METHOD FOR MULTIPHYSICAL CHARACTERIZATION OF ELECTROCHEMICAL OR PHOTOELECTROCHEMICAL CELL STACKS

Abstract

A multiphysics characterization method for stacks of electrochemical or photoelectrochemical cells, sites of endergonic or exergonic reactions, of finite or infinite capacity, intended to determine qualitatively or quantitatively the level of performance homogeneity of each cell placed in the stack, and consisting of: - operating the cell stack preferentially in steady state, - sampling one (or more) characteristic physical property of a mechanical, electrical, electrochemical, chemical, fluidic or thermal type, for a duration dependent on the sampling rate in order to obtain samples of sufficient size to ensure the robustness of the measurement, - converting the sample acquired on each cell at the end of the test into a distribution representing the number of occurrences of the measured property in each of the chosen classes.- to compare the distributions of each cell with each other at a given moment in order to diagnose the homogeneity of the stack, or to compare the distributions of each cell measured at different times during the reactor's life in order to identify degradation processes and identify cells in a phase of rapid degradation.

Description

Multiphysics characterization method for stacks of electrochemical or photoelectrochemical cells Background and State of the Art The present invention relates to electrochemical or photoelectrochemical reactors in the broadest sense. In this document, the term electrochemical reactor generally refers to a reactor containing at least one elementary electrochemical cell or any stack (at least one) of elementary electrochemical cells. An elementary electrochemical cell contains at least one galvanic chain, that is, in the simplest case, the series connection of three electrically conductive phases (a phase is a material in the Gibbs sense and in the sense of systems thermodynamics, i.e., an isotropic space whose physical and chemical properties are identical at every point) with different charge carriers: the first phase is generally an electronically conductive metallic material,The first layer may be modified at its extreme surface by one or more layers with electrocatalytic properties; the second is generally an electrolytic solution, solid or liquid, that is, a solvent in which one or more electrolytes are dissolved, often one or two redox couples, within which the transport of electrical charges is ensured by ions via an ionic conduction mechanism; the third is generally a metallic material with electronic conduction, modified at its extreme surface by one or more layers with electrocatalytic properties. A simple elementary electrochemical cell therefore comprises at least two metal-electrolyte interfaces. It generally includes other additional cell components. For example, in some cells,A separator is inserted in the middle of the interpolar space: it ensures ionic transport while preventing direct contact between the reactants or reaction products of each electrode-electrolyte interface. In practice, the basic electrochemical cell consists of any number of components (for example, and not exhaustively, an electrolytic solution (often incorrectly called an electrolyte), a first electrode called the anode (porous or non-porous), a second electrode called the cathode (porous or non-porous), one or more spacers, one or more current distributors, one or more current collectors, one or more bipolar plates), that is to say, a set of materials, each having a shape, thickness, size, and weight that varies according to the technology and the intended application. In this document, the term photoelectrochemical reactor generally refers to an electrochemical reactor containing at least one elementary photoelectrochemical cell. An elementary photoelectrochemical cell comprises at least one galvanic chain in which at least one (or both) of the two electrodes is replaced by a photoelectrode. The photoelectrochemical reactor may also consist of any (at least one) stack of elementary photoelectrochemical cells. Like an elementary electrochemical cell, an elementary photoelectrochemical cell also consists of any number of additional components. However, it differs from the elementary electrochemical cell in that it contains at least one, or even two, photoelectrodes. The term photoelectrode is generally a semiconductor material in contact with an electrolyte.The photoelectrode is exposed to a natural or artificial light source. The semiconductor material that constitutes it, characterized by its band gap energy, absorbs a greater or lesser portion of this light flux and generates electron-hole pairs in situ. These pairs act as charge carriers and can produce redox reactions when the semiconductor is brought into contact with an electrolytic solution. These charge carriers of opposite electrical charge are separated (to prevent their direct recombination) by migration under the influence of the electric field that develops upon contact with an electrolytic solution. There is a wide variety of electrochemical reactors and photoelectrochemical reactors. These can be classified into two main families. The first family of reactors comprises those that carry out so-called endergonic, or non-spontaneous, transformations. These reactors are used to perform a non-spontaneous chemical reaction by supplying external electrical energy (current). The electrochemical reactor in this family is called an electrolyzer, and the photoelectrochemical reactor in this family is called a photolyzer. The hybrid form between an electrolyzer and a photolyzer is called a photo-assisted electrolyzer or an electro-assisted photolyzer, depending on the degree of hybridization.The reactors in this family use a direct current power supply, whether it operates in voltage regulation (we then speak of potensiostatic mode if the reactor is supplied by a constant voltage invariant over time or potentiodynamic if the reactor is supplied by a voltage variable over time), in current regulation (we then speak of intensiostatic mode if the reactor is supplied by a constant current invariant over time or intensiodynamic. (if the reactor is powered by a time-varying current) or by power regulation (referred to as constant or variable power supply). An example of an electrolyzer is a water electrolyzer, for liquid or vapor water, which transforms water (the reactant) into molecular hydrogen and oxygen (reaction products) in a single cell or a stack of cells by supplying electrical energy using a direct current generator. But many others exist, for example, chlor-sodium electrolyzers, electrolyzers for aluminum production, and for the production of alkali or alkaline earth metals.An example of a photolyzer is a water photolyzer (or water photoelectrochemical dissociation cell), liquid or vapor, which transforms water (reactant) into molecular hydrogen and oxygen (reaction products) in a single cell or a stack of cells by supplying light and / or electrical energy using a natural or artificial light source and / or a direct current generator. An example of a hybrid reactor is a hybrid reactor combining a water electrolyzer and a water photolyzer. The second family of reactors comprises those that undergo so-called exergonic, or spontaneous, transformations. These reactors are used to produce electrical energy from a spontaneous chemical reaction. The electrochemical reactor in this family is called a direct current generator, and the photoelectrochemical reactor in this family is called a specific type of cell, for example, a dye-sensitized solar cell. Reactors in this family are powered by chemical reactants which, instead of reacting directly with each other through contact, exchange electrons via an external electrical circuit. A distinction is made between batteries (non-rechargeable) and accumulators or accumulator batteries (rechargeable). The first known example is Volta's pile. A very wide variety exists (for example, lead-acid and lithium batteries).It is worth noting that a rechargeable battery, when charging, corresponds to an electrochemical reactor of the first family. During charging, the current is controlled by the power supply (potensiostatic, potentiodynamic, intensiostatic, or intensiodynamic mode). During discharging, the flow rate of the cell or battery is controlled by adjusting the impedance of the electrical load (generally a resistor or an electric motor). The flow rate can be regulated by voltage (this is called potensiostatic mode if the reactor delivers a constant voltage that remains constant over time, or potentiodynamic mode if the reactor delivers a voltage that varies over time), or by current regulation (this is called intensiostatic mode if the reactor delivers a constant current that remains constant over time). Intensiodynamics (if the reactor delivers a current that varies over time) or power regulation (referred to as constant or variable power flow). A modern example of a fuel cell is the hydrogen-oxygen fuel cell, which transforms hydrogen and oxygen (reactants) into water (the reaction product, liquid or vapor depending on the operating temperature and pressure) in a single cell or a stack of cells, and produces electrical energy in the form of direct current or voltage. An example of a photoelectrochemical reactor that converts light energy into electrical energy is provided by dye-sensitized cells. Electrochemical and photoelectrochemical reactors are open systems that exchange energy or matter with the surrounding environment. However, a second useful distinction for classifying reactors is whether the active chemical material is stored inside or outside the reactor. The amount of available active material defines the reactor's electrochemical capacity. Generally, when the active material is stored inside the reactor, the capacity is lower than when it is stored outside because, in the latter case, the reservoir size can be very large, even infinite. An example of a reactor in the first category, with active material stored inside the reactor, is a charging battery. An example of a reactor in the first category, with active material stored outside the reactor, is a water electrolyzer.An example of a second-family reactor with active material stored inside the reactor is a discharged storage battery. An example of a second-family reactor with active material stored outside the reactor is a fuel cell or a redox battery. An example of an infinitely large reservoir is the Earth's atmosphere, which can store oxygen, or an ancient aquifer, which can store a large quantity of hydrogen under pressure. From a technological standpoint, there are three main types of elementary electrochemical or photoelectrochemical cells: cells in which the electrode-separator or photoelectrode-separator distance is non-zero (known in the scientific and technical literature as gap cells); cells in which the electrode-separator or photoelectrode-separator distance is zero (zero-gap cells); and solid electrolyte cells (SPE cells). An elementary electrochemical or photoelectrochemical cell operates over a variable temperature and pressure range. It uses an electrolyte with a variable pH, whether liquid, solid, or gaseous, and consumes liquid chemical reagents. Solid, gaseous, or in solution, and produces liquid, solid, gaseous, or solution chemical reaction products. Several elementary electrochemical or photoelectrochemical cells can be interconnected to adapt the reactor size to the needs of the intended application: this extreme modularity is a characteristic of the electrochemical or photoelectrochemical reactor. There are different modes of electrical interconnection (series, parallel, compact stacking, with central anode or cathode), fluidic interconnection (series or parallel supply of reactants and collection of reaction products), and thermal interconnection (cooling method using a heat transfer fluid). It should be noted that the term "stack" refers to a number of individual cells ranging from one to a large number, N. For example, in the case of water electrolyzers, N is currently around 200. Interconnecting several stacks can increase the value of N to a thousand or even more. It should also be noted that fluids circulate in some operating reactors, particularly those that use external tanks to store the active chemical material. For example, in a water electrolyzer, liquid water or vapor stored outside the electrolyzer is injected into the inlet of each cell, and two-phase liquid-gas mixtures (water-hydrogen in the cathodic circuit and water-oxygen in the anodic circuit) or gas mixtures (water-hydrogen vapor in the cathodic circuit and water-oxygen vapor in the anodic circuit) are collected at the outlet of these cells.Another example is in fuel cells, where hydrogen is introduced at the inlet of the so-called "negative" cells and oxygen or air at the inlet of the so-called "positive" cells. To prevent leaks, the individual cells must be sealed. This sealing can be achieved in various ways: for example, by placing a polymer seal around the periphery of the cells. The seal is then obtained by axially compressing all the cells contained in the stack. The mechanical properties of the stack are dictated by the individual mechanical properties of each cell. It is desirable to obtain a homogeneous elastic deformation to avoid undesired displacements of the internals during tightening (for example, but not limited to, accordion effects, rotation, or displacement of the internal components).Poor control of stack mechanics can lead to electrical, electrochemical, chemical, fluidic, and thermal malfunctions of varying degrees, the negative effects of which will be felt more or less quickly (poor performance, rapid performance loss). Due to the series interconnection of the elementary cells in a stack, it only takes one... If any cell in the stack experiences degraded performance (beyond a critical threshold specific to each reactor type), the stack becomes unusable. It should be noted that similar situations are encountered in reactors that use internal tanks to store the active chemical material. However, there is no reliable technique for qualitatively or quantitatively measuring the degree of mechanical, electrical, electrochemical, chemical, fluidic, and thermal homogeneity of these stacks. Limitations of the State of the Art: Those skilled in the art understand that as the surface area of ​​the elementary cell increases and the number N of elementary cells in the stack increases, the challenge lies in obtaining the most homogeneous stacks possible, whether the assembly of the N cells is performed manually or automatically using dedicated assembly tools. For proper and sustainable operation, electrochemical and photoelectrochemical reactors require multi-physics optimization: the homogeneity criteria concern mechanical aspects (distribution of pressure forces or compression field, characteristic dimensions of the components), electrical, electrochemical, chemical, fluidic, and thermal aspects. In the state of the art, homogeneity is assumed to be achieved by stacking elementary cells of the same type, containing the same components, and having the same geometric dimensions.In reality, components do not all have strictly identical dimensions. The elastic deformation of each cell during compression leads to various deformations which, even when small, can have significant consequences. Compressing several elementary cells can lead to substantial heterogeneity in properties from one elementary cell to another, even when the assembly process is automated. This is because each cell possesses its own mechanical characteristics (e.g., yield strength). There may be displacements and compressive deformations that exceed the yield strength of some components, altering their mechanical behavior and potentially affecting their corrosion resistance. As a matter of current practice, the homogeneity of an assembly of N elementary cells is generally assumed to be achieved by performing various types of tests.An example of a test is the leak test (e.g., with air, helium or water). As long as at least one cell of the stack leaks, the pressure applied on both sides of the stack to ensure the seal of the assembly is gradually increased until there is no more leak. of leaks. Over-compressing a set of N elementary cells can lead to inhomogeneous mechanical deformations, more or less pronounced, possibly exceeding the elastic limit of the components, on any cell of the stack, sometimes on those near the edges, or in the middle, sometimes on several cells, whether adjacent or not. The consequence is that when the reactor thus assembled is put into electrochemical operation, not all the elementary cells of the reactor function identically and homogeneously. For example, the fluid flow rate through each elementary cell may differ from one cell to another, or the current flowing transversely across the stack may be different at different points on the surface. This can lead to more or less pronounced local heating, and is detrimental to the performance and lifespan of the reactor as a whole.Another example of a test is the impedance test. The electrochemical impedance of each cell is measured over a wide frequency range (typically between 100 kHz and 10 MHz at a rate of 8 or 10 points per decade), at different current densities. The disadvantage of this type of test is that it requires expensive, specialized equipment, is time-consuming, is limited to low current densities (especially on large reactors), and is poorly suited for characterizing individual cells operating under non-faradic conditions (in the absence of current). Need beyond the state of the art The expert therefore needs robust and reliable tools enabling him, at different stages of the manufacturing and assembly process of individual cells or during the operation of the reactor made up of several assembled cells, to make a diagnosis on the homogeneity of the performance of each elementary cell present in any stack. The tools in question must meet a set of technical and economic criteria: 1. In practice, the homogeneity diagnosis is carried out through measurement individual of a significant physical property (mechanical, electrical, electrochemical, chemical, fluidic, thermal) must be non-intrusive, non disruptive, in situ, easy to implement, requiring simple equipment, non- specific, inexpensive. The method proposed by the invention makes this possible. 2. The homogeneity diagnosis must be able to be performed during operation of the reactor (in operando), without interfering with its normal operation. 3. The diagnosis of homogeneity must be able to be carried out either externally by means of to the use of specific equipment, but preferably directly during the process of reactor operation, using the instrumental resources already in place in the balance-of-plant, if possible based on a treatment carried out in time actual by the process monitoring system, without any consequences in terms of safety, processing time and cost. 4. The homogeneity diagnosis must be quick, simple, easy to export or to communicate to operators or supervisors, clearly and explicitly, easy to interpret for the user, small in size (amount of data to acquire to wear a diagnostic and quantity of data, analog or digital, contained in the function that allows for diagnosis). 5. Data analysis must be possible over time (using tools (analysis of the fluctuations of a stationary physical property as a function of time) or in the frequency space (e.g., Fourier) using the tools of classic transformations such as the fast Fourier transform. The object of the invention Specifically, the present invention concerns a technical method for characterization multiphysics of individual electrochemical or photoelectrochemical cells or stacks of electrochemical or photoelectrochemical cells. The method allows for the qualitative or quantitative determination of the level of homogeneity of performance in operation of individual electrochemical or photoelectrochemical cells, whether they are the site of endergonic or exergonic reactions and whether they possess a finite or infinite capacity for storing the active material, when they are placed in stacks having any number (at least one) of individual cells. In another embodiment the process is characterized in that the homogeneity criterion used is a characteristic physical quantity of a mechanical, electrical, electrochemical, chemical, fluidic or thermal nature, or any combination of one or more of these characteristic physical quantities. In another embodiment the process is characterized in that the characteristic physical quantity is measured on a single individual cell or on any subset of individual cells, adjacent or non-adjacent, or on all the cells of the stack. In another embodiment, the process is characterized by a sequence of operations consisting of: - to operate the reactor under steady-state conditions, that is, at constant and time-invariant flow rates of matter and energy, to use a metrological chain including a sampler comprising a analog-to-digital converter per characteristic physical quantity and per individual cell, and a scanning automaton or multiplexer, and comprising a specific sensor per characteristic physical quantity, - to sample one (or more) characteristic physical quantity, each with a a characteristic sampling frequency and for a characteristic duration so as to obtain statistically robust samples consisting of a large number of points, at least one hundred, and preferably several thousand, - to convert the sample(s) thus obtained into distributions, that is to say by defining a set of classes, each class being characterized by a lower bound, an upper bound and an interval, and to count the number of occurrences of the characteristic physical quantity in each of the classes during the duration of the test, - to compare the distributions of each individual cell or any combination of individual cells of a stack in the form of graphs so as to be able to perform a qualitative or quantitative diagnosis on the homogeneity of the stack. In another embodiment, the process is characterized in that the characteristic physical quantity is measured when the reactor operates under steady-state conditions, i.e., in a steady state characterized by time-independent operating variables, such as: - from an electrical point of view, reactor operation in potentiostatic mode or intensiostatic mode, - From a fluidic point of view, the reactor operates with a steady flow rate of fluids circulating in the different parts of the reactor. - from a thermal point of view, operation of the reactor with stationary inlet and outlet temperatures of the fluids circulating in the cell stacks and in the different parts of the reactor, - from a chemical point of view, operation of the reactor with a stationary chemical composition of the fluids in the different parts of the reactor, - from a mechanical point of view, operation of the reactor with a stationary pressure of the fluids circulating in the different parts, - from a mechanical point of view, operation of the reactor with a stationary pressure or force or clamping torque applied to the cell stack. In another embodiment, the process is characterized in that the characteristic physical quantity measured is: - either an individual cell electrical voltage, - or the current flowing through the cell stack, - or the individual fluid inlet temperature inside an individual cell, - or the average fluid inlet temperature inside the cell stack, - or the individual fluid outlet temperature of an individual cell, - or the average fluid outlet temperature inside the cell stack, - or the fluid inlet pressure inside an individual cell, - or the fluid inlet pressure inside the cell stack, - or the fluid outlet pressure of an individual cell, - or the fluid outlet pressure of the cell stack, - or the fluid inlet flow rate inside an individual cell.- either the fluid inlet flow rate within the cell stack, - or the fluid outlet flow rate of an individual cell, - or the fluid outlet flow rate of the cell stack, - or the chemical composition of the fluid inlet within an individual cell, - or the chemical composition of the fluid inlet within the cell stack, - or the chemical composition of the fluid outlet of an individual cell, - or the chemical composition of the fluid outlet of the cell stack, - or the clamping force applied to the cell stack, - or the clamping torque applied to the cell stack, - or the clamping pressure applied to the cell stack, - or any combination of these characteristic physical quantities. In another embodiment, the method is characterized in that the characteristic physical quantity is sampled so as to obtain a sufficient number of points to ensure the robustness of the measurement, characterized by: - ​​each point being measured using a metrological measurement chain having an accuracy on the order of one percent or more, - each point being measured as synchronously as possible using either an analog-to-digital converter per cell or a multiplexer, the synchronization criterion depending on the choice of the sampling rate, - each sample contains several hundred points and preferably several thousand, typically ten thousand, - the duration of the sampling test and the sampling rate are chosen so as to obtain the sample having the satisfactory number of points.Preferably, but not exclusively, one point every 100 milliseconds for 100 seconds, or one point every 10 seconds for 28 hours. In another embodiment the process is characterized in that the sample of the characteristic physical quantity is converted into a distribution characterized by: - ​​a set of identical or different classes, of values ​​defined each by a characteristic interval, a minimum value, a maximum value, a number of classes and an interval value of the characteristic physical quantity, - a number of occurrences corresponding to the number of times the characteristic physical quantity took the value of the considered interval in the class during the duration of the diagnostic test. In another embodiment, the method is characterized in that the distributions of each individual cell within the stack or of any subset of individual cells, obtained from the measurement of a characteristic physical quantity, are compared by plotting not the entire distribution but a particular value of each distribution, as a function of the cell number in the stack, this particular value being: - either the initial value of the distribution, - or the maximum value of the distribution, - or the final value of the distribution. - either the mean value of the distribution, - or the standard deviation of the distribution, - or any other characteristic value of the distribution chosen by the user. In another embodiment, the process is characterized in that the distributions of each individual cell within the stack or of any subset of individual cells, obtained from the measurement of a characteristic physical quantity, are measured at regular or irregular time intervals and used to perform any mathematical processing useful for the analysis of the information they contain, in particular to produce Cartesian graphs characterized by: - ​​a two-dimensional representation consisting of plotting on the same graph the distributions with the occurrence on the ordinate and the classes on the abscissa, or vice versa, - a three-dimensional representation consisting of plotting on the same graph the distributions with the occurrence on the ordinate, the classes on the abscissa and on the third axis the number of each cell in the stack, or any permutation of axes. The invention applies for example to the multiphysics characterization of fuel cells, electrolyzers of any type or accumulator batteries made up of one or more elementary cells, that is to say without limitation of the number of elementary cells, of nature, of type of constituent materials, of size and of shape (for example planar, cylindrical or spherical). When several elementary cells are interconnected, for example by stacking them in series (to form what is known as a stack), in order to size these reactors according to the intended applications (choice of chemical production rate or electrical power consumption in the case of reactors of the first type, and electrical power output in the case of reactors of the second type), the challenge lies in manufacturing and then stacking individual cells with the same properties and performance. Ideally, all the cells in the same stack have the same characteristics and the same multiphysics responses (mechanical, electrical, electrochemical, fluidic, and thermal). The engineer's role is to produce reactors that come as close as possible to this ideal case.But in real systems, there are dispersions, fluctuations: the characteristics and multiphysics responses are not identical, whether the elementary cells are. manufactured and / or assembled manually or automatically. In the case of electrolyzers, these fluctuations can be due to the type of power supply used (for example, switch-mode power supplies generate Gaussian current distributions). In the case of discharge generators, they can be due to fluctuations in the impedance of the flow load. But more importantly, these fluctuations vary from one cell to another depending on the degree of control over the manufacturing and assembly processes of the individual components. The higher the machining precision of the cell components, the greater the homogeneity of the stack, but also the higher the cost of the cell. The expert seeks a compromise between cost and performance.The stacking of elementary cells with non-homogeneous multi-physical characteristics (mechanical, electrical, electrochemical, chemical, fluidic, thermal) has several consequences. In particular, when the reactor is operating, the mechanical, electrical, electrochemical, chemical, fluidic, and thermal properties fluctuate to varying degrees around a pivot value (for example, an average value). Pivot values ​​can be very different from one cell to another, reflecting significant heterogeneity in the stacking. This leads to degraded efficiency (energy or faradaic), more or less accelerated aging of individual cells, which limits the overall lifespan and necessitates more or less frequent, more or less costly, repair and maintenance operations. Therefore, there is a need for the researcher, manufacturer or user of this type of reactor to have technical processes for multiphysics characterization (mechanical, electrical, electrochemical, fluidic chemical, thermal) of elementary electrochemical or photoelectrochemical cells or of stacking elementary electrochemical or photoelectrochemical cells.The invention is therefore of interest in the fields of research and development (allowing the optimization of the design of elementary cells and stacks), in assisting in the design of automated manufacturing and assembly tools for elementary cells, in assisting in the adjustment of the mechanical clamping applied to the stack in order to ensure its sealing and homogeneity, in quality control during manufacturing (manual or automatic) of elementary cells and stacks, in assisting with the activation phases and testing of stacks (indeed, it is often necessary to verify the proper functioning of stacks from the production lines before selling them to customers and to guide the operator in their activation procedure), in diagnostics and remote diagnostics on elementary cells and stacks. stacking, in the management (forecasting, anticipation) of maintenance operations, in the monitoring of reactors on customer site, in variable industrial or consumer environments and the management of safety associated with the operation of these reactors, whether the power is low or high (there are for example electrolysis units of several MW, several tens of MW, several hundred MW or even more), in the comparison of technologies and manufacturers of these technologies, and in the qualification of these manufacturers (of their technology, their manufacturing processes, their teams) to access specific markets. To this end, the invention makes it possible to obtain key qualitative and quantitative information on the degree of homogeneity, whether close to or far from an ideal or reference state, of a set of key physical properties (mechanical, electrical, electrochemical, chemical, fluidic, thermal). This key information is obtained by sampling, using a suitable measurement chain, one or more physical properties characteristic of the stack. It is then processed digitally to facilitate comparison, from a qualitative or quantitative point of view. For example, the invention makes it possible to obtain a 3D visualization of the individual performance of each elementary cell located anywhere in any stack of elementary cells. In practice, the invention relies on the acquisition of one or more physical quantities characteristic of the stack (mechanical, electrical, electrochemical, chemical, fluidic, thermal) and their processing, in order to make various types of diagnoses (scientific, technical, technological, economic) on the mechanical, electrical, electrochemical, chemical, fluidic, and thermal aspects. These diagnoses can also be used for comparing electrochemical or photoelectrochemical technologies of the same or different nature, produced by the same manufacturing or research teams or different teams, for comparing offers from various suppliers, and for qualifying equipment according to its use. The invention is based on several guiding and innovative ideas: The first innovative idea stems from observing electrochemical reactors in operation. Homogeneity diagnosis is relevant, robust, and easy to perform when carried out on a reactor in steady-state operation, which is not intuitive because steady-state operation seems to provide little information. This is because, in reactors of this type, the characteristic physical quantities (for example, and not exhaustively, the individual voltage of each elementary cell in a stack of elementary cells, or the temperature of the fluids entering or leaving the individual cells or reactors, or the pressure of the fluids entering or leaving the individual cells or reactors) are not completely constant but fluctuate over time around pivot values.Even with great experimental precautions (thermostatically controlled reactor, potensiostatic or intensiostatic power supply controlled by a precision generator), these quantities fluctuate because these are complex, multiphysics systems with strong coupling between characteristic quantities. It should be noted that individually homogeneous cells can be interconnected in a non-homogeneous manner, leading to heterogeneous operation. The differences in fluctuation from one elementary cell to another depend directly on the homogeneity of the environment (mechanical, electrical, electrochemical, chemical, fluidic, thermal) from one cell to another. Observation has shown, and this is what led to the invention, that the characteristic measurements providing diagnostic information must be carried out on the reactor operating in steady state, which implies the implementation of specific procedures to perform these diagnostics.The information carried by the fluctuations of each cell provides limited information about that cell's characteristics because these characteristics are modified by the fluctuations of surrounding cells. The measured information is not an absolute characteristic of the cell but a characteristic of the cell that includes the characteristics of surrounding cells. We were thus able to observe that these interactions between cells occur at greater or lesser distances. The characteristics of a cell are convolved with those of neighboring cells, with a decreasing contribution with increasing distance. However, it is likely that in stacks of a few hundred cells, the first cell in the stack still interacts with the last despite the distance separating them. The important information for making a useful homogeneity diagnosis is therefore not contained in the characteristics of each fluctuation (measured via...).an average value, a standard deviation, or other statistical quantities) but in the comparison of fluctuations from one cell to another. It follows from these various considerations that the relevant test leading to the diagnosis must be carried out under the most stationary conditions possible (searching for conditions leading to invariance over time or the smallest possible fluctuations), in an established fluctuation mode, regardless of the reactor's usual operating mode, so that the comparison is possible and facilitated. 2.The second innovative idea stems from the observation of physical quantities measured on electrochemical reactors in steady-state operation: the individual measurement of a significant physical property (mechanical, electrical, electrochemical, chemical, fluidic, thermal) of an individual cell or a group of individual cells (or a subset, adjacent or not, of individual cells in a stack of individual cells) at a given time t is insufficient to provide a reliable and informed diagnosis, due to the coupling between cells. Similarly, measuring the average value of a set of measurements does not provide sufficient information.These measurements or averages must be replaced by distributions (calculated from a large sample size, obtained through periodic measurements over a longer or shorter period depending on the sampling rate used, and with appropriate precision) in order to make the observable more robust. It should be noted that the concept of distribution is only meaningful when measurements are performed on a reactor operating under globally stationary or at least periodic conditions: periodicity distorts the morphology of the distributions, but since each elementary cell operates under the same conditions as the others, comparing distributions of arbitrary morphologies (i.e., non-symmetrical, non-Gaussian) still provides useful information on the homogeneity of the stacks.Therefore, in this type of reactor, the shape of the distributions (not necessarily centered around a pivot value) provides the relevant information for establishing the desired homogeneity diagnosis. 3. The third innovative idea is that the measurement of the distributions of interest is modular: it can be performed on cells containing (or not) all the key elements of the stack. For example, although it is not intuitive, the electrolyte layer in the individual cells can be replaced by a non-conductive plastic film of the same thickness. By applying a potential difference across the terminals of such a cell stack, a [missing information] appears. The electrical potential difference across each individual cell, which then behaves like a charged capacitor, is a key component. Even in the absence of a key component in each elementary cell, it is still possible to assess the homogeneity of the cells in the stack. For example, when an electrical potential difference is applied to the terminals of a cell stack, each individual cell develops an individual potential difference across its terminals, the sum of which equals the total potential difference. In this case, measuring the electrical voltage of each cell allows us to test the homogeneity of each cell by focusing the test on a particular cell component, without having to integrate into the cells all the components necessary for the actual operation of the reactor.This is useful when one wants to selectively and progressively optimize a particular component or subset of components, or when experimenting and seeking to optimize certain inexpensive components before integrating the most complex and / or expensive component into the cells. The fourth innovative idea is that the measurement of certain interesting properties (e.g., electrical and electrochemical) can be carried out during the compression of the stack, in order to find the optimal compression ratio—that is, the one that does not compromise the optimum of the other key characteristics of the stack, which is never easy to achieve. The proposed innovative approach also makes it possible to verify the homogeneity of the mechanical compression from one cell to another by analyzing the stationary distributions derived from electrical measurements, based on the multi-physics coupling of the characteristic quantities. The implementation of the invention allows for a precise diagnosis on The stacking of elementary cells requires a series of successive steps: 1. Defining the technical means implemented: this step allows us to define the tools necessary to carry out the test on which the diagnosis will be based (choice of sensors used, characteristics in terms of precision and operating conditions, type and characteristics of the sampler, connection method on the stack of elementary cells). 2. Definition of operating conditions: This step defines the stationary conditions under which the test must be performed to arrive at a relevant diagnosis. Several physical parameters must be set for the test to run correctly; their specific values ​​depend on the user's needs. In particular, the nature of the fluid in the cells (air or water); the flow rate of this fluid (stationary or circulating fluid); the test temperature (the temperature of the atmosphere surrounding the stack or the temperature of the fluid in the cells; preferably, there should be thermostating of the circulating fluids or the surrounding atmosphere, but this is not essential); the fluid pressure in the cells; and the operating mode (potensiostatic or intensiostatic). 3.Definition of sampling conditions: choice of the physical quantity to be sampled (in a preferred mode of the invention, the quantity to be monitored is the individual electrical voltage in volts of each elementary cell), sampling rate, sampling duration, real-time or delayed processing of the acquired data, and storage means and methods. Typically, a sample of 10,000 points provides the necessary robustness for the observable. This corresponds, for example, to a measurement every 10 milliseconds for 100 seconds or to a measurement every 10 seconds for approximately 28 hours. 4. Performing the test to acquire the samples necessary for constructing the diagnostic.The test can be more or less rapid depending on the needs (between one minute and several tens of hours); this simply requires choosing a sampling rate that is faster or slower depending on the characteristics of the equipment used. However, the sample must be of sufficient size (for example, ten thousand voltage measurements per cell) to produce distributions that have physical meaning and provide a sound diagnosis. An example of rapid sampling is measuring cell voltage every 10 milliseconds or less; a typical sampling rate is one point every second (or every 10-60 seconds). An example of slow sampling is measuring voltage every minute or longer. 5. Sample processing: each sample is converted into a distribution.A distribution is a mathematical function in which the space (minimum value to maximum value of the measured quantity) is subdivided into classes, each class containing an integer equal to the number of times the physical quantity being measured took the particular value of the interval considered during the time. of the test. The samples are digitally transformed into distributions by real-time or delayed processing; the choice of classes (lower bound, upper bound, interval) used for calculating the distributions depends on the quantity being monitored. For example, when dealing with a distribution associated with the measurement of the individual voltage of a cell in the case of a water electrolyzer, the classes can range from zero to 2 volts in steps of 1 millivolt. It should be noted that a continuous sample obtained by sampling over long periods can be processed using sliding windows. It should also be noted that the acquired samples do not need to be stored in memory (RAM or on storage media); they can be converted into distributions as they are acquired. 6.Graphical or mathematical processing of the distributions obtained cell by cell: for example, the 2D or 3D plotting of the distributions from the test provides visual information (supporting a qualitative diagnosis) or, through digital post-processing, a quantitative measurement of the differences between distributions for a quantitative analysis of the stack's homogeneity. 7. Analysis of distributions, interpretation, and diagnosis: based on the plots and analyses from the previous step, the engineer can diagnose the quality of the stack. A comparison of the results with a reference result (for example, a close alignment within a few millivolts of the distributions of each cell) allows, in relative terms, either a comparison (in terms of homogeneity or performance) or the observation of a change. 8.Analysis of the temporal evolution of distributions: this analysis makes it possible to track the evolution of the individual performance of each elementary cell in a stack of elementary cells, to observe any degradation processes, to identify cells with problems, and to extrapolate to determine the time remaining before a maintenance operation, according to criteria (performance, safety, cost) specific to each user of the invention. It should be noted that samples of any given characteristic quantity associated with any given elementary cell of the stack do not necessarily have to be stored on mass storage. Any instantaneous measurement taken on an elementary cell can be processed in real time and used to increment the corresponding class by one. Only the distribution, which is built up as sampling progresses, must be preserved and memorized, so that a diagnosis can be made at the end of the test. It should be noted that the characteristic operating sequence of the invention can be repeated as many times as necessary, at different points in the reactor's life (at any time between the Beginning of Life (BoL) and the End of Life (EoL)), regardless of the operating conditions. The evolution of the distributions over time allows for the analysis of aging processes and the anticipation of maintenance operations. Depending on the reactor's operating conditions, more or less rapid aging can be observed. For example, an electrolyzer powered by a steady voltage or current will age more slowly (and therefore have a longer lifespan) than an electrolyzer powered by a non-steady voltage or current (for example, from a photovoltaic panel or a wind turbine). Examples. By way of example, we describe here how to characterize, using the present invention, the homogeneity of stacks of elementary cells in electrochemical reactors of the first family, water electrolyzers of PEM technology (this is an electrolyzer technology in which the electrolyte is a proton-conducting membrane, for example a perfluorosulfonated membrane). It should be noted that the method described in the invention is also applicable to characterize other types of water electrolyzer technology, for example alkaline electrolyzers (which are characterized by the use of a liquid electrolyte that allows the transfer of hydroxyl ions (OH-) from the cathode to the anode), and high-temperature electrolyzers, in which the electrolyte is a ceramic, but also any type of electrolyzer.The invention can be used to diagnose any type of electrochemical or photoelectrochemical cell stack. List of figures: Figure 1: Schematic representation of various types of stacking of individual electrochemical or photoelectrochemical cells: (a) monopolar configuration (cells connected in parallel). (b) bipolar configuration (cells connected in series), (c) bipolar configuration (series stacking with a central anode). Figure 2: Schematic cross-sectional view of a basic water electrolysis cell (PEM technology). 1 — proton-conducting polymer membrane; 2 / 2' — electrocatalytic layers; 3 / 3° — porous current distributors; 4 / 4” — spacers (titanium grids); 5 / 5' — bipolar plates; 6 — hydrated proton flow. (a): liquid water inlet; (b): liquid water inlet; (a'): liquid water + hydrogen gas biphasic mixture outlet; (b°): liquid water + oxygen gas biphasic mixture outlet. Figure 3: 3D graph (number of occurrences on the vertical axis of the ordinates, voltage classes in volts on the horizontal axis of the abscissas and number of the cell in the stack on the depth axis) of the distributions measured by applying a total voltage of 5 Volts (non-faradic conditions) to a stack of 12 PEM water electrolysis cells in which the AME of each cell has been replaced by a non-conductive PTFE film in the presence of stagnant air at 22°C. Figure 4: 3D graph (number of occurrences on the vertical axis of the ordinates, voltage classes in volts on the horizontal axis of the abscissas and number of the cell in the stack on the depth axis) of the distributions measured by applying a total voltage of 5 Volts (non-faradic conditions) to a stack of 12 PEM water electrolysis cells in which the AME of each cell has been replaced by a non-conductive PTFE film, in the presence of water circulating thermostatically controlled at 35°C. Figure 5: 2D graph (number of occurrences on the vertical ordinate axis, and voltage classes in volts on the horizontal abscissa axis) of the distributions measured by applying a total voltage of 5 volts (non-faradic conditions) to a stack of 12 PEM water electrolysis cells in the presence of water circulating thermostatically controlled at 45°C. The numbers indicated in the figure are those of the individual cells in the stack. Figure 6: 3D graph (number of occurrences on the vertical axis of the ordinates, voltage classes in volts on the horizontal axis of the abscissas and number of the cell in the stack on the depth axis) of the distributions measured by applying a total voltage of 5 Volts (non-faradic conditions) to a stack of 12 PEM water electrolysis cells in the presence of a water circulation thermostatically controlled at 45°C. Figure 7: 2D graph (number of occurrences on the vertical ordinate axis, and voltage classes in volts on the horizontal abscissa axis) of the distributions measured by applying a total voltage of 76 volts (faradic conditions) to a stack of 44 PEM water electrolysis cells in the presence of water circulating thermostatically controlled at 72°C. The numbers indicated in the figure are those of the individual cells in the stack. Figure 8: 3D graph (number of occurrences on the vertical axis of the ordinates, voltage classes in volts on the horizontal axis of the abscissas and number of the cell in the stack on the depth axis) of the distributions measured by applying a total voltage of 76 Volts (faradic conditions) to a stack of 44 PEM water electrolysis cells in the presence of a water circulation thermostatically controlled at 72°C. Figure 9: Quantitative analysis of the performance degradation of a 100-cell PEM water electrolyzer. The value on the ordinate is the difference between the cell voltage at maximum distribution after 12 months of operation and the cell voltage at maximum distribution at the start of operation. Figure 1 schematically illustrates some possible examples of interconnecting individual cells to form an electrochemical or photoelectrochemical reactor comprising several individual cells. Figure 2 shows a general example of a basic water electrolysis cell using PEM technology. The key (fragile and expensive) component of this type of cell is element number one, located in the center. This is the MEA (Membrane-Electrode Assembly), consisting of a planar, proton-conducting polymer membrane (generally 100 to 200 microns thick) on the surface of which two catalytic layers are deposited. In modern electrolyzers, several hundred basic cells of this type are connected in series according to configuration (b) in Figure 1. To ensure ! P. Millet, doctoral thesis, Preparation and optimization of electrode-membrane-electrode assemblies. Application to water electrolysis, National Polytechnic Institute of Grenoble, 1989. ? C. Rozain, P. Millet, Electrochemical characterization of polymer electrolyte membrane water electrolysis KES, ELECIOCEAN, Acta L3C (20 (4) Lo0-167.. For sealing (not shown in Figure 2), it is necessary to use gaskets (for example, but not limited to, flat, O-ring, or knife-edge gaskets) generally placed around the periphery of the cells. To ensure fluid circulation (not shown in Figure 2), injection channels are required at the cell inlet and collection channels at the cell outlet. The shape of these channels and their cross-sectional areas depend on the flow rate of the circulating fluid. When these gaskets are made of plastic, during the compression of a set of N elementary cells to form a stack, some undesirable mechanical deformations may obstruct these channels to a greater or lesser extent. This is a significant source of heterogeneity from one cell to another.This results in reduced fluid circulation during operation in cells with small cross-sections, and therefore different electrochemical performance. Cells exhibiting this type of problem must be identified as they are susceptible to premature aging. In the following examples, the specific physical property used for diagnosis is chosen, for illustrative purposes, to be the individual voltage in volts of each individual cell in the stack. Similar reasoning can be used to diagnose any type of characteristic property of the stack, whether at the level of an individual cell, a subset of individual cells (adjacent or not) in the stack, or the entire stack. When the individual voltage applied to each cell is less than the thermodynamic voltage at which current flows in the stack, the test is referred to as being performed under non-faradic conditions. When the individual voltage applied to each cell is greater than the thermodynamic voltage at which current flows in the stack, the test is referred to as being performed under faradic conditions.The advantage is that information obtained under non-faradic conditions provides insight into what will happen under faradic conditions without having to perform (irreversible) tests under faradic conditions. The invention therefore allows for the validation of an assembly before it is powered. Example | of use of the invention In this first example, the objective is to diagnose the mechanical and electrical homogeneity of the stack without risking damage to the key and expensive component of individual PEM electrolysis cells, the membrane-electrode assembly (AME). To this end, the AMEs of each individual cell are replaced by electrically non-conductive polymer films (e.g., PTFE-type films). Despite this modification, and somewhat unintuitively, the method proposed by the invention enabled a relevant diagnosis of the stack of elementary cells. Indeed, by applying a constant potential difference (potensiostatic mode) across the stack, despite the presence of insulating discs, each cell acquires, more or less rapidly (the speed at which these individual voltages reach a steady-state value depends on the surface oxidation state of the cell components, information that can also be used to diagnose the internal condition of a given elementary cell), a characteristic voltage that depends on its homogeneity.The physical quantity examined in this example is the individual electrical voltage of each elementary cell in a stack of elementary cells. No electrical current flows through the stack because of the PTFE resistive films, regardless of the total potential difference applied across the terminals: the breakdown voltages on this type of material are much higher than the typical operating voltages of the individual cells.The test allows, however, (i) to demonstrate that it is possible to make a useful diagnosis in the absence of AME; (ii) to show, by comparison with the results of the following examples, that the information obtained without AME is revealing of what happens when AMEs are installed in the cells; (iii) to determine the optimal clamping pressure of a stack of elementary cells; (iv) to highlight the role of the fluid (stagnant air or circulating thermostatically controlled water) present in the stack on the shape and position of the cells. individual cell distributions. The chronological steps outlined above were followed to perform the test: 1. Definition of the technical means implemented: the reactor used to carry out this The test comprises twelve cells; the individual voltage of each cell in the stacking is measured and converted into a numerical value using a system acquisition (a commercial multimeter) including an acquisition card having 20 input channels (unipolar or bipolar), and equipped with a multiplexer for reading and the analog-to-digital conversion of each channel. Each input channel on The unipolar acquisition card is connected to an elementary cell by a cable small cross-section electrical wire (0.1 mm) with a length of up to a few meters. The voltage measured on each cell is, for this type of reactor, between 0 and 3 volts, depending on the total voltage applied to the stack (faradic or non-faradic conditions). 2. Definition of operating conditions: the test was performed with cells filled with two types of fluids; either under stagnant air or under circulating thermostatically controlled water; the lag bolts used to compress the stack and ensure proper stack compression were tightened with a torque wrench applying a torque between 40 and 80 Nm. 3. Definition of sampling conditions (example): each individual voltage was measured once every 10 seconds for 24 hours. 4. Test execution: a first test was performed with cells filled with stagnant air at 22°C and a second test was performed with cells filled with circulating thermostatically controlled water at 35°C. The individual voltage of each cell in the stack of elementary cells was measured every ten seconds for 24 hours.A sample of 8640 points was measured for each cell during each of the two tests. 5. Sample processing: Each cell produced a voltage sample; each sample was converted into a distribution by digital processing using a computer program or commercial spreadsheet. 6. Figure 3 shows the cell-by-cell distribution plots when the stack was tightened with a torque of 70 Nm and when the test was performed with cells filled with stagnant air at 22°C. The classes used for each cell range from 0 to 0.6 volts in steps of 0.001 volts (1 millivolt). 7. Figure 4 shows the cell-by-cell distribution plots when the stack was tightened with a torque of 70 Nm and when the test was performed with cells filled with circulating water thermostatically controlled at 35°C. The classes used for each cell range from 0 to 2.5 volts in steps of 0.001 Volt (1 millivolt).The presence of several distributions on the voltage axis stems from the fact that the water temperature was changed during the test. 8. Test conclusions: the two tests provide several useful pieces of information: (i) it is demonstrated that the invention can be used to perform a diagnosis in the absence of an AME and in the presence of a non-conductive element (the PTFE film) in the stack; the heterogeneities between the individual distributions measured under these conditions (absence of electric current) already indicate what will happen under current flow, very useful information for the operator. manufacturing of the assembly; (ii) the electrical percolation threshold of the stack is obtained from a torque of approximately 40 Nm; tightening up to 80 Nm does not improve homogeneity; (iii) plotting distributions in the form of 3D graphs show significant differences between the distributions of a cell elementary to the other, which translates into strong mechanical and electrical heterogeneities in the stacking; the position of the distributions on the tension axis does not vary by monotonously from one cell to the next; the distributions are more or less tighter, which reflects a more or less homogeneous average operation; this type stacking was therefore not carried out correctly; (iii) differences in distributions are observed according to the fluid present in the elementary cells and their temperature; therefore, it is necessary to carry out the tests under the most favorable conditions. Isothermal distributions are possible, but the values ​​of the electrical distributions depend on the fluid. present in cells. Example 2 of use of the invention In this second example, the objective of measuring the homogeneity of the cells' electrical responses remains the same, but this time, the AMEs are mounted in the stack. Given their cost and fragility, the challenge is to determine whether the cell stack built during the assembly and compression phase is sufficiently homogeneous before deciding whether or not to pass current, as passing current is an irreversible step in the reactor activation process. An unfavorable diagnosis allows the operator to re-tighten the stack, or even dismantle it to correct the position or replace defective or poorly assembled cell components, without risking damage to the AME (a key and expensive component of individual cells).The physical quantity examined in this example is again the individual electrical voltage of each elementary cell in a stack of elementary cells powered by a direct current voltage. No electric current flows through the stack because the voltage applied to each cell is less than the thermodynamic start-up voltage of water electrolysis (1.23 volts per cell under standard temperature and pressure conditions). The chronological steps indicated above were followed: 1. Definition of the technical means implemented: The reactor used to perform this test comprises twelve cells; the individual voltage of each cell in the stack is measured and converted into a digital value using a data acquisition system (a commercial multimeter) with a data acquisition card having 20 input channels (unipolar or bipolar), and equipped with a multiplexer for reading and converting each channel to analog-to-digital. Each input channel on the unipolar data acquisition card is connected to an individual cell by a thin electrical cable (0.1 mm²) up to a few meters long. The voltage measured on each cell, for this type of reactor, is between 0 and 3 volts, depending on the total voltage applied to the stack (faradic or non-faradic conditions). 2. Definition of operating conditions: the test was carried out under circulating water thermostatically controlled in each compartment (anodic and cathodic); the water flow rate is 300 liters / hour; the test temperature is 45°C; the operating pressure is 2 bars; the operating mode is potensiostatic (total voltage of 5 Volts); the individual voltage of each cell (approximately 5 / 12 = 0.417 volts) is less than the thermodynamic voltage of electrolysis and therefore there is no passage of electric current. 3. Defining sampling conditions (example): each individual voltage has was measured once every 10 seconds for 24 hours. 4. Test procedure: The individual voltage of each cell in the stack of elementary cells was measured every ten seconds for 24 hours. A sample of 8640 points was measured for each cell. 5. Sample processing: Each cell produced a voltage sample; each sample was converted into a distribution by digital processing using a computer program or commercial spreadsheet; the classes used for each cell range from 0 to 0.6 volts in steps of 0.001 volts (1 millivolt). 6. Figure S shows the plot of the cell-by-cell distributions as a graph. 2D. 7. Figure 6 shows the plot of the cell-by-cell distributions in graph form. 3D. 8. Conclusions of the test: the test provides several useful pieces of information: (i) it is demonstrated that it is possible to obtain homogeneity information on individual cells containing AMEs without having to pass a faradaic current; The information obtained, similar to that obtained in example 1, already indicates what the situation will be under current flow (example 3); (ii) the plotting of the distributions in the form of 2D graphs clearly shows the spread of performance along the voltage axis, from one cell to another, whether the measurements were taken under strictly stationary conditions or not; significant differences between the distributions from one elementary cell to another, which reflects strong mechanical and electrical heterogeneities in the stacking; the position of the distributions on the voltage axis does not vary monotonically from one cell to another; the distributions of each cell are more or less tightly packed, which reflects a more or less homogeneous operation of each cell; this type of stacking was therefore not carried out correctly;(iii) plotting the distributions in the form of 3D graphs provides an overview of the stacking behavior and allows for the qualitative identification of the cells containing the main heterogeneities. Example 3 of Use of the Invention In this third example, the objective of measuring the homogeneity of cell responses remains the same. The AMEs are mounted in the stack, and the stack operates in production, in intensiostatic mode, with the circulation of a faradaic current. The physical quantity sought in this example is again the individual electrical voltage of each elementary cell in a stack of elementary cells. The chronological steps indicated above were followed: 1. Definition of the technical means implemented: The reactor used to perform this test comprises forty-four cells; the individual voltage of each cell in the stack is measured and converted into a digital value using a data acquisition system (a commercial multimeter) with a data acquisition card having 60 input channels (unipolar or bipolar), and equipped with a multiplexer for reading and converting each channel to analog-to-digital. Each input channel on the unipolar data acquisition card is connected to an individual cell by a thin electrical cable (0.1 mm²) up to a few meters long. The voltage measured on each cell, for this type of reactor, is between 0 and 3 volts, depending on the total voltage applied to the stack (faradaic or non-faradaic conditions). 2. Definition of operating conditions: The test was performed with thermostatically controlled water circulating in each compartment (anodic and cathodic); the fluid flow rate was 800 liters / hour; the fluid temperature during the test was 72°C; the operating pressure was 15 bar; the operating mode was current-intensity (applied current density of 1.2 Amperes per cm²); this resulted in a total stacking voltage of 76 Volts (i.e., an individual cell voltage of 1.73 Volts). 3. Definition of sampling conditions (example): Each individual voltage was measured once every 30 seconds for 24 hours. 4. Test execution: Over the 24 hours, a sample of 8640 points was measured for each of the 44 cells. 5. Sample processing: each cell produced a voltage sample of 8640 points; each sample was converted into a distribution by digital processing using a commercial spreadsheet; the classes used for each cell range from 1.6 to 1.8 volts in steps of 0.001 volts (1 millivolt). 6. Figure 7 shows the plot of the cell-by-cell distributions in graph form. 2D. 7. Figure 8 shows the plot of the cell-by-cell distributions in graph form. 3D. 8. Test Conclusions: The test provides several useful pieces of information: (i) it demonstrates that it is possible to obtain homogeneity information on individual cells containing AMEs under the passage of a faradaic current, i.e., when the reactor is operating under nominal conditions; (ii) the plotting of the distributions in the form of 2D graphs clearly shows the spread of performance along the voltage axis, from one cell to another, whether the measurements were made under strictly steady-state conditions or not; the shape of the distributions depends on the particular conditions of the test; the presence of points on the high-voltage parts of the distributions comes from the phases during which cooler water is injected into the reactor to replenish the stock of water that has been electrolyzed;the distributions from one elementary cell to another are tightly packed, which reflects a relative mechanical and electrical homogeneity in the stacking; the position of the individual distributions on the voltage axis does not vary monotonically from one cell to another; this type of stacking has therefore been carried out in a way; satisfactory; (iii) cells whose distribution appears at the lowest voltages are those for which the temperature of the water circulating in the cell is the high: these are therefore the cells in which the water flow is the lowest; Conversely, the cells whose distribution appears at the highest voltages are those for which the temperature of the water circulating in the cell is the low: these are therefore the cells in which the water flow is strongest; (iv) the Plotting the distributions in the form of 3D graphs provides an overview of the stacking behavior and allows for qualitative identification of the main heterogeneities and problematic cells. Example 4 of Use of the Invention: In this fourth example (Figure 9), the voltage distributions measured on a stack of 100 PEM-type water electrolysis cells at two different times (after twelve months of operation and at the beginning of their service life = BoL) are compared. A group of about twenty cells appears in the middle of the stack whose voltage distributions have shifted towards higher voltages, while for the other cells, the distributions have remained virtually unchanged. These twenty or so cells have aged faster than the others. These will be the ones that need to be replaced first. Extrapolating the degradation rate (linear from two distributions or non-linear when several distributions are available at different times during reactor operation) makes it possible to anticipate when these cells will have reached a critical voltage requiring maintenance.

Claims

Demands 1. A multiphysics characterization method for an electrochemical or photoelectrochemical reactor, consisting of operating said reactor in steady state, sampling a characteristic physical quantity of interest using a measurement chain for a sampling time and with a particular sampling frequency, converting the sample thus obtained into a distribution and representing this distribution in the form of 2D or 3D graphs, in order to determine the degree of homogeneity of the stacking and to identify the cells that are malfunctioning.

2. A method according to claim 1 characterized in that the reactor is made up of the stacking of any number of individual electrochemical or photoelectrochemical cells, each of which is the site of endergonic or exergonic type electrochemical reactions, having a finite or infinite capacity for storing the active material, and having any geometric shape (plane or cylindrical) and dimensions.

3. A method according to any one of the preceding claims characterized in that the reactor operates in steady state, i.e. with average values ​​of temperature, pressure, current density, voltage and incoming and outgoing material flux, which are constant and independent of time, during the sampling period.

4. A method according to any one of the preceding claims characterized in that the characteristic physical quantity is measured on a single individual cell or on any subset of individual cells of the stack, adjacent or non-adjacent, or on all individual cells contained in the stack.

5. A method according to any one of the preceding claims characterized in that the measurements obtained by sampling on the individual cells are samples consisting of a set of pairs {value of the characteristic physical property; time at which the measurement is made}, acquired during sampling, which is characterized by a sampling duration of any duration, and by any sampling frequency, so as to obtain a sample having a (non-exclusive) size between a few tens and a few thousand pairs.

6. A method according to any one of the preceding claims characterized in that the characteristic physical quantity sampled during the sampling time to form the sample is a mechanical, electrical, electrochemical, chemical, fluidic, or thermal physical quantity, and more particularly any one or any combination thereof, of: - the electrical voltage measured across each individual cell in the stack, or across each subset of adjacent cells in the stack, or across the entire stack of individual cells, - the individual current that flows through each individual cell in the stack, or the total current that flows through the entire stack of individual cells, - the temperature of the fluid(s) at the inlet of an individual cell, - the average temperature of the fluid(s) at the inlet of any subset of individual cells or at the inlet of the entire stack of individual cells, - the temperature of the fluid(s) at the outlet of an individual cell, - the average temperature of the fluid(s) at the outlet of any subset of individual cells or at the outlet of the entire stack of individual cells, - the pressure of the fluid(s) at the inlet of an individual cell, - the pressure of the fluid(s) at the inlet of any subset of individual cells or of the total stack of individual cells, - the pressure of the fluid(s) at the outlet of an individual cell, - the pressure of the fluid(s) at the outlet of any subset of individual cells or of the entire stack of individual cells, - the flow rate of the fluid(s) at the inlet of an individual cell, - the flow rate of the fluid(s) at the inlet of any subset of individual cells or of the entire stack of individual cells, - the flow rate of the fluid(s) at the outlet of an individual cell, - the flow rate of the fluid(s) at the outlet of any subset of individual cells or of the entire stack of individual cells, - the chemical composition of the fluid(s) entering an individual cell, - the chemical composition of the fluid(s) at the inlet of any subset of individual cells or of the total stack of individual cells, - the chemical composition of the fluid(s) exiting an individual cell, - the chemical composition of the fluid(s) at the outlet of any subset of individual cells or of the total stack of individual cells, - the clamping force applied to any subset of individual cells or to the entire stack of individual cells, - the clamping force torque applied to any subset of individual cells or to the entire stack of individual cells, - the clamping pressure applied to any subset of individual cells or to the entire stack of individual cells, 7. A method according to any one of the preceding claims characterized in that the measurement chain enabling the sampling of a particular property and obtaining the associated sample consists of a specific sensor per characteristic property having, non-exclusively, an accuracy of at least 5%, an analog-to-digital converter per cell or a multiplexer-type acquisition system, and a digital storage and archiving system for the sample.

8. A method according to any one of the preceding claims characterized in that the sample of claim 7 is converted into a distribution, by subdividing the range of values ​​of the characteristic quantity being examined, between the minimum value and the maximum value of this characteristic quantity contained in the sample, into a set of adjacent classes of any number, and by counting for each class the number of occurrences of the characteristic quantity present in the sample obtained during sampling.

9. Method according to claim 8 characterized in that the distribution is represented in the form of 2D or 3D graphs allowing comparison of the homogeneity of the characteristic property of each individual cell in the stack in order to make a diagnosis for the purpose of quality control operations during the manufacture of the reactor or maintenance during reactor operation.

Images