heteroelement types also includes phosphorus, boron, chlorine, and silicon.

[0073] The maximum number of heteroelements in the compound is typically predefined separately for each of the heteroelement types in the list. As a purely indicative example, the maximum number of heteroelements in the compound is set equal to 2 for nitrogen, equal to 6 for oxygen and equal to 2 for sulfur. When the list of heteroelement types includes phosphorus, boron, chlorine, the maximum number of heteroelements in the compound is for example set equal to 2 for phosphorus, equal to 1 for boron, equal to 1 for chlorine, and equal to 8 for silicon.

[0074] The maximum error tolerance between the theoretical mass-to-charge ratio and the experimental mass-to-charge ratio of the compound is, for example, between 1 ppm and 3 ppm, and preferably between 1.5 ppm and 2 ppm.

[0075] As an optional addition, the predefined conditions include an ion configuration, selected from a radical configuration and a deprotonated configuration.

[0076] As an optional addition, the predefined conditions include a range of authorized values ​​for a ratio of the number of hydrogen atoms divided by the number of carbon atoms within the compound, also called the H / C range. In addition to the maximum error tolerance and the maximum number of heteroelements, this optional additional condition makes it possible to limit a number of unsaturation and an aromaticity of the molecular formulas of the compounds. The range of authorized values ​​is, for example, between a minimum value equal to 0 and a maximum value equal to 3.

[0077] The determination module 24 is then configured to determine the compound corresponding to each peak 45 of the mass spectrum 40 from the mass to charge ratio associated with the respective peak 45 and according to the predefined conditions described above. In other words, for each peak 45, the determination module 24 is configured to determine, as the compound associated with said peak 45, the compound whose exact mass to charge ratio is equal, to within the maximum error tolerance, to the mass to charge ratio associated with said peak 45 and which further complies with the other predefined conditions relating to the types of heteroelements and the maximum number of heteroelements.In the event that several compounds meet all the predefined conditions for a given peak 45, then the determination module 24 is configured to determine, as the compound associated with said peak 45, the compound whose exact mass-to-charge ratio is closest to the mass-to-charge ratio associated with said peak 45.

[0078] As an optional addition, the determination module 24 is then configured to determine the compound corresponding to each peak 45 of the mass spectrum 40, further taking into account one or other of the additional predefined conditions, namely, further depending on the configuration of the ions, then chosen to be equal to a radical configuration or a deprotonated configuration, and / or further depending on the range of authorized values ​​for the ratio of the number of hydrogen atoms divided by the number of carbon atoms.

[0079] These predefined conditions are typically configurable by a user of the electronic classification device 20.

[0080] The determination module 24 is thus configured to determine a molecular formula corresponding to each peak 45 of the mass spectrum 40, from an exact mass based on the sum of the mass of each atom present in a molecule. The data are aligned from the moment when the exact mass of a compound of a sample is equal to the exact mass of the compound of another sample, for example of a reference sample. They have the same molecular formula, and therefore necessarily the same exact mass.

[0081] Alternatively, the determination module 24 does not determine the molecular formula of the compounds, and the determination module 24 is then configured to determine an experimental mass corresponding to each peak 45 of the mass spectrum 40. The data are then aligned from the moment when the mass difference between a compound present on a spectrum, and a compound present on another spectrum is less than or equal to an error tolerance set by an operator. These are the same compounds, but they do not necessarily have the same experimental mass.

[0082] The generation module 26 is then configured to generate the table comprising, for each of the products in the set, an ionic flux intensity value for each compound present in said product.

[0083] The intensity value contained in the table for each compound present in the respective product is preferably a relative intensity, expressed as a ratio between the ion flux intensity value for said compound and the total ion flux intensity value for the respective product. The total ion flux intensity value is the cumulative ion flux intensity value for the entire mass spectrum 40 obtained with said product. The intensity value is one of the variables taken into account by the multivariate statistical algorithm for assigning a respective class to each product in the set, and taking into account a relative intensity, rather than an absolute intensity, then makes it possible to limit the effects of a possible measurement deviation due to the spectrometry device 12, from one product analyzed to another.

[0084] In addition, the generation module 26 is configured to position in the table a zero value for each compound absent from a respective product. The table then comprises, for each respective product, a zero value for each compound absent from said product.

[0085] By way of example, the table is in the form of a matrix consisting of rows and columns, and each row corresponds to a respective product, each column corresponding to a respective compound, and each cell of the matrix then containing the intensity value, such as the relative intensity, for the compound corresponding to the column considered and for the product corresponding to the row considered. In other words, according to this example, each row of the table corresponds to a respective mass spectrum 40, and each cell of non-zero value of the row corresponds to a peak 45 detected for said mass spectrum 40, the non-zero value contained in said cell then being the intensity value, such as the relative intensity, corresponding to the compound which was determined by the determination module 24 for said peak 45.

[0086] The allocation module 28 is then configured to assign, to each product of the set, a respective class C1, C2 from among the plurality of classes, by applying the multivariate statistical algorithm to the generated table.

[0087] The multivariate statistical algorithm applied to obtain this classification of each of the products referenced in the table is typically a partial least squares regression algorithm.

[0088] The multivariate statistical algorithm is for example an orthogonal regression algorithm of partial least squares, also noted O-PLS, or an orthogonal regression algorithm of partial least squares with discriminant analysis, also noted O-PLS-DA.

[0089] The multivariate statistical algorithm makes it possible to classify the products of the set from qualitative variables on the one hand and quantitative variables on the other hand. A person skilled in the art will then observe that the qualitative variables correspond to the mass-to-charge ratios of the ions obtained by the ionization of each of the products analyzed, and that the quantitative variables correspond to the intensities, preferably the relative intensities, which are attributed to them.

[0090] The operation of the electronic classification device 20 will now be described with reference to [Fig. 3] representing a flowchart of the method of classifying the set of products among the plurality of classes, said classification method being implemented by the electronic classification device 20.

[0091] During an initial step 100, the classification device 20 acquires, via its acquisition module 22 and from the spectrometry device 12, the respective mass spectrum 40 for each of the products in the set. In other words, the acquisition module 22 acquires all the mass spectra 40 obtained via the spectrometry device 12 for all the products in the set.

[0092] The acquisition step 100 typically comprises the detection, by thresholding, of the peak(s) 45 of the respective mass spectrum 40, each detected peak 45 then having an ion flux intensity greater than the threshold, said peak alone preferably corresponding to a predefined signal-to-noise ratio.

[0093] In addition, during this initial acquisition step 100, the acquisition module 22 performs an internal calibration of the acquired mass spectra 40, with the respective mass to charge ratios of the predefined collection of reference elements. When the products analyzed are lubricating oils, the reference elements used for this internal calibration are preferably phenols and fatty acids.

[0094] At the end of the acquisition step 100, the classification device 20 moves on to a next step 110 during which it determines, via its determination module 24 and from each mass spectrum 40, the group of compound(s) included in the product associated with the respective mass spectrum 40. Each compound corresponds to a peak 45 of the mass spectrum 40, and is determined from the value of the mass to charge ratio m / z for said peak 45.

[0095] The determination step 110 comprises, for example, the determination, for each peak 45 detected and from the mass to charge ratio associated with said peak 45, of the corresponding compound as a function of the predefined conditions, as described above.

[0096] During the following step 120, the classification device 20 generates, via its generation module 26, the table comprising - for each of the products in the set - an ion flux intensity value, such as a relative intensity, for each compound present in said product. The relative intensity is equal to the ratio, i.e. the ratio, of the ion flux intensity for said compound, divided by the value of the total ion flux intensity. In other words, the relative intensity is equal to the ratio of the ion flux intensity for peak 45 associated with the determined compound, divided by the cumulative value of the ion flux intensity for the entire mass spectrum 40 comprising said peak 45.

[0097] During generation step 120, the table is preferably generated with zero values ​​for each of the compounds absent from a respective product.

[0098] At the end of the generation step 120, the classification device 20 moves on to the last step 130 during which it assigns, via its assignment module 28 and to each product of the set, a respective class from among the plurality of classes, this by applying a multivariate statistical algorithm to the table generated during the previous generation step 120.

[0099] During the allocation step 130, the qualitative variables of the multivariate statistical algorithm are then the exact mass-to-charge ratios, corresponding to the compounds determined during the determination step 110; and the quantitative variables of the multivariate statistical algorithm are the intensities which are attributed to these mass-to-charge ratios, i.e. to these compounds. The intensities taken into account are preferably the relative intensities, as defined above.

[0100] The multivariate statistical algorithm used during the allocation step 130 is ty using a partial least squares regression algorithm, such as the O-PLS algorithm or the O-PLS-DA algorithm.

[0101] The classification thus obtained by the classification device 20 is then particularly effective, as shown - by way of example - in a diagram 200 visible in [Fig.4]. In the example of [Fig.4], the products analyzed are lubricating oils.

[0102] In the example of [Fig.4], each of the tablets shown corresponds to a respective product, and the string of alphanumeric characters indicated near each tablet then corresponds to a code name of each respective product, these code names having been kept in [Fig.4] only in order to facilitate the distinction of the tablets from one another, and having no impact on the classification of the products.

[0103] In [Fig.4], diagram 200 shows an example of classification obtained, with the products associated with the first class Cl corresponding to negative abscissas, all approximately equal to -30; and the products associated with the second class C2 corresponding to positive abscissas, generally approximately equal to +13.

[0104] Those skilled in the art will then understand that, on this diagram 200, the classification carried out is quite distinct, with a substantial difference between the abscissas of the two groups, that is to say of the products of the two distinct classes.

[0105] Furthermore, in order to verify the accuracy of the classification carried out by the classification device 20, each of the products analyzed was also identified by an identifier of its true class among the first class Cl and the second class C2, this identifier naturally not being known to the classification device 20. In [Fig. 4], the identifier indicating the first class Cl corresponds to the light-colored pellets, and the identifier indicating the second class C2 corresponds to the darker-colored pellets, and the person skilled in the art will then be able to observe that the classification carried out by the classification device 20 is perfectly accurate, all the light-colored pellets actually being classified in the first class Cl, and all those of a darker color being classified in the second class C2.

[0106] In addition, several other products analyzed were not previously identified by a class identifier, these products being designated by the reference ND in [Fig.4].

[0107] In diagram 200 of [Fig.4], the abscissa axis corresponds to intergroup statistical variability, and the ordinate axis corresponds to intragroup variability.

[0108] Those skilled in the art will further observe that the ND products not previously identified are classified as belonging to the second class C2, their abscissas being on the diagram 200 close to those of the other products of the second class C2.

[0109] Thus, the measurement system 10 according to the invention makes it possible to efficiently assign, to each product in the set, a class from among the plurality of classes.

[0110] The application of the multivariate statistical algorithm, taking into account as variables on the one hand the intensity, and on the other hand the mass to charge ratio, makes it possible to carry out a precise classification of each of the products in the set.

Claims

Claims

1. Measuring system (10) comprising: - a spectrometry device (12) comprising: + an ionization source (14) inside which a sample of each product is capable of being ionized, + a separation module (16) connected to the output of the ionization source (14) and capable of separating in an electrostatic field ions of the ionized sample according to their mass to charge ratio, and + a detector (18) connected to the output of the separation module (16) and capable of measuring an ionic flux of the sample, - an electronic device (20) for classifying a set of products among a plurality of classes, the electronic classification device (20) being connected to the spectrometry device (12) and comprising: + an acquisition module (22) configured to acquire, from the spectrometry device (12), a respective mass spectrum (40) for each of the products of the set,each mass spectrum (40) representing an intensity (I) of the ionic flux as a function of a mass to charge ratio (m / z); + a determination module (24) configured to determine, from each mass spectrum (40), a group of compound(s) included in the product associated with the respective mass spectrum (40), each compound corresponding to a peak (45) of the mass spectrum (40) and being determined from the mass to charge ratio (m / z) for said peak (45); + a generation module (26) configured to generate a table comprising, for each of the products of the set, an intensity value of the ionic flux for each compound present in said product; + an assignment module (28) configured to assign, to each product of the set, a respective class (Cl, C2) from among the plurality of classes, by applying a multivariate statistical algorithm to the generated table.,

2. The system (10) of claim 1, wherein the ionization source (12) is an electrospray ionization source.

3. System (10) according to claim 1 or 2, wherein the spectrometry device (12) is an Orbitrap type mass spectrometer.

4. System (10) according to any one of the preceding claims, wherein the spectrometry device (12) is calibrated with at least one reference product, prior to obtaining mass spectra (40) for the set of products.

5. The system (10) of any preceding claim, wherein the multivariate statistical algorithm is a partial least squares regression multivariate statistical algorithm; the multivariate statistical algorithm preferably being selected from the group consisting of: a partial least squares orthogonal regression algorithm, and a partial least squares orthogonal regression algorithm with discriminant analysis.

6. System (10) according to any one of the preceding claims, wherein each mass spectrum (40) comprises several peaks (45), each corresponding to a respective value of the mass to charge ratio m / z, and the acquisition module (22) is configured to detect each peak (45) of a corresponding mass spectrum (40) by thresholding, a peak (45) being detected when its intensity is greater than a threshold; the threshold preferably corresponding to a predefined signal to noise ratio, such as a signal to noise ratio greater than or equal to 3; the threshold being more preferably equal to a multiple of an average noise of the mass spectrum, such as a multiple greater than or equal to 3.

7. The system (10) of any preceding claim, wherein the determination module (24) is configured to determine the compound corresponding to each peak (45) of the mass spectrum (40) based on predefined conditions including a list of heteroelement types, a maximum number of heteroelements in the compound, and a maximum error tolerance between a theoretical mass-to-charge ratio and an experimental mass-to-charge ratio of the compound; the predefined conditions preferably further including an ion configuration selected from a radical configuration and a deprotonated configuration; and a range of permitted values ​​for a ratio of the number of hydrogen atoms divided by the number of carbon atoms within the compound.

8. System (10) according to any one of the preceding claims, wherein the generation module (26) is configured to position in the table a zero value for each compound absent from a respective product.

9. System (10) according to any one of the preceding claims, in which each product is a lubricating oil.

10. A system (10) according to any preceding claim, wherein the plurality of classes comprises an original product class (Cl) and a recycled product class (C2).

Images