WO2022084103A1 - Method for electronic sensory analysis of gases, particularly co2, used in foodstuffs or pharmaceutical applications - Google Patents

Abstract

Method for sensory analysis of industrial gas samples delivered to an industrial consumer site, characterised by the following steps: • - providing a hermetically sealed enclosure in which a detection system is placed which comprises one or each of the following sensor systems: • i) one or more electrochemical cells; • j) one or more MOS ("Metallic Oxides Semiconductor") sensors; • - inserting the gas sample into the chamber; • - the cells and/or sensors present in the enclosure react in the presence of any organic compounds present in the gas sample, and provide a detection response; • - providing a data acquisition and processing system; • - performing, beforehand, a phase of training of the analysis by a human panel, and performing, by means of the data acquisition and processing system, a step of reconciling the two types of data, i.e. the results of the tests performed by the human panel and the results provided by the sensors and/or cells in the enclosure.

Description

Process for the electronic sensory analysis of gases, in particular CO2, used in food or for pharmaceutical applications

The present invention relates to the field of quality control of industrial gases delivered to an industrial consumer site, particularly in the food and pharmaceutical industry. She is more specifically interested in the applications of CO2, particularly in the food industry. We know that CO2 is used in particular for the carbonation of beverages.

We detail in what follows the problems of CO2 in particular.

We then know that the regulatory authorities of this industry have issued standards for the various market players (producers, transporters, bottlers) on the various raw materials that go into the composition of "carbonated" drinks (water, carbon dioxide, sugar, flavourings, etc.).

As far as gas is concerned, the CO2 quality specifications apply to the entire distribution chain, from the product production plant upstream to the gas storage container (bottle or tank) on the beverage bottler website.

This technical field has notably recommended carrying out organoleptic tests on CO2 (liquid and gaseous). These tests are therefore carried out at the various stages of production and distribution of CO2 (production plant, road tanker, gas bottle, tank and installation on the food production site).

These tests are well described in the literature and the predominant element of measurement is the human, through his senses (vision, smell, taste). These are mainly qualitative measurements that are difficult to quantify from a metrological point of view, with a binary assessment in the end (compliant, non-compliant).

An odor and a taste are the result of the combination of an association of molecules (currently about twenty in the standard) all with variable concentration levels for each molecule. Concretely in the industry currently, the taste and odor tests are carried out in parallel, on a sample to be evaluated on the one hand, and on a reference CO2 considered as "pure" in the sense that it would be completely devoid of aroma. and/or flavor (or that it is considered to have the “normal” CO2 odor).

And so each test ends with the comparison between these two products and any appearance of minor difference with the reference CO2 should end with the systematic downgrading of the analyzed CO2. In fact, it happens that this sensory judgment is impacted by economic issues, even unconsciously.

Indeed, the test currently carried out by the technicians in the factory or in the laboratory constitutes the last milestone before the release of the product.

It happens then that all the analyzes that are carried out with conventional equipment show that the product is blockable because an odor is smelled. But in practice, the consequences of such a blockage are very serious (shutdown of a plant, qualification of a reservoir/road tank, shutdown of a customer site using this fluid, modification of the delivery circuit, triggering of a quality incident, etc.), any consequences that may generate significant direct and indirect costs.

And so the human brain under stress may tend to increase its level of tolerance of an anomaly by responding, for example, “it smells a little”.

The electronic nose, on the contrary, is not emotional and therefore its response is purely factual.

Moreover, in absolute terms this reference product does not exist and given the consumption of liquid CO2 necessary in this type of beverage application, it is impossible to qualify a source and then to keep a reserve of CO2 cylinders in stock. liquid called “pure”.

The second difficulty lies in the fact that the environment (especially in an industrial environment) affects the measurement methodology.

Indeed, when an operator is authorized to carry out this type of test, it is because he has succeeded in recognizing odors during an "examination", except that this qualification is true at a time t (depending on his state physiological) and that it also depends on its capacity:

• to memorize the recognition of these odors over time

• to maintain their level of nasal sensitivity over time

In fact, these people have a sensitivity that varies, among other things, because of their state of health (sleep, fatigue, memory, ability to concentrate, etc.) The environment factor is also a risk.

The CO2 production sites are more generally located in industrialized areas in which other companies have industrial processes generating odorized discharges.

This "polluted" environment affects the olfactory capacity of the operators on a daily basis and this background noise minimizes the perception of the operators. Depending on the weather and the operation of other industrial processes, this impact is variable.

It is then proposed according to the present invention to carry out a sensory analysis of industrial gases delivered to an industrial consumer site, particularly in the food and pharmaceutical industry, and in particular of CO2, an analysis which is not dependent/influenced by its environment and/or or by the individual, which would make it possible to:

• Streamline the validation of the sensory quality of food CO2.

• Limit divergent interpretations (methods, means, people, environment, etc.) related to the sensory capacity which in essence is unique for each individual.

• Standardize the process throughout the chain and over time.

As will be seen in more detail in the following, the present invention proposes to analyze a sample of gas, for example CO2, gaseous, in an enclosure equipped with electronic sensors to detect molecules (or family of molecules): either those identified according to the legislation (benzene, H2S and acetaldehyde) or those known for their sensory specificity (alkenes for example).

To this end, this proposal implements:

- a hermetic enclosure in which is placed a detection system, a detection system which comprises one or each of the following sensor systems:

- one or more electrochemical cells, cells which are each of them specific and which react to a given molecule such as SO2, H2S, or even NH3;

- One or more MOS (“Metallic Oxides Semiconductor”) type sensors, each sensor reacting to a chemical oxidation-reduction reaction. - the sample of gas, for example of CO2, to be evaluated, approaches the enclosure in the form of a gas stream (in a gaseous form or in the form of solid dry ice in the case of CO2), it is then positioned in the enclosure. In the case of dry ice, the ice will gradually sublimate into gas and thus release the potential "odorous" impurities initially "trapped" in the solid CO2.

- as will be described in more detail below, the cells and sensors present in the enclosure will react in the presence of organic compounds, according to a type of response which will not necessarily be directly linked to a molecule in particular but which will give much more a "profile" which will be considered either as normal/usual or as abnormal/unusual. To better understand this, the cells will give a quantifiable electrical response associated with a molecule which is either the one that the sensor is supposed to detect or a molecule which gives an interference type response but which clearly establishes the presence of an unusual impurity. MOS sensors will also give an answer, but rather in the form of a "profile or a fingerprint" which cannot be directly quantified but which says that the "fingerprint" is either identical or different from that of the reference product. the enclosure is advantageously equipped with a system of pressure/temperature/hygrometry sensors to control the operating conditions of the analysis.

- it is advantageous to have a flow meter to regulate a flow rate for purging/scavenging the enclosure.

Each molecule of interest will have been tested beforehand, a signature combining the response of several sensors on this molecule being identified.

Indeed, for the data provided by the device to be usable and interpretable, it is necessary to have previously learned the machine based on a reference method.

Consequently, according to the invention, a series of blind olfactory tests are carried out beforehand with a representative panel of people (typically a panel of several dozen people) following the protocol currently followed for the human nose, from so as to obtain a consensus, a tendency, an average of the answers.

Once the odor has been characterized by the human panel, the same test (ie on the same sample) is carried out on the device of the invention. Then comes the work of reconciling the two types of data, so as to associate a human representation of an odor with the response obtained using one or more of the sensors present in the device.

These tests are carried out both qualitatively (characterization of the smell) and quantitatively (intensity of the smell).

For each "guinea pig" person, a response grid system is proposed to help the "guinea pig" to associate his feelings with his experience.

To avoid unconsolidated responses, odorless or duplicate samples can be added to avoid false positive or uncertain positive phenomena.

Examples of implementation are described below in order to better understand the invention.

First example of application: continuous monitoring of CO2 production.

On a given CO2 production plant, a systematic sensory analysis is usually carried out on each tank once filled, in order to proceed with its qualification and then its subdivision. These operations are carried out by technicians from the production site, which poses two major constraints for the staff, who must be:

• Qualified (on a voluntary basis) and followed,

• Physically present (and able) to qualify and subdivide the reservoir in real time.

The method according to the invention makes it possible to automate the operation and therefore to limit the solicitation of local operators for critical situations.

Indeed, carrying out an odor test in a factory is an operation that requires time, a form of serenity, concentration.

But the common daily conditions of a factory operator is to manage unexpected situations (a process that goes wrong, a driver who does not have his clearances, an incident on his tank, regulatory inspections, costly maintenance operations and risky, customer audits, installation start-up and shutdown operations and all on call 24 hours a day).

Under these conditions, when a priority event (safety, quality, process) occurs, the odor test will pass last because perceived as "less critical". Second application example: periodic monitoring of CO2 production.

According to an established provisional schedule, liquid CO2 samples are usually sent to a central laboratory in charge of carrying out the same tests but in a non-industrial environment and with a view to inter-comparison of the various gas production sites or on samples from third-party sources, with the main drawbacks:

• An environment and people different from the industrial site

• A time lag between the date of sampling and the date of the laboratory test (potentially several weeks)

The method according to the invention makes it possible to implement a metrologically demonstrable staff qualification system and to characterize the more specific degrees of perception of the actors in this analysis process.

In other words, the method makes it possible to ensure the reproducibility of the test over time and which does not depend on the person chosen within the laboratory. Indeed each person has specific sensory capacities, and not necessarily constant over time:

• A test can be positive by one person and negative by another

• A person can, several weeks/months/years apart, have an altered or exacerbated perception

Third example of application: monitoring during the delivery of CO2 to a food industry site, which uses this CO2.

Currently, with each delivery, before each unloading, a sensory test is carried out on site by the personnel of the site delivered in gas, in order to authorize the transfer and use of the delivered product. The main pitfall is the possible difference in perception between the operating/material/personal modes of the user site and the company delivering the gas, which, as a precautionary principle requires, systematically ends with the validation or otherwise of the delivered site. The risk is therefore both economic and commercial.

The method according to the invention allows the extension of the test on the containers (tank, road tank, tank, evaporator, food user network) in order to optimize the monitoring of the entire distribution chain. The present invention therefore relates to a process for the sensory analysis of industrial gas samples delivered to an industrial consumer site, particularly in the food or pharmaceutical industry, and in particular CO2, characterized by the implementation of the following measures:

- we have a hermetic enclosure in which is placed a detection system, detection system which comprises one or each of the following sensor systems: i) one or more electrochemical cells, cells which are each of them specific and which react to a given molecule such as SO2, H2S, or even NH3; j) one or more MOS (Metallic Oxides Semiconductor) type sensors;

- the gas sample, for example CO2 , to be evaluated, is introduced into the enclosure in the form of a gas stream (in a gaseous form or in the form of solid dry ice in the case of CO2).

- the cells and/or sensors present in the enclosure will react in the presence of organic compounds possibly present in the gas sample, and provide a detection response;

- the enclosure is advantageously fitted with a system of pressure, temperature and hygrometry sensors to enable the operating conditions of the analysis to be controlled;

- a data acquisition and processing system is available;

- we carried out beforehand, a learning phase of the analysis, based on a reference method implementing human noses, learning where we carry out with a representative panel of people a series of blind olfactory test on the sample in question;

- using said data acquisition and processing system, an operation for reconciling the two types of data is carried out, i.e. the results of the tests carried out by the human panel and the results provided by the sensors and/or cells in the enclosure.

An example of implementation of the invention is described below. - The sample evaluated is made up of 99.9999% ultra-pure CO2 but which sometimes may contain ppm or ppb of one or more chemical molecules which have a very strong odor power (for example, it is said that the human nose would be able to detect about 5 ppb of H2S). - A human panel made up of ordinary people and personnel qualified for this type of analysis, makes it possible to provide a statistical response (representative survey of the population) on the characterization of odors and their intensity so that when the machine will be subjected to exactly the same test then his answer will be correlated to the answers of the panel and memorized as such.

- The enclosure and its devices reproduce in a way the hands that close around the nose when you want to smell an "aroma" without being disturbed by the environment.

- Subsequently, when in the conditions of operational use, the machine will detect an odor, it will be able to translate this odor directly in the form of the representation of the image which will have been programmed for it by the human (fruit odor, smell of earth, etc...) and not a chemical formula or an obscure machine language.

Claims

9

CLAIMS Process for the sensory analysis of industrial gas samples delivered to an industrial consumer site, particularly in the food or pharmaceutical industry, and in particular CO2, characterized by the implementation of the following measures:

- we have a hermetic enclosure in which is placed a detection system, detection system which comprises one or each of the following sensor systems: i) one or more electrochemical cells, cells which are each of them specific and which react to a given molecule such as SO2, H2S, or even NH3; j) one or more sensors of the MOS (“Metallic Oxides Semiconductor”) type;

- the gas sample, for example CO2, to be evaluated is introduced into the enclosure, in the form of a gas stream, or in a gaseous form or in the form of solid dry ice in the case of CO2.

- the cells and/or sensors present in the enclosure will react in the presence of organic compounds possibly present in the sample, and provide a detection response;

- a data acquisition and processing system is available;

- we carried out beforehand a learning phase of the analysis, based on a reference method implementing human noses, learning where we carry out with a panel of people a series of olfactory tests blindly on the sample considered;

- using said data acquisition and processing system, an operation for reconciling the two types of data is carried out, i.e. the results of the tests carried out by the human panel and the results provided by the sensors and/or cells in the enclosure.