Method and arrangement for determining the charge imbalance of a redox flow battery

Abstract

In order to allow simple and precise determination of a charge imbalance (L) between the electrolyte liquids (5a, 5b) of a redox flow battery (1) over a long period of time, a present correlation (25) between the determined cell voltage (VOCV) and the determined at least one concentration potential (Ea, Eb) is ascertained for the period of time and the correlation (25) is evaluated in order to determine the charge imbalance (L).

Description

[0001] Method and arrangement for determining the charge imbalance of a redox flow battery

[0002] The present invention relates to a method for determining the charge imbalance between a positive electrolyte and a negative electrolyte of a redox flow battery. During operation of the redox flow battery, a cell voltage between the positive electrolyte and the negative electrolyte and a concentration potential of at least one of the positive and negative electrolyte are determined during a measurement period. The invention also relates to a corresponding arrangement for determining the charge imbalance.

[0003] A redox flow battery is an electrochemical energy storage device. It typically consists of storage tanks for electrically positive and negative electrolyte fluids, as well as pumps and pipes for circulating the electrolyte fluids (electrolytes) through one or more cell stacks, each containing a number of individual cells. The individual cells of the cell stack are each formed by a positive half-cell and a negative half-cell arranged side by side, with the positive and negative half-cells of an individual cell separated from each other by a semipermeable membrane, typically an ion exchange membrane. The semipermeable membrane is, for example, a cation and / or anion exchange membrane, e.g., Nation®. The positive half-cell contains a positive electrode located in a frame, through which the positive electrolyte fluid flows.The negative half-cell contains a negative electrode located in a frame, through which the negative electrolyte flows. The positive and negative electrolyte circulate separately through the half-cells. The positive and negative electrodes are usually made of porous graphite felts, through which the electrolyte flows.

[0004] The half-cell frames are arranged next to each other and connected to form the cell stack, with negative and positive half-cells alternating. Electrode plates, such as bipolar plates, are arranged between adjacent individual cells of the cell stack as current collectors. These plates are usually made of a composite material made of carbon and plastic. Current collectors are located on the axial outer sides of the axially outer individual cells of the cell stack. An electrical contact is made via these current collectors to the outside in order to tap an electrical voltage across the entire cell stack (discharging the redox flow battery) or to apply an electrical voltage to the cell stack (charging the redox flow battery). The cell stack is closed off on each axial outer side by an end plate, which holds the cell stack together.

[0005] In a vanadium-based redox flow battery, the positive electrolyte fluid in the charged state consists of a redox pair in the form of vanadium with the oxidation number +4 (also known as V lv or V 4+ and vanadium with the oxidation number +5 (also known as V v or V 5+ The negative electrolyte fluid in the charged state consists of a redox pair in the form of vanadium with the oxidation number +2 (also known as V" or V 2+ referred to) and vanadium with the oxidation number +3 (also known as V IH or V 3+ (referred to as "negative electrolyte") - meaning that the negative electrolyte fluid has a more negative electrochemical potential than the positive electrolyte fluid. Both the positive and negative electrolyte fluids can also contain sulfuric acid and other additives.

[0006] This structure and function of a redox flow battery is well known, for example from WO 2018 / 087270 A1 or WO 2014 / 1331702 A1 .

[0007] During operation of a redox flow battery, electrical energy is delivered to an electrical load or absorbed from an energy source. In the example of a vanadium-based redox flow battery, the following well-known chemical reaction (redox reaction) occurs during discharge in the redox flow battery: y5 + _|_ y2 + y4 + _|_ y3 +

[0008] During the charging process, this reaction is reversed with the help of an external current or voltage source that applies an electrical voltage to a single cell (or the entire cell stack), and the reactions shown above occur in the opposite direction. A charging process requires a voltage increased by the cell voltage (overvoltage).

[0009] In addition to the desired charge / discharge reactions, parasitic reactions can also occur, which lead to an ineffective charge / discharge process.

[0010] In addition to impurities, hydrogen reduction in an aqueous electrolyte can cause a significant parasitic effect and lead to an imbalance in the charging process, for example in the form of whereby hydronium ions HaO +to hydrogen gas H2. Impurities that are practically always present in the electrolyte, such as elements Cu, Sb, Ag, Ni, or platinum group metals such as Pt, Pd, Ir, Rh, etc., act as catalysts for hydrogen reduction and enhance the parasitic effect of hydrogen formation. This leads, for example, to a reduced charge of the negative electrolyte during charging. After a redox flow battery undergoes numerous charge / discharge cycles, the resulting charge imbalance between the electrolytes increases.

[0011] It is also possible that water electrolysis of the aqueous electrolyte occurs at both electrodes, producing both hydrogen gas and oxygen gas.

[0012] Such parasitic reactions consequently lead to an imbalance of the charge states of the charge carriers (redox pair) in the two electrolyte liquids, and thus, for example, to different ratios of V 2+ / V 3+ and V 4+ / V 5+An imbalance of the charge states in the electrolyte fluids leads to a charge imbalance in the redox flow battery and consequently to a limited battery capacity of the redox flow battery and is fundamentally undesirable. A charge imbalance can be expressed as an average oxidation number of the redox elements in the electrolyte fluids. In the case of a positive charge imbalance (negative electrolyte fluid undercharged and / or positive electrolyte fluid overcharged), increased oxidative stress on battery components can also occur, which can lead to increased degradation of such battery components and battery failure.

[0013] In the event of a charge imbalance in a redox flow battery, the redox flow battery can be serviced to rebalance or at least reduce the charge imbalance. A redox flow battery can also be equipped with a charge balancing unit to balance or at least reduce a charge imbalance in the electrolyte fluids. If a certain charge imbalance is detected, such a charge balancing unit can also be activated. An example of a charge balancing unit can be found in WO 2014 / 198546 A1. Last but not least, other measures could also be taken to protect the redox flow battery from damage in the event of a charge imbalance. For example, the maximum permissible cell voltage of an individual cell could be reduced to prevent damage to the cell stack.

[0014] In order to compensate for a charge imbalance between the electrolyte fluids during operation of a redox flow battery, the extent of the charge imbalance must be known.

[0015] However, these different ratios of the charge carriers in the half-cells cannot be determined using the cell voltage alone, which could be measured and used as the state of charge (SoC) of the redox flow battery, since cell voltage can only determine the potential differences between the two electrodes of the cell, and thus between the two electrolyte fluids, which does not allow any conclusions to be drawn about a possible charge imbalance between the electrolyte fluids.

[0016] To measure the electrical concentration potentials of electrolyte fluids, which can also be used as the SoC of the positive or negative electrolyte fluid, a reference electrode is typically used, which is immersed, for example, in one of the storage tanks in a redox flow battery. A reference electrode provides a (preferably constant) reference potential to which another electrical potential, for example, of an electrolyte fluid, can be referenced. The classic reference electrode is the standard hydrogen electrode (SHE). Although it provides a very stable reference potential, it is cumbersome due to its requirement for gaseous hydrogen and is difficult to integrate into existing systems, particularly because gaseous hydrogen must be supplied.Therefore, so-called second-type reference electrodes, such as the silver / silver chloride electrode (SCE) or the mercury / mercury sulfate electrode (MSE), are used today. The so-called dynamic hydrogen electrode (DHE) is also used, in which hydrogen gas is generated in situ at the electrode via an applied current, thus eliminating the need for an external gas reservoir as with SHE. However, there are a variety of other reference electrodes that can be used depending on the system under investigation.

[0017] However, reference electrodes have the disadvantage that, while they generate a stable reference potential over short periods of time, the reference potential is still subject to drift, i.e., a change in the reference potential, over long periods of time. If the reference potential is subject to drift, the measurement of a concentration potential with the reference electrode is negatively influenced and distorted. With an expected service life of a redox flow battery of up to 20 years, such drift in a reference electrode is a major problem.

[0018] The reason for such a drift is, for example, contamination of the reference electrode with the active redox elements of the redox flow battery. Osmotic water loss or water gain in the reference electrode, which leads to a change in the internal chemical equilibrium conditions of the reference electrode, can also cause drift. The second type of reference electrodes mentioned above, for example, have the disadvantage that they are subject to degradation during normal use. For example, a semipermeable membrane in a reference electrode can be attacked by an electrolyte or medium, but the internal electrolyte in the reference electrode can also evaporate, for example due to thermal stress.

[0019] Reference electrodes, especially reference electrodes of the second type, often also introduce foreign elements into the redox flow battery system, which can accelerate self-discharge and / or hydrogen evolution reaction (HER).

[0020] During operation of a redox flow battery, a reference electrode must therefore be regularly maintained, calibrated, or even replaced if it loses its function. However, this is difficult in a stationary system, which is often used in remote locations, and involves significant personnel and equipment expenditure. Furthermore, the redox flow battery cannot be used during such maintenance.

[0021] Since the determination of the concentration potential of an electrolyte fluid with a reference electrode is affected and, in particular, becomes inaccurate due to a drift in the reference electrode's reference potential over extended periods, charge imbalances between the electrolyte fluids can also only be determined to a limited extent or inaccurately. This leads to a limited capacity of the redox flow battery, but can also cause damage to the redox flow battery or its components.

[0022] From WO 2015 / 073286 A1, it is known to determine the change in the concentration potential of an electrolyte liquid as a function of a charging current during charging of the electrolyte liquid. For this purpose, the concentration potential can be measured at two points in time using a reference electrode. The charging current flowing during this period must be known, as must the concentration of the redox elements of the redox pair in the electrolyte liquid. From this, the change in the concentration of the redox elements due to the flowing current between the two points in time can be determined. From the change in the concentration potential and the change in the concentration of the redox elements, a change in the concentration potential can be determined as a function of the change in the concentration of the redox elements, which in turn can be correlated with the charge state of the electrolyte liquid.Since only a change in the concentration potential is considered, this method is independent of a drift of the reference potential of a reference electrode.

[0023] However, additional parameters of the electrolyte fluid, such as the concentration of redox elements, the volume of the electrolyte fluid, and the duration of the current flow, must be known, which in turn makes the method more complex than a simple voltage measurement. Therefore, there is a need for a simple and accurate determination of charge imbalance between the electrolyte fluids of a redox flow battery over a long period of time.

[0024] The objective technical problem is solved by determining a current relationship between the determined cell voltage and at least one determined concentration potential for the measurement period, and evaluating the current relationship to determine the charge imbalance. It was recognized that the relationship between the cell voltage and the concentration potential of an electrolyte fluid, apart from possible drift, depends only on the charge imbalance between the two electrolyte fluids and can thus be used to determine the charge imbalance of the redox flow battery. For this purpose, only the two voltages need to be determined; no further knowledge of the electrolyte fluids is required.With the redox flow battery, there is no need to run separate charge and discharge cycles to determine the charge imbalance, which would interrupt the normal operation of the redox flow battery.

[0025] The current relationship can advantageously be determined for the period of the measurement from a curve of the cell voltage versus at least one concentration potential. Alternatively, this curve can be derived with respect to the cell voltage and the derivative used as the relationship. Alternatively, a curve of the cell voltage versus the concentration potential of the positive electrolyte fluid and a curve of the cell voltage versus the concentration potential of the negative electrolyte fluid can be determined for the period of the measurement, the determined curves can be derived with respect to the cell voltage and a ratio of the derivatives of the curves can be used as the relationship. Using the derivative has the advantage that the relationship is independent of any possible drift in the reference voltage of a reference electrode. This means that any reference electrode can be used, even one subject to drift.

[0026] To easily determine the charge imbalance, the determined current relationship is compared with a predefined reference curve for a specific charge imbalance, and the charge imbalance underlying the reference curve is used as the charge imbalance of the redox flow battery if the deviation between the reference curve and the current relationship is smaller than a predefined limit. This makes it possible to use a reference curve, for example for a maximum permissible charge imbalance. If this charge imbalance is reached, appropriate countermeasures can be initiated. To easily determine the charge imbalance, the determined current relationship can also be compared with predefined reference curves for different charge imbalances. This makes it possible to determine the reference curve that minimizes the deviation between the relationship and the reference curve.The charge imbalance assigned to this reference curve then represents the current charge imbalance of the redox flow battery. With this method, multiple reference curves can be used for different charge imbalances, for example, at intervals of 1% or 0.5%. This always yields a current charge imbalance of the redox flow battery.

[0027] It is even more advantageous to use a predefined charge imbalance-dependent reference curve. This allows the charge imbalance of the reference curve to be varied, and the charge imbalance that minimizes the deviation between the relationship and the reference curve can be determined. This charge imbalance, which minimizes the deviation, is used as the charge imbalance of the redox flow battery. This allows the charge imbalance to be determined using a single known reference curve.

[0028] The present invention will be explained in more detail below with reference to Figures 1 to 11, which show exemplary, schematic and non-limiting advantageous embodiments of the invention.

[0029] Fig. 1 shows a schematic structure of a single cell of a cell stack of a redox flow battery,

[0030] Fig. 2 the measurement of the cell voltage and a concentration potential of an electrolyte liquid,

[0031] Fig. 3a shows the time course of the cell voltage and the concentration potential of an electrolyte liquid during a measurement period,

[0032] Fig. 3b shows a relationship between the cell voltage and the concentration potential of an electrolyte liquid during a measurement period, Fig. 4 shows a drift of such a relationship due to the drift of the reference electrode,

[0033] Fig. 5 shows a relationship in the form of the derivation of the cell voltage according to the concentration potential of the electrolyte liquid,

[0034] Fig.6 a simple reference electrode for measuring the concentration potential of the electrolyte liquid,

[0035] Fig.7 a measuring cell for determining the cell voltage,

[0036] Fig.8 and 9 each show a combination of a measuring cell and a reference electrode, Fig.10a and Fig.10b show the result of an inventive determination of the charge imbalance of a redox flow battery and Fig.11 shows an arrangement for determining the charge imbalance of a redox flow battery.

[0037] Fig. 1 shows a schematic structure of a redox flow battery 1 using a single cell 18 of a cell stack 10.

[0038] For better explanation and illustration, only a single cell 18 of a cell stack 10 of a redox flow battery 1 is shown in Fig. 1. The term "battery" is applied in the present application to secondary cells that allow at least one recharging, i.e., a charging process. Secondary cells are often also referred to as accumulators. In a practical implementation of a redox flow battery 1, a plurality of such individual cells 18 are naturally often combined in a known manner to form a cell stack 10, for example as described above. Such cell stacks 10 can generate high electrical voltage and high electrical currents and are designed differently depending on the application. A redox flow battery 1 can also contain multiple cell stacks 10, which can be connected in series and / or parallel.Individual cells 18 and cell stacks 10 and their structure and modes of operation are well known to those skilled in the art and are not crucial for explaining the invention, which is why they are only briefly described below.

[0039] A single cell 18 consists of two half-cells 18a, 18b, which form a positive reaction chamber 11a and a negative reaction chamber 11b. The two half-cells 18a, 18b, or the positive reaction chamber 11a and the negative reaction chamber 11b, are separated by a semipermeable, in particular ion-selective, membrane 19. The reaction chambers 11a, 11b are formed, for example, in recesses of frames. Typically, an electrode is arranged in each of the reaction chambers 11a, 11b. Electrolyte fluids 5a, 5b with different electrical charges (positive and negative electrolyte fluid) flow through the reaction chambers 11a, 11b. Each of the electrolyte fluids 5a, 5b contains a redox pair with specific, time-varying concentrations (depending on the state of charge) of redox elements.The semipermeable, in particular ion-selective, membrane 19 can be made, for example, of polytetrafluoroethylene, with the trade name Nation™, and allows ions, for example hydronium ions HaO. + , oxygen ions O 2 ' (usually bound in complex ions) or chloride ions CI create a charge balance between the positive reaction space 11a and the negative reaction space 11b (or between the electrolyte liquids 5a, 5b contained therein). The individual cell 18 is closed off on both sides by an electrode plate 14, 15. In a cell stack with several adjacent individual cells 18, an electrode plate 14, 15, such as a bipolar plate, is arranged between each two adjacent individual cells 18. At the outer ends of the cell stack, a current collector 6, 7 can be attached to the electrode plates 14, 15 or to the half-cells 18a, 18b, which can be electrically contacted from the outside.

[0040] There are a variety of redox-active elements or ions, or combinations of redox-active elements and / or ions, which can be used as redox pairs in a redox flow battery 1. A redox pair consists of a first redox element and a second redox element in the form of a redox-active element or ion, or combinations of redox-active elements and / or ions with different electrical charges (oxidation numbers). The redox elements of a redox pair are usually contained in electrolyte liquids 5a, 5b, for example, dissolved in the electrolyte liquid. Depending on the cell voltage Vocv and the design of the redox flow battery 1, a wide variety of combinations of redox pairs can be used. Some non-exhaustive examples of known combinations of redox pairs are V 2+ / V 3+ in the negative electrolyte fluid vs. V 4+ / V 5+in the positive electrolyte (in a vanadium redox flow battery), V 2 7V 3+ vs. Bi7CIBr2, Br2 / Br vs. S / S 2 -, Br7Br2vs. Zn 2+ / Zn, Ce 4+ / Ce 3+ vs. V 2+ / V 3+ , Fe 3+ / Fe 2+ vs. Br2 / Br, Mn 2+ / Mn 3+ vs. Br2 / Br, Fe 3+ / Fe 2+ vs. Ti 2+ / Ti 4+ and others in other redox flow batteries. Depending on the charge state of the redox flow battery 1, the redox elements of the redox pairs are present in the electrolyte fluids 5a, 5b in different concentration ratios.

[0041] For a simple understanding of the invention, the explanation will only refer to the vanadium-based redox flow battery 1, which contains the redox elements V as redox pairs. 2+ / V 3+ and V 4+ / V 5+However, the invention is of course not limited to this embodiment.

[0042] When discharging the redox flow battery 1, the reduction of the oxidant takes place in the positive reaction chamber 11a. In the case of the vanadium-based redox flow battery 1, V 5+ in the positive electrolyte liquid to V 4+ reduced and one electron e- is consumed. V +4 occurs, for example, in the form of the vanadyl ion [VO(H2O)S] 2+ and vanadium V 5+ in the form of the vanadate ion [VO2(H2O)a] 1+ in the aqueous electrolyte (e.g., sulfurous acid). The oxidation of the reducing agent takes place in the negative reaction space 11 b. In the case of the vanadium-based redox flow battery 1, V 2+ becomes V 3+oxidized, and an electron e- is released to the electrode plate 14 of the negative half-cell 18b. During the chemical reaction, an electron e- is released to the electrode plate 14 of the negative half-cell 18a (anode) and passes via electrical lines to an electrical load 16 and then to the electrode plate 15 of the positive half-cell 18b (cathode), where the reaction described above takes place in the positive reaction chamber 11a. When charging with an electrical charging voltage (instead of the load 16), the corresponding reverse reactions take place.

[0043] During the charging process of the redox flow battery 1, the load 16 is replaced, for example, by an electrical voltage source. An electrical charging voltage is applied that corresponds to the cell voltage Vocv, including an overvoltage. In the vanadium-based redox flow battery 1 used as an example, the cell voltage Vocv is approximately 1.45 V. The overvoltage results from kinetic, ohmic, and diffusive inhibitions during charging of the redox flow battery 1 and is also dependent on the charging current.

[0044] An electrical load 16 can take any form. Based on the voltage, current, or power requirements of the electrical load 16, a cell stack can be configured in a redox flow battery 1 to provide the necessary electrical voltage and / or the necessary electrical current. Redox flow batteries 1 are often used as stationary energy storage devices, for example, to serve as emergency power systems for industrial plants, storage systems for renewable energy (photovoltaics, wind power), and the like. Consequently, depending on the application, a person skilled in the art can design a cell stack, or a parallel and / or serial connection of multiple cell stacks, and redox pairs in a redox flow battery 1.

[0045] To enable a continuous reaction in the half-cells 18a, 18b, the electrolyte liquids 5a, 5b with the respective redox pairs are circulated through the half-cells 18a, 18b. For example, pumps P can generate a flow of the respective electrolyte liquid 5a, 5b through a half-cell 18, 18b of an individual cell 18. The pumps P convey the respective electrolyte liquids 5a, 5b via lines 17 from respective storage tanks 12, 13 of the electrolyte liquids 5a, 5b into the half-cells 18a, 18b of an individual cell 18 and from there back into the storage tanks 12, 13. Typically, there is one storage tank 12, 13 for each electrolyte liquid 5a, 5b. However, it may also be possible to have separate storage tanks 12, 13 for used (discharged) and unused (charged) electrolyte fluids 5a, 5b.

[0046] During the discharge process via the consumer 16, the concentrations (expressed by a square bracket [.]) of the redox elements in the positive electrolyte liquid 5a shift, for example from [V 5+ ] towards [V 4+ ] , in a positive half-cell 18a and in the associated storage tank 12 for the positive electrolyte liquid 5a. At the same time, the concentrations of the redox elements in the negative electrolyte liquid 5b shift, for example from [V 2+ ] towards [V 3+], in a negative half-cell 18b and in the associated storage tank 13 for the negative electrolyte liquid 5b. This shift caused by the chemical redox reaction can continue until an electrochemical potential equilibrium has been established between the two half-cells 18a, 18b, i.e., until all electrochemical potentials of all species involved in the redox reaction are equal. This completely discharges the redox flow battery 1. During the charging process, the concentrations shift in the opposite direction. The concentrations of the redox elements in an electrolyte liquid 5a, 5b thus change between 0% and 100%. The concentrations of the redox elements can therefore also be used as the state of charge of the electrolyte liquids 5a, 5b.

[0047] Changes in the concentrations of redox elements of the respective redox pair lead to a change in the electrical potential E in a half-cell 18a, 18b, also known as the concentration potential. The well-known Nernst equation can be used to calculate such concentration potentials E:

[0048] Where E° is the standard potential of the redox couple according to the electrochemical series, R is the known ideal gas constant, T is the temperature in Kelvin, F is the known Faraday constant, and z is the number of transferred electrons e-. The terms [Red] and [Ox] refer to the oxidized and reduced forms of the redox element of the redox couple in half-cells 18a and 18b, respectively, and are usually expressed as activity. However, at low concentrations in half-cells 18a and 18b, the activity can be assumed to be equal to the concentration.

[0049] Using the example of the negative electrolyte liquid 5b in the vanadium-based redox flow battery 1, [Red] would be the concentration of [V 2+ ] in the negative electrolyte liquid 5b, while [Ox] is the concentration [V 3+ ] If one inserts the two concentrations, combines the constants and converts the natural logarithm into the decimal logarithm, the following version of the Nernst equation results at standard conditions (298 K, 1 bar) and an electron e-transition (i.e. z=1), where [V] stands for the unit volt of electrical potential: E [V] = -0.250 [V] - 0.059

[0050] This results in a difference of 59 mV per decade difference in the concentrations of [V 2+ ] to [V 3+ ], For example, with a standard potential E° of -0.250 V and a concentration of [V 2+ ] = 0.01 mol (M) and [V 3+] = 0.1 M a concentration potential E of -0.191 V. Thus, based on the concentration differences, for example due to the parasitic reactions mentioned above, a shift of the concentration potential E occurs. The knowledge of a concentration potential E can therefore be used to obtain knowledge about the concentration ratio of the redox elements of a redox pair in an electrolyte liquid 5a, 5b, and thus in particular also about the charge state of the electrolyte liquids 5a, 5b.

[0051] In principle, various parasitic reactions can occur in a redox flow battery 1. While hydrogen electrolysis can occur in an aqueous electrolyte, other reactions can also occur specifically at the anode (negative half-cell 18b). An example is the reaction of V 2+ with hydronium ions HaO +during charging of the negative electrolyte liquid 5b with the formation of hydrogen gas: v 2+ + H3O + -> y 3+ + 0.5 H2+ H2O

[0052] This reaction depends on the concentration of [V 2+ ] and the pH of the electrolyte liquid 5b, as the negative decimal logarithm of the hydronium ions [HaO + ]. In an aqueous electrolyte there are always hydronium ions HaO + so that this parasitic reaction always occurs. This reaction leads to a change in the [V 2+ ] to [V 3+ ] Concentration during charging of the redox flow battery 1. Because hydrogen gas H2 escapes, this reaction is also virtually irreversible.

[0053] Due to such parasitic reactions, a change in the concentrations of the redox elements in at least one electrolyte liquid 5a, 5b can occur and thus a charge imbalance between the electrolyte liquids 5a, 5b and consequently also a change in a concentration potential E a , Eb of an electrolyte liquid 5a, 5b.

[0054] However, a charge imbalance can also occur if the negative electrolyte liquid 5b is oxidized by the ingress of air, resulting in less reducible redox element in the negative electrolyte liquid 5b than the oxidizable counterpart in the positive electrolyte liquid 5a. This also changes the concentration potential E a , Eb in at least one electrolyte liquid 5a, 5b.

[0055] Of course, there may also be other reasons for a charge imbalance between the electrolyte fluids 5a, 5b.

[0056] Such concentration potentials E a , Eb of the electrolyte liquids 5a, 5b are not accessible by measuring the cell voltage Vocv alone across a single cell 18 in the redox flow battery 1, or generally between the electrolyte liquids 5a, 5b. To determine a concentration potential E a , Eb, reference electrodes 20 are therefore generally used, as shown in Fig.2, which provide a reference potential VR that is as stable as possible, relative to which a concentration potential E a , Eb of an electrolyte liquid 5a, 5b can be measured.

[0057] In a redox flow battery 1, the electrical cell voltage Vocv between the positive electrolyte 5a and the negative electrolyte 5b is typically measured, particularly as a measure of the SoC of the redox flow battery 1 (Fig. 2). The cell voltage Vocv depends on the charge state of the two electrolyte 5a, 5b. If the charge states of the two electrolyte 5a, 5b were equal, i.e., if no charge imbalance existed, the measured cell voltage Vocv would be an accurate estimate of the SoC of the redox flow battery 1.However, if a charge imbalance exists between the two electrolyte fluids 5a, 5b in the redox flow battery 1, particularly due to a concentration imbalance of the redox elements in one or both electrolyte fluids 5a, 5b, then the cell voltage Vocv alone cannot be used to accurately determine the SoC or provide information about the charge state of the electrolyte fluids 5a, 5b or the charge imbalance. In the case of a charge imbalance, the concentration potential E would have to be determined. a , Eb of each electrolyte liquid 5a, 5b relative to a reference potential VR of a reference electrode 20 in order to enable an estimation of the SoC and the charge imbalance in the redox flow battery 1.

[0058] The cell voltage Vocv is therefore dependent on the individual charge states of the electrolyte liquids 5a, 5b and thus on the concentration potentials E a, Eb of the electrolyte liquids 5a, 5b. Each value of the cell voltage Vocv can be determined by different concentration potentials E a , Eb of the electrolyte liquids 5a, 5b. The concentration potential E a , Eb of an electrolyte liquid 5a, 5b can be measured with a reference electrode 20, as shown in Fig.2.

[0059] The invention now uses this relationship 25 between the cell voltage Vocv and a concentration potential E a , Eb of an electrolyte liquid 5a, 5b to detect a charge imbalance in the redox flow battery 1.

[0060] For this purpose, the cell voltage Vocv between the electrolyte liquids 5a, 5b and the concentration potential E a, Eb of at least one electrolyte liquid 5a, 5b relative to a reference electrode 20, preferably the negative electrolyte liquid 5b, during a predetermined measuring time T and the resulting relationship 25 between the cell voltage Vocv and the concentration potential E a , Eb of the electrolyte liquid 5a, 5b were evaluated.

[0061] The relationship 25 between the cell voltage Vocv and the concentration potential E a , Eb of the electrolyte liquid 5a, 5b results from the respective value of the cell voltage Vocv and the value of the concentration potential E a , Eb of an electrolyte liquid 5a, 5b at each time t (usually at time-discrete times) during the measurement period T.

[0062] In Fig.3a, the relationship 25 is visualized in the form of the temporal progression of the cell voltage Vocv and the concentration potential Eb of the negative electrolyte liquid 5b during the measurement period T. The relationship 25 can also be expressed as cell voltage Vocv vs. concentration potential E a , Eb of an electrolyte liquid 5a, 5b curve can be visualized as shown in Fig.3b.

[0063] The visualization in Fig.3a and Fig.3b, and also the visualization explained below in Fig.4 and Fig.5, only serves to explain the invention and is not required in a practical application of the invention.

[0064] In Fig.3b, the relationship 25 is additionally shown in the form of the cell voltage Vocv vs. concentration potential E a, Eb curve for different charge states of the electrolyte liquid 5a, 5b, shown in Fig. 3b for the concentration potential Eb of the negative electrolyte liquid 5b. Depending on the charge state, the relationship 25, here the course of the curve, changes.

[0065] The relationship 25 between the cell voltage Vocv and the concentration potential E a , Eb of the electrolyte liquid 5a, 5b depends on the charge imbalance between the electrolyte liquids 5a, 5b of the redox flow battery 1 and a reference potential VR of the reference electrode 20, as well as on a possible drift D of the reference potential VR of the reference electrode 20.

[0066] In Fig.4, the effect of a drift D of the reference potential VR of the reference electrode 20 on the relationship 25 is visible. In this illustration, the relationship 25 shifts in the form of the curve cell voltage Vocv vs. concentration potential E a, Eb of an electrolyte liquid 5a, 5b due to the drift D in the direction of the concentration potential E a , Eb (as in the case shown).

[0067] In the embodiment shown in Fig. 2, the cell voltage Vocv between the two electrolyte liquids 5a, 5b is measured directly. Alternatively, but equally, the concentration potential E a of the positive electrolyte liquid 5a and the concentration potential Eb of the negative electrolyte liquid 5b are measured and the cell voltage Vocv is determined from this, for example as the difference between the two concentration potentials E a , Eb. However, since a redox flow battery 1 usually always implements a measurement of the cell voltage Vocv as a measure of the SoC, preferably only the concentration potential E aof the positive electrolyte liquid 5a or, preferably, the concentration potential Eb of the negative electrolyte liquid 5b must be measured additionally. After the relationship 25 of the charge imbalance L between positive and negative electrolyte liquid 5a, 5b, and thus also of the charge state of the electrolyte liquid 5a, 5b, for which the concentration potential E a , Eb is measured, the relationship 25 can be evaluated to determine an existing charge imbalance between the electrolyte liquids 5a, 5b, which can be done in various ways.

[0068] In order to evaluate the relationship 25 with regard to the charge imbalance, the relationship 25 determined during the measurement period T can be compared with predetermined reference curves R for different charge imbalances L, whereby each reference curve R represents a specific charge imbalance L. The reference curve R that approximates the relationship 25 with the smallest deviation, for example in the sense of the least squares error, then determines the charge imbalance L. It may be sufficient to use at least one reference curve R, for example for a critical charge imbalance L that should not be exceeded.

[0069] For example, the determined current relationship 25 is compared with known reference curves R for various charge imbalances L, and the deviation of the relationship 25 from the reference curve R, for example, based on the measured values, is determined. The deviations are squared and summed. The reference curve R of a specific charge imbalance L, which results in the smallest sum of squares, then determines the current charge imbalance L of the redox flow battery 1.

[0070] The reference curves R could, for example, be determined empirically by measuring the cell voltage Vocv and the desired concentration potential E on a redox flow battery 1 for certain predetermined or set charge imbalances L (at least one) and predetermined charge and / or discharge cycles. a , Eb is measured.

[0071] The reference curve R could also be determined by inserting a given mathematical function, for example a polynomial of a certain degree, for example a polynomial of the 2nd, 3rd or 4th degree, with given function parameters resulting from the definition of the mathematical function, such as polynomial parameters, into measurement data for the cell voltage Vocv and the at least one concentration potential E a , Eb at a certain known charge imbalance L. This can also be done for different known charge imbalances L in order to obtain different reference curves R for the different charge imbalances L. The measured data can be available or can again be measured empirically as described above. The measured data are available for a measurement period T in the form of values ​​of the cell voltage Vocv and a concentration potential E a , Eb, or both concentration potentials E a, Eb. The given mathematical function is then adapted to the measured data as best as possible. The best possible approximation of the mathematical function to the measured data can be achieved using known methods of fitting calculations. The approximation is carried out by changing the function parameters of the mathematical function, such as the polynomial parameters, so that the measured data are approximated as best as possible, for example in the sense of a small square error, by the mathematical function. The approximated mathematical function then represents the reference curve R for the respective charge imbalance L. After different concentration potentials E a , Eb and thus relationships 25, reference curves R for different charge imbalances L can be determined in this way.

[0072] In order to infer a specific charge imbalance L using the reference curves R and the available current measurement data of a relationship 25, the reference curve R must be determined that best approximates the available measurement data in terms of the smallest deviation. This can be done using known optimization methods, for example, the least squares method.

[0073] From Fig.4 it can be seen that the drift D starting from an initial state causes a shift of the relationship 25 in the direction of the axis of the concentration potential E a , Eb. Such a drift D could also be taken into account in an optimization method by using the drift D as an additional optimization parameter. In this case, the optimization is extended by various shifts of the relationship 25 determined during the measurement period T in the direction of the axis of the concentration potential E a, Eb, to compensate for the influence of drift D.

[0074] However, it is easier to use a drift-independent relationship 25. For this, relationship 25 can be the derivative of the concentration potential E a , Eb according to the cell voltage Vocv, which is known as dE a / dVocv or dEb / dVocv. Since the drift D can be considered a constant offset, the drift D disappears in such a derivation, and only the dependence of relationship 25 on the charge imbalance L remains.

[0075] The derivative of the curve in Fig.3b or in Fig.4 with respect to the cell voltage Vocv can be determined analytically or by means of numerical differentiation, for example by the method of forward difference quotients or central difference quotients; other numerical differentiation methods are of course also applicable.

[0076] The result of such a differentiation for relationships 25 for various charge imbalances L is shown in Fig. 5, again using the example of the concentration potential Eb of the negative electrolyte liquid 5b. In Fig. 5, the negative value of the derivative -dEb / dVocv is used, but this is merely an arbitrary choice. The positive value dEb / dVocv could equally well be used.

[0077] In the example of Fig. 5, the second curve from the left represents the state of charge equilibrium between the two electrolyte liquids 5a, 5b. For a vanadium redox flow battery 1, for example, this would correspond to an average oxidation state of 3.5. To the left of this, a charge imbalance L with an undercharged negative electrolyte liquid 5b is shown. For a vanadium redox flow battery 1, this could correspond to an average oxidation state of 3.45. To the right of the charge-balanced curve, charge imbalances with an overcharged negative electrolyte liquid 5b are shown. For a vanadium redox flow battery 1, this could correspond to an average oxidation state of 3.55 and 3.6.

[0078] These curves of relationship 25 in the form of the derivatives for different charge imbalances L could also be best approximated by given mathematical functions, such as a straight line or a 2nd or 4th degree polynomial with function parameters, for example again using methods of best-fit calculations, in order to determine reference curves R. Here, too, the function parameters of the mathematical function are determined so that relationship 25 obtained from the measured data is best approximated by the mathematical function. In this way, reference curves R for different charge imbalances L could again be obtained, which could then be compared with a current relationship 25 during a measurement period T in order to find the reference curve R, and thus the current charge imbalance L, that best approximates the current relationship 25.

[0079] However, it is more advantageous if the charge imbalance L is considered as an additional function parameter of the mathematical function and for the various relationships 25, as in Fig.5, a reference curve R is determined which is additionally dependent on the charge imbalance L.

[0080] The function parameters of the individual reference curves R for the various charge imbalances L can be considered. Different function parameters result for different charge imbalances L. The function parameters are therefore dependent on the charge imbalance L. This dependence of the function parameters on the charge imbalance L can again be modeled by a mathematical function, such as a straight line or a polynomial of a certain degree, for example, again using methods of best fit calculation. This results in a function that describes the function parameters as a function of the charge imbalance L. The reference curve R is thus described by a mathematical function that has at least one function parameter that depends on the charge imbalance L.

[0081] In one embodiment, the relationships 25 are approximated as in Fig.5 by a reference curve R of the following form, which is dependent on the charge imbalance L. In this, the relationships 25 for the various charge imbalances L are represented by a mathematical function in the form of a second-degree polynomial with the function parameters (polynomial parameters) A, B, C, where the function parameters are linearly dependent on the charge imbalance L, in the form

[0082] A = A1-L + A2 B = B1-L-B2. C = C1-L + C2

[0083] The function parameters A, B, C are modeled as linear (as a straight line) depending on the charge imbalance L, with the straight line parameters A1 , A2, B1 , B2, C1 , O2.

[0084] The charge imbalance L can therefore be considered as an additional function parameter of the mathematical function describing the relationship 25.

[0085] The above reference curve R is written in the form of four equations, whereby the reference curve R could of course be summarized in one equation, i.e. in a mathematical function.

[0086] In a concrete embodiment, the function parameters could, for example, be

[0087] A = 34.853 -L + 5.3087 = -119.69-1-14.71 0 = 99.449-1 + 10.6.

[0088] Of course, there are also other charge imbalance L-dependent

[0089] Reference curves R (mathematical functions) are conceivable, i.e., reference curves R with the charge imbalance L as an additional function parameter. For example, the curve of a relationship 25 can also be modeled with a reference curve R in the form of a polynomial with a higher or lower degree than two, for example, with a 4th-degree polynomial. Likewise, the function parameters of the reference curve R can be modeled in a manner other than linearly dependent on the charge imbalance L, for example, with a 2nd-degree polynomial.

[0090] If a current curve of the derivative -dEb / d ocv is determined as relationship 25 using current measurement data during a measurement period T, then an optimization procedure can be used to determine the charge imbalance L of the underlying mathematical model that best approximates the current measurement data. To do this, the charge imbalance L of the reference curve is varied in order to minimize the deviation, e.g., in the form of the least squares error, between the reference curve R and the current measurement data. The charge imbalance L with the minimal deviation between the reference curve and the measurement data is then the desired current charge imbalance L of the redox flow battery 1. In this way, the current charge imbalance L can be easily determined.

[0091] A relationship 25 between the cell voltage Vocv and the determined at least one concentration potential E a, Eb can be determined in such a way that the concentration potentials E a , Eb of both electrolyte liquids 5a, 5b are determined with a reference electrode 20. Both concentration potentials E a , Eb can then be derived from the cell voltage Vocv, as explained above. Then the ratio of both derivatives can be dE a / dVocv and dEb / dVocv are formed, and the ratio, dE / dl' ,.,. dE h / dy nr so — - - or — - - , can be used as context 25. This gives dE b !dV ocv dE a ! dV ocvagain a curve as relationship 25, where the course of the curve depends on the charge imbalance L. Such a relationship 25 is of course also independent of a possible drift D of the reference electrode 20. For such curves as relationship 25, as described above, reference curves R for different charge imbalances L can be created, or a reference curve R dependent on the charge imbalance L can be created.

[0092] A reference curve R depending on the charge imbalance L can of course also be used for a relationship 25 in the form of a curve of the cell voltage Vocv versus at least one concentration potential E a , Eb (as shown in Fig.3b or Fig.4) can be used.

[0093] The reference curves can be determined in advance for a redox flow battery 1 and are then available for determining the charge imbalance L. To determine the charge imbalance during operation of the redox flow battery 1, a known reference curve R can be used to determine the current charge imbalance L from a comparison of a current relationship 25, which is determined on the basis of current measurement data, with at least one reference curve R. A known reference curve R can be compared with a current relationship 25 resulting from current measurement data in order to determine a current charge imbalance L of the redox flow battery 1. In this way, a reference curve R can be determined for a specific charge imbalance L which comes closest to the current relationship, i.e., for example, minimizes a deviation.Or the deviation, for example, as a least squares error or another suitable measure, between a reference curve R for a specific charge imbalance L and the current relationship 25 can be determined, and the charge imbalance L can be assumed if the deviation becomes smaller than a predetermined limit. Or a reference curve R dependent on the charge imbalance L can be used, and by varying the charge imbalance, the reference curve can be compared with a current relationship 25 resulting from current measurement data, for example, until the deviation is minimized.

[0094] The charge imbalance L can be expressed, for example, in percent, with 0% for electrolyte liquids 5a, 5b with balanced charge states, or in the form of an average oxidation number.

[0095] Although self-evident, it should be noted that the reference curve R naturally shows the same relationship between cell voltage Vocv and at least one concentration potential E a , Eb describes how the current relationship 25, which results during the measurement period T, can naturally not be compared, for example, with a relationship according to Fig.3b with a relationship according to Fig.5.

[0096] For the invention, it is advantageous if the temporal drift D of the reference electrode 20 is significantly slower than the measurement duration T. The measurement duration T can, for example, comprise several charge / discharge cycles of the redox flow battery 1, which typically covers a period of several hours. The temporal drift D of a reference electrode 20 is typically considerably slower and therefore not noticeable in a measurement during the measurement duration T. The temporal drift D usually only becomes noticeable over periods of days or weeks.

[0097] Likewise, it is advantageous for the invention that a charge imbalance L between the electrolyte liquids 5a, 5b develops only very slowly, typically no more than a few percent per year. Thus, the charge imbalance L can also be considered constant during a measurement period T.

[0098] The advantage of evaluating the relationship 25 between the cell voltage Vocv and a concentration potential E a , Eb of at least one electrolyte liquid 5a, 5b is that, on the one hand, only electrical voltages need to be measured and, on the other hand, the influence of a possible drift D of the reference electrode 20 can be easily eliminated from the evaluation. Either because the drift D has already been eliminated from the relationship 25, as in the case of the derivative dEb / dVocv or dE a / dVocv, or because the drift D is taken into account when comparing the measured data with a reference curve R. This, in turn, allows for great flexibility in the choice of the reference electrode 20. In particular, a reference electrode 20 with a non-stable reference potential VR can also be used. The reference potential VR of the reference electrode 20 does not even have to be clearly defined or known.

[0099] Thus, very simple, cost-effective reference electrodes 20 can be used with the invention and the reference electrodes mentioned at the beginning with all their disadvantages can be dispensed with.

[0100] Such a simply constructed reference electrode 20 is described with reference to Fig.6.

[0101] The reference electrode 20 consists of a preferably straight line section 23, which branches off at a branch 22 from a line 21 carrying the electrolyte liquid 5a, 5b, through which the electrolyte liquid 5a, 5b is circulated. The line 21 can, for example, be a line 17 of the redox flow battery 1. The end of the line section 23 opposite the branch 22 is closed by a first, electrically conductive reference electrode element 24. Thus, there is no flow of the electrolyte liquid 5a, 5b through the line section 23, but rather a stationary column of electrolyte liquid forms in the line section 23.A change in the charge state (concentration ratio of the redox pair) of the electrolyte liquid 5a, 5b in the line 21 will only become noticeable very slowly at the first reference electrode element 24 because the charge state spreads only slowly by diffusion through the electrolyte liquid column to the first reference electrode element 24. This achieves a sufficiently stable reference potential VR at the first reference electrode element 24, at least during the measurement period T.

[0102] To measure the concentration potential E a, Eb relative to the reference potential VR, a second electrically conductive reference electrode element 26 can be arranged in the line 21, preferably in the region of the branch 22, in contact with the flowing electrolyte liquid 5a, 5b, as shown in Fig.6. The current concentration ratio of the redox pair used in the electrolyte liquid 5a, 5b is always present at the second reference electrode element 26.

[0103] The concentration potential E a , Eb can then be measured, for example, with a voltage sensor between the first reference electrode element 24 and the second reference electrode element 26.

[0104] A first and second reference electrode element 24, 26 can consist of any electrically conductive material that is sufficiently stable with respect to the electrolyte liquid 5a, 5b, for example plastic-carbon composite materials, graphite, graphite felt, etc.

[0105] To measure the cell voltage Vocv, a measuring cell 30 is often used, which is essentially constructed like a single cell 18 of the cell stack 10, as schematically shown in Fig. 7. The measuring cell 30 thus also consists of a positive half measuring cell 30a, in which a positive measuring cell electrode is arranged and through which the positive electrolyte liquid 5a flows, and a negative half measuring cell 30b, in which a negative measuring cell electrode is arranged and through which the negative electrolyte liquid 5a flows. A semipermeable, in particular ion-selective, membrane 31 is arranged between them. The cell voltage Vocv between the two electrolyte liquids 5a, 5b can be tapped via the measuring cell electrodes, for example with a voltage sensor.

[0106] A measuring cell 30 for measuring the cell voltage Vocv, as shown in Fig. 7, can advantageously be combined with a reference electrode 20, in particular a reference electrode 20 as described in Fig. 6, as shown in Fig. 8. The negative half measuring cell 30b is flowed through by the negative electrolyte liquid 5b. In this embodiment, the discharge line of the negative electrolyte liquid 5b from the negative half measuring cell 30b is simultaneously used as line 21 from which the line section 23 of the reference electrode 20 branches off. It is also advantageous if the branch 22 of the line section 23 is arranged as close as possible to the negative half measuring cell 30b in order to avoid electrical voltage propagation effects along the line 21 between the negative half measuring cell 30b and the branch 22. However, the reference electrode 20 could also be arranged in the supply line of the negative electrolyte liquid 5b to the negative half measuring cell 30b.In this way, the concentration potential E could also be determined. a the positive electrolyte liquid 5a.

[0107] Instead of a second reference electrode element 26 in the reference electrode 20, as in Fig. 6 or 7, the negative measuring electrode of the measuring cell 30 could also be used, as shown in Fig. 9. At the measuring cell 30, an electrical contact must be made between the negative measuring cell electrode in the negative half-cell 30b and the outside as a negative measuring electrode. The negative concentration potential Eb can then be measured directly against this negative measuring electrode. In this embodiment, too, the line 21 of the reference electrode 20 could again be arranged in the supply line of the negative electrolyte liquid 5b. Likewise, it is again advantageous if the branch 22 of the line section 23 of the reference electrode 20 is arranged as close as possible to the negative half-measuring cell 30b in order to avoid electrical voltage propagation effects along the line 21 between the negative half-measuring cell 30b and the branch 22.Likewise, the concentration potential E could of course also be affected. a the positive electrolyte fluids 5a are measured.

[0108] To determine a charge imbalance L between a positive electrolyte liquid 5a and a negative electrolyte liquid 5b of a redox flow battery 1, a cell voltage Vocv between the positive electrolyte liquid 5a and the negative electrolyte liquid 5b and a concentration potential E a , Eb of at least one of the positive and negative electrolyte liquids 5a, 5b. The concentration potential E a , Eb is measured with a reference electrode 20 with a reference potential VR. The cell voltage Vocv is measured, for example, with a measuring cell 30 as described in Figs. 7, 8, 9, or based on the two concentration potentials E, each measured with a reference electrode 20, preferably with the same reference potential R.a , Eb of the positive electrolyte liquid 5a and the negative electrolyte liquid 5b are determined. From the measurement data during the measurement period T, the current relationship 25 between the cell voltage Vocv and the at least one concentration potential E a , Eb are determined. This current relationship 25 is evaluated to determine the charge imbalance L between the electrolyte fluids 5a, 5b of the redox flow battery 1. One of the methods described above can be used for this purpose. The charge imbalance L thus reflects the state of the redox flow battery 1 at the end of the measurement period T.

[0109] The invention was tested on a 10 kW vanadium redox flow battery 1. The vanadium redox flow battery 1 is equipped with a measuring cell 30 for measuring the cell voltage Vocv as described in Fig. 7. The concentration potential Eb of the negative electrolyte liquid 5b was measured using a reference electrode 20, as shown in Fig. 6, which was combined with the measuring cell 30 as described with reference to Fig. 8. In this embodiment, the line section 23 of the reference electrode 20 was 12 cm long with an inner diameter of 6 mm. At the end of the line section 23, a cylindrical rod made of a plastic-graphite composite material with an outer diameter of 6 mm was arranged as the first reference electrode element 24. The second reference electrode element 26 was also made of a plastic-graphite composite material with a diameter of 6 mm and was arranged in the circulation line 17 of the negative electrolyte liquid 5b as line 21.The vanadium redox flow battery 1 had been in operation for 8 years at the time of the experiment and showed a charge imbalance L between the electrolyte liquids 5a, 5b of +7.7 ±0.5% (determined for the experiment by UV / VIS spectroscopy); the electrolyte liquids 5a, 5b were therefore overoxidized.

[0110] To determine the charge imbalance L, a relationship 25 in the form of the derivative of the negative concentration potential Eb with respect to the cell voltage Vocv, dEb / dVocv, was used, as shown in Fig. 5. A reference curve dependent on the charge imbalance L in the form of a second-order polynomial for the derivative dEb / dVocv and with polynomial parameters A, B, C linearly dependent on the charge imbalance L, as described above, was used to determine the charge imbalance L as described above.

[0111] A sliding period of 48 hours was set as the measurement duration T. The cell voltage Vocv and the concentration potential Eb were measured every second. The time intervals between the measurements could, of course, also be chosen differently. In each case, 2 hours of measurement data were added to the end of the sliding period of 48 hours, and at the same time 2 hours of measurement data were removed from the beginning of the 48-hour period. The measurement data for the inventive evaluation of the charge imbalance was thus updated 50 times over a period of 100 hours, which corresponded to approximately 2.5 charge-discharge cycles, and the charge imbalance was determined in each case. For all 50 measurement data with a measurement duration of 48 hours, the derivative dEb / dVocv and thus the relationship 25 were determined. The measurement data were compared with the reference curve R by adjusting the reference curve R as best as possible to the current measurement data, i.e., to the 50 current relationships 25 obtained, by varying the charge imbalance L.This resulted in 50 charge imbalances L of the redox flow battery 1. This process is shown in Fig.10a and 10b.

[0112] Fig. 10a shows the temporal progression of the cell voltage Vocv and the concentration potential Eb of the negative electrolyte liquid 5b over a period of 100 hours, as well as a sliding period of 48 hours as the measurement duration T. Fig. 10b shows the progression of the charge imbalance L determined according to the invention for the 50 measurement durations within a period of 100 hours. The charge imbalance L was determined in the range of 7% to 8%, with an average charge imbalance L of 7.5%. This corresponds to the charge imbalance of +7.7 ±0.5% determined by UV / VIS spectroscopy.

[0113] Fig.11 shows an arrangement for determining the charge imbalance L between a positive electrolyte liquid 5a and negative electrolyte liquid 5b of a redox flow battery 1, wherein a cell voltage detection unit 2 is provided, which is configured to determine a cell voltage Vocv between the positive electrolyte liquid 5a and negative electrolyte liquid 5b during operation of the redox flow battery 1 during a measurement period T, and a concentration potential detection unit 3 is provided, which is configured to determine a concentration potential E during operation of the redox flow battery 1 a, Eb of at least one of the positive and negative electrolyte liquids 5a, 5b. The cell voltage detection unit 2 is, for example, a measuring cell 30 as shown in Fig.7, 8 or 9. The concentration potential detection unit 3 is, for example, any reference electrode 20, for example as shown in Fig.6. Furthermore, an evaluation unit 4 is provided in the arrangement, which is preferably a microprocessor-based hardware with evaluation software or an integrated circuit, such as a Field Programmable Gate Array (FPGA) or an Application Specific Integrated Circuit (ASIC). The evaluation unit 4 receives the cell voltage Vocv from the cell voltage detection unit 2 and the at least one concentration potential E a, Eb from the concentration potential detection unit 3. In addition, the evaluation unit 4 can also receive a reference curve R or a reference curve R dependent on the charge imbalance L, wherein the reference curve R can also be stored in the evaluation unit 4. The evaluation unit 4 determines for the period of the measurement duration T a current relationship 25 between the determined cell voltage Vocv and the determined at least one concentration potential E a , Eb and evaluates this current relationship 25 to determine the current charge imbalance L of the redox flow battery 1.

Claims

Patent claims 1 . Method for determining the charge imbalance (L) between a positive electrolyte liquid (5a) and a negative electrolyte liquid (5b) of a redox flow battery (1), wherein during operation of the redox flow battery (1) during a predetermined measuring period (T) a cell voltage (Vocv) between the positive electrolyte liquid (5a) and the negative electrolyte liquid (5b) and a concentration potential (E a , Eb) of at least one of the positive and negative electrolyte liquids (5a, 5b), characterized in that for the period of the measurement duration (T) a current relationship (25) between the determined cell voltage (Vocv) and the determined at least one concentration potential (E a , Eb) is determined and the current relationship (25) is evaluated to determine the charge imbalance (L).

2. Method according to claim 1, characterized in that for the period of the measurement duration (T) a curve of the cell voltage (Vocv) versus the at least one concentration potential (E a , Eb) is determined and this curve is used as the current relationship (25).

3. Method according to claim 1, characterized in that for the period of the measurement duration (T) a curve of the cell voltage (Vocv) versus the at least one concentration potential (5a, 5b) is determined, that the determined curve is derived according to the cell voltage (Vocv), and that the derivative of the curve is used as the current relationship (25).

4. Method according to claim 1, characterized in that for the period of the measurement (T) a curve of the cell voltage (Vocv) versus the concentration potential (E a) of the positive electrolyte liquid (5a) and a curve of the cell voltage (Vocv) versus the concentration potential (5b) of the negative electrolyte liquid (5b) is determined, that the determined curves are each derived according to the cell voltage (Vocv), and that a ratio of the derivatives of the curves is used as the relationship (25).

5. Method according to claim 2 or 3 or 4, characterized in that the relationship (25) is compared with a predetermined reference curve (R) for a specific charge imbalance (L) and the charge imbalance (L) underlying the reference curve (R) is used as the charge imbalance (L) of the redox flow battery (1) if a deviation between the reference curve (R) and the current relationship (25) is smaller than a predetermined limit value.

6. Method according to claim 2 or 3 or 4, characterized in that the relationship (25) is compared with predetermined reference curves (R) for different charge imbalances (L) and the reference curve (R) is determined which minimizes a deviation between the current relationship (25) and the reference curve (R) and the charge imbalance (L) of the redox flow battery (1) results from the underlying charge imbalance (L) of the determined reference curve (R).

7. Method according to claim 2 or 3 or 4, characterized in that a predetermined reference curve (R) dependent on charge imbalance (L) is used and the charge imbalance (L) of the reference curve (R) is varied, and the charge imbalance (L) is determined which minimizes a deviation between the current relationship (25) and the reference curve (R) and this charge imbalance (L) is used as the charge imbalance (L) of the redox flow battery (1).

8. Arrangement for determining the charge imbalance (L) between a positive electrolyte liquid (5a) and a negative electrolyte liquid (5b) of a redox flow battery (1), wherein a cell voltage detection unit (2) is provided which is configured to determine a cell voltage (Vocv) between the positive electrolyte liquid (5a) and the negative electrolyte liquid (5b) during operation of the redox flow battery (1) during a predetermined measuring period (T), and a concentration potential detection unit (3) is provided which is configured to determine a concentration potential (E a , Eb) of at least one of the positive and negative electrolyte liquids (5a, 5b), characterized in that an evaluation unit (4) is provided which is designed to determine, for the period of the measurement duration (T), a current relationship (25) between the determined cell voltage (Vocv) and the determined at least one concentration potential (Ea , Eb) and evaluate the current relationship (25) to determine the charge imbalance (L).

9. Arrangement according to claim 8, characterized in that the evaluation unit (4) for the period of the measurement duration (T) generates a curve of the cell voltage (Vocv) versus the at least one concentration potential (E a , Eb) and this curve is used as the current relationship (25).

10. Arrangement according to claim 8, characterized in that the evaluation unit (4) for the period of the measurement duration (T) generates a curve of the cell voltage (Vocv) versus the at least one concentration potential (E a , Eb), the determined curve is derived with respect to the cell voltage (Vocv), and the derivative of the curve is used as the current relationship (25). 11 . Arrangement according to claim 8, characterized in that the evaluation unit (4) for the period of the measurement duration (T) generates a curve of the cell voltage (Vocv) versus the positive concentration potential (E a ) and a curve of the cell voltage (Vocv) versus the negative concentration potential (Eb) is determined, the determined curves are each derived with respect to the cell voltage (Vocv), and a ratio of the derivatives of the curves is used as the current relationship (25).

12. Arrangement according to claim 9 or 10 or 11, characterized in that the evaluation unit (4) is set up to compare the current relationship (25) with a predetermined reference curve (R) for a specific charge imbalance (L) and to determine a deviation between the reference curve (R) and the current relationship (25) and to use the charge imbalance (L) underlying the reference curve (R) as the charge imbalance (L) of the redox flow battery (T) if the deviation is smaller than a predetermined limit value.

13. Arrangement according to claim 9 or 10 or 11, characterized in that the evaluation unit (4) is set up to compare the current relationship (25) with predetermined reference curves (R) for different charge imbalances (L) and to determine the reference curve (R) which minimizes a deviation between the relationship (25) and the reference curve (R) and the charge imbalance (L) of the redox flow battery (1) results from the determined reference curve (R).

14. Arrangement according to claim 9 or 10 or 11, characterized in that the evaluation unit (4) is set up to use a predetermined reference curve (R) dependent on charge imbalance (L) and to vary the charge imbalance (L) of the reference curve (R), and to determine the charge imbalance (L) which minimizes a deviation between the relationship (25) and the reference curve (R) and this charge imbalance (L) results in the charge imbalance (L) of the redox flow battery.

15. Arrangement according to one of claims 8 to 14, characterized in that a reference electrode (20) is provided as the concentration potential detection unit (3), and the reference electrode (20) has a line piece (23) which is connected to a branch (22) of a line supplied from the electrolyte liquid (5a, 5b) for which the concentration potential (E a, Eb) is to be determined, wherein at an end of the line section (23) opposite the branch (22) a first, electrically conductive reference electrode element (24) is arranged, which closes the line section (23) and a second, electrically conductive reference electrode element (26) is arranged in the electrolyte liquid (5a, 5b), so that between the first reference electrode element (24) and the second reference electrode element (25) the concentration potential (E a , Eb).

16. Arrangement according to claim 15, characterized in that the second reference electrode element (26) is arranged in the region of the branch (22).

Images