WO2022128191A1 - Method and apparatus for determining the electrolyte system state in a redox flow battery - Google Patents

Abstract

A method for determining an electrolyte system state, in particular a shift of an oxidation state of the electrolyte, in a redox flow battery is disclosed, in particular a VRFB battery, comprising the following steps: (a) filling each half-cell of a sensing cell with an anolyte and a catholyte; (b) connecting a voltage source in series with the sensing cell; (c) closing a current circuit of the sensing cell across a resistor; (d) discharging the sensing cell while monitoring current and voltage of the sensing cell as a function of time; and (e) determin- ing the electrolyte system state, particularly the shift in oxidation state, from the current and voltage values during discharge, from the concentrations of anolyte and catholyte, from the volumes of half cells (16, 18), and from the volumes of anolyte and catholyte.

Description

METHOD AND APPARATUS FOR DETERMINING THE ELECTROLYTE SYSTEM STATE IN A REDOX FLOW BATTERY

[0001] The invention relates to a method and apparatus for determining the electrolyte system state in a redox flow battery.

[0002] In a redox flow battery, electrical energy is stored in the chemical energy of the electrolyte system consisting of an anolyte and a catholyte. Both electrolytes are located in separate tanks. The electrolytes contain ions of different oxidation states. When the battery is charged, ions of the anolyte are reduced and ions of the catholyte are oxidized. When discharging, the opposite happens. In operation, the anolyte from the anolyte tank and the catholyte from the catholyte tank are each pumped through the two half-cells of a work cell by means of a pump. During discharging, ions are exchanged at the membrane between the two half-cells of the work cell and electrical energy is delivered to the electrodes, and during charging, energy is stored in the anolyte and the catholyte by reversing the exchange processes. A combination of several working cells into a multicellular unit is called a stack. An RFB unit (redox flow battery unit) consists at least of a working cell with its electrical contacting and the two electrolyte circuits with the tanks of anolyte and catholyte, which are usually actively flowed through.

[0003] If, during charging, as many charge carriers are available for reduction in the anolyte as charge carriers are available for oxidation in the catholyte, then the electrolyte system is balanced. If more charge carriers are available for reduction in the anolyte than charge carriers are available for oxidation in the catholyte, then the electrolyte system is positively (oxidatively) shifted. If more charge carriers are available for oxidation in the catholyte than charge carriers are available for oxidation in the anolyte, then the electrolyte system is negatively (reductively) shifted. The number of charge carriers available for reaction on the anolyte and catholyte sides depends on the volume of the anolyte and catholyte, the respective charge carrier concentration and the respective oxidation state of the charge carriers. The electrolyte system state is described by the volume of the anolyte and catholyte, and by the concentration and oxidation state of the charge carriers in the anolyte and catholyte. Another term for the electrolyte system state is "State of Health" (SOH). A complete determination of the electrolyte system state includes the determination of the volume and concentration difference of the anolyte and catholyte as well as the shift of the oxidation state. If no volume and concentration differences occur, the shift of the electrolyte system state corresponds to that of the oxidation state. Another term for the average oxidation state of the electrolyte system is "Average Oxidation State" (AOS).

[0004] Both a positive and negative shift reduces the amount of charge (capacity) that can be stored in a flow battery because not every potential reactant on one side of the electrolyte has a matching reactant on the other side of the electrolyte. The product of electrochemical potential difference and charge quantity is the energy quantity. Since a shift occurring in practice only slightly changes the average electrochemical potential difference between anolyte and catholyte in the usable range, a reduced storable charge quantity immediately follows from a reduced storable energy quantity.

[0005] There are various flow batteries with different reversible redox systems. A frequently used flow battery is the VRFB battery (Vanadium Redox Flow Battery). This takes advantage of the fact that vanadium can occur as divalent, trivalent, tetravalent and pentavalent vanadium. A VRFB battery uses redox pairs of vanadium in the two half cells. In this process, the vanadium ions of different oxidation states are dissolved in an aqueous medium. Common are sulfuric and hydrochloric acid media as well as mixtures. If the medium is sulfuric acid, then the electrolyte contains vanadium sulfate (VOSO₄) on the positive side, which can be oxidized to pentavalent ions.

[0006] Positive electrode, V^lv and V^v:

VO²⁺ + H₂O O VO₂ ⁺ + 2H⁺ + e". [0007] A sulfuric acid vanadium electolyte of the negative pole side contains vanadium (III) sulfate, which can be reduced to the divalent vanadium salt: Negative electrode, negative pole:

[0008] During regular operation of flow batteries, different transport rates of ions and electrolyte across the membrane, which cause concentration and volume differences, as well as gasgenerating side reactions, especially during charging or in the charged state, can cause shifts between the two half-cells, which, as mentioned above, lead to capacity loss.

[0009] Furthermore, in some electrolyte systems, positive (oxidative) shifts of the oxidation state may occur over time during practical operation, which may lead not only to a loss of capacity, but also to damage of the battery by oxidation of the bipolar plates or other materials in the stack, which may be associated with further serious damage (e.g., electrolyte contamination, stack leakage). Furthermore, gas developments can occur, which have a damaging effect. The extent to which an electrolyte shifts positively or negatively in the long term depends on the electrolyte system, the materials used in the stack, the system design and the operating parameters.

[0010] Therefore, when operating a flow battery, it is useful to monitor whether a shift occurs and then compensate from time to time by adding reactants to ensure the battery has the highest possible capacity and to avoid damage.

[0011] In the prior art, various methods are known for determining a shift in the oxidation state of a flow battery.

[0012] Basically, according to potentiometry, electrolyte samples can be taken with a pipette and then analyzed offline with a potentiometer for this purpose. This is a very accurate method, but can only be realized in the environment of a chemical laboratory. Other known methods are LIV/VIS spectroscopy, monitoring of viscosity changes, monitoring of density changes or using model calculations from the development of the OCV voltage during charging (Open Circuit Voltage). [0013] From WO 2018/047079 A1 it is known for the determination of the oxidation shift to first completely reset the electrolyte to the initial state before a measurement. For this purpose, the anolyte tank and the catholyte tank are completely mixed with each other. This causes a complete loss of the electrical energy still stored in the electrolyte and an unnecessary heat input by discharging the entire tank. In addition, the process requires intensive mixing of the anolyte and catholyte volumes. Depending on tank size and system design, this can take as little as a few hours or days to complete. After mixing of the two electrolytes, pre-charging is followed by charging of the battery. By monitoring the potential jumps that occur during this process, the shift in oxidation state can be determined.

[0014] As mentioned above, as a result of the necessary mixing of the anolyte and catholyte prior to measurement, the method is hardly applicable in practice, especially above a certain tank size, and leads to considerable energy losses and a considerable expenditure of time. Mixing compensates for concentration differences between anolyte and catholyte. If volume differences between anolyte and catholyte are not compensated, a shift in electrolyte system state is measured based on both the shift in oxidation state and the shift due to volume differences. If a redox flow battery uses ions on the anolyte and catholyte sides, which may not be intermixed, this method is not applicable anyway.

[0015] From WO 2012/135473 A2 it is further known to use in a redox flow battery a sensing cell through which the electrolytes flow and wherein a charge or discharge takes place. Here, a colorimetric measurement method is used, wherein the supplied charge is taken into account with a known charge current in an OCV (open circuit voltage) measurement. Here, the measurement result is influenced by both the concentration and the oxidation state of the ions on the anolyte and catholyte sides.

[0016] Furthermore, a measurement method using a reference electrode is disclosed. Even without a reference electrode, a measurement can be made wherein only the catholyte or only the anolyte is placed in a test cell and then charged in a controlled manner while the OCV voltage is measured over time. Here, different samples of the anolyte and the catholyte are placed in the two chambers of a test cell and then loaded in a controlled manner to specific end points while the OCV voltage is monitored. The concentrations of the two electrolytes are calculated from the measured charging times, and finally the displacement between the two electrolytes is determined from this.

[0017] In another measurement method, a chrono-potentiometric method is used wherein the OCV voltage in a test cell is monitored over time. This is a very time-consuming and relatively inaccurate procedure. Another method consists in the evaluation of light measurements.

[0018] In view of this, it is an object of the invention to disclose a measurement method which is simple and inexpensive and which is suitable for monitoring the electrolyte system state, in particular for monitoring the shift in the oxidation state, in a flow battery and which permits measurement which is as precise as possible with high repeatability.

[0019] Furthermore, a method for compensating a shift of an electrolyte system state, in particular an oxidation state, in a redox flow battery shall be disclosed. Further a suitable device for determining an electrolyte system state, in particular a shift of the oxidation state, shall be disclosed.

[0020] According to the invention, this task is solved by a method for determining an electrolyte system state, in particular a shift of an oxidation state of the electrolyte, in a redox flow battery, in particular a VRFB cell, comprising the following steps:

(a) filling one half-cell each of a sensing cell with an anolyte and a catholyte;

(b) connecting a voltage source in series with the sensing cell;

(c) closing a circuit of the sensing cell;

(d) discharging the sensing cell while monitoring the current and voltage of the sensing cell as a function of time; and

(e) determining an electrolyte system state, particularly the shift in oxidation state, from the current and voltage values during discharge, from the concentrations of anolyte and catholyte, and from the volumes of the half cells. [0021] With respect to the device, the invention is further solved by a device for determining an electrolyte system state, in particular a shift of an oxidation state of the electrolyte, in a redox flow battery (working battery), in particular a VFRB battery, comprising: a sensing cell with two half cells separated by a membrane, one of which can be filled with the anolyte and one with the catholyte from the working battery; a voltage source which can be connected in series with the sensing cell and can be connected via a resistor to a closed circuit for discharging and subsequently recharging the sensing cell; detector for detecting a voltage change and for detecting the charge of the sensing cell when discharging; and computer for determining an electrolyte system state, in particular the shift of the oxidation state, from the current and voltage values during discharge, from the concentrations of anolyte and catholyte, and from the volumes of the half cells.

[0022] The invention is completely solved in this way.

[0023] According to the invention, by monitoring the current and voltage of the sensing cell during a discharge process, a possible shift of the average oxidation state can be determined from the current and voltage values during the discharge, from the concentrations of anolyte and catholyte, and from the volumes of the half cells.

[0024] This is a very simple and reliable method that can be applied "in-line" (i.e. , during operation) and wherein automated evaluation can be performed.

[0025] According to one embodiment of the invention, the discharge of the sensing cell is performed up to a termination criterion that ensures that the occurrence of a first and usually second inflection point of the voltage over time (equivalence point) is part of the measurement, whereby the electrolyte system state, resp. in particular the shift of the oxidation state, is calculated from the charge transferred between the equivalence points, the given or additionally determined concentrations of anolyte and catholyte and the volumes of the half cells and, if necessary, from the given or additionally determined volumes of anolyte and catholyte. [0026] Here, for the special case that no second equivalence point should occur for equal halfcell volumes, it is determined that there is no shift in oxidation state; where for the general case that no second equivalence point should occur for unequal half-cell volumes, it is determined that the shift in oxidation state has been unambiguously determined by a measurement.

[0027] Part of the termination criterion is preferably that a [dependent on the specific design (current load, internal resistance of the sensing cell)] minimum voltage difference is crossed during discharge, so that it is ensured that not only the first equivalence point, but also a second equivalence point is detected, if it exists at all.

[0028] According to a further embodiment of the invention, the discharge of the sensing cell is terminated when there is no further change over a period of time greater than a discharge shoulder, i.e. no further change in voltage over time for a period of time, and/or a minimum discharge voltage difference has been traversed, or when a period of time has elapsed during which the discharge shoulder(s) is (are) known to occur.

[0029] According to a further embodiment of the invention, the discharge of the sensing cell is terminated when a period of time has elapsed that is greater than discharge shoulders to be detected and their time intervals until they occur are long.

[0030] According to a further embodiment of the invention, in step (b) a constant current source is used as a voltage source, which is capable of supplying a current sufficiently large for a defined discharge of the sensing cell over the entire voltage range occurring during the measurement and whose voltage range is preferably larger than the OCV voltage of the sensing cell, which in the current-loaded case has a higher voltage than the OCV voltage of the sensing cell. In this way, it is ensured that in any case a discharge can take place until either a second equivalence point occurs or there is no further voltage change over a period of time greater than a discharge shoulder can theoretically be wide for a given current, electrolyte and given design of the sensing cell, and/or a minimum discharge voltage difference has been crossed. [0031] According to another embodiment of the invention, the electrolyte system state for a redox flow battery is determined according to the equations:

wherein:

^■t,ges,i and Zt,ges,2, are the two possible average oxidation states of the entire electrolyte system of the RFB unit,

ZK,SOC O% and z_{A S}oc o%, are the theoretical oxidation states of the catholyte and anolyte ions, respectively, in a discharged battery (state of charge (SOC): 0%)

V is the volume,

K refers to the catholyte

A refers to the anolyte, c is the concentration of redox ions, t refers to the amount of electrolyte in the tanks and electrolyte circuits of the RFB unit, s refers to the sensing cell,

Aq1 , Aq2 are the charge differences from the start of discharge to the first equivalence point and to the second equivalence point, respectively, and

F is the Faraday constant. [0032] With this general displacement formula, which is derived from the balancing of charge Q, volume V and the amount of substance n, where it holds that a charge is always coupled to a charge carrier (Q = n - z - F), and the amount of substance can be written as the product of concentration and volume (n = c - V), the average oxidation state of the entire electrolyte system of the RFB unit Zt,_ges,i or z_t,g_es,2 for the electrolyte in the tanks and electrolyte circuits of the RFB unit. If this differs from the average oxidation state, where exactly one reactant is assigned to anolyte and catholyte, i.e. , the electrolyte system is balanced, then the difference results in the shift of the electrolyte system state, which takes into account the shift of theoxidation states, the concentrations and the volumes of anolyte and catholyte.

[0033] If the VRFB battery is one with ZK.SOC O% = 4 and z_A, soc o%,= 3 and assuming further that the volumes of the half cells of the sensing cell are the same for the anolyte and the catholyte (V_SA = V_SK = V_Hz), that the volumes of anolyte and catholyte in the tanks and electrolyte circuits of the VRFB unit are the same (VA = V_tK), and that the ion concentrations in anolyte and catholyte in the tanks and electrolyte circuits of the VRFB unit and in the half-cells of the sensing cell are the same for catholyte and anolyte (CSK = C_SA = Ct_K = CtA = c), then the two possible oxidation states are obtained as:

and

[0034] wherein Aq1 and Aq2 are the charge differences from the beginning of the discharge to the first equivalence point and to the second equivalence point, respectively, and F is the Faraday constant. [0035] For a vanadium electrolyte system (VRFB battery), the average oxidation state of the balanced system is 3.5 for the same anolyte and catholyte volumes and concentrations.

[0036] Correspondingly simplified shift formulas can be given for other redox flow systems.

[0037] The two possible oxidation states of the entire electrolyte system of the RFB unit z_t,_ges,i and z_t,ges,2 depend on whether the electrolyte system is oxidatively (positively) or reduc- tively (negatively) shifted.

[0038] According to a further feature of the invention, the process steps (a) to (e) are repeated at least with changed parameters of the sensing cell, and/or with a changed SOC, and further parameters are determined from a comparison of the different measurements.

[0039] By repeating the measurement with a changed volume of the half cells of the sensing cell and/or with a changed SOC charge state, it can thereby be determined from a comparison with the value calculated in at least one other measurement for the shift of the average oxidation state of the entire electrolyte system of the RFB unit whether it is an oxidative or a reductive shift.

[0040] According to a further embodiment of the invention, a complete determination of all independent state variables comprising concentrations, volumes and oxidation states in the anolyte and catholyte (z_Ko, c_K,V_tK, z_Ao, c_A, V_tA) of the electrolyte system are performed, where the state variables (z_Ko, c_K,V_tK, z_A0, c_A, V_tA) are determined stepwise by measurements with changed state variables or changed SOCs and comparison with the results of the previous measurements.

[0041] By performing a sufficient number of measurements with changed state variables or changed SOCs, all state variables of the electrolyte system can thus be determined, as far as they are not already known. This is a considerable advantage compared to conventional methods, with which no complete determination of the state variables was possible up to now. [0042] In this case, the process steps (a) to (e) can be repeated at least with changed parameters of the sensing cell, and/or with a changed SOC, and further parameters are determined from a comparison of the different measurements.

[0043] By repeating the measurement with a changed volume of the half cells of the sensing cell and/or with a changed SOC charge state, it can thereby be determined, for example, from a comparison with the value of the shift of the oxidation state calculated in at least one other measurement, whether it is an oxidative or a reductive shift.

[0044] According to another feature of the invention, the sensing cell is disconnected from the RFB unit, and the step (d) of discharging is continued until a recharging of the sensing cell is performed by means of the voltage source and a third and a fourth equiva-lence point are obtained. The further charge differences Aq3 between the first and the third equivalence point and Aq4 between the second and the third equivalence point and between the second-third and the fourth equivalence point measured during this process are used to determine the concentrations of the anolyte (c_A) and the catholyte (c_K).

[0045] According to another feature of the invention, a pulsed measurement is performed. For this purpose, steps (c) and (d) are performed in a pulsed manner by performing the discharge or recharge according to step (c) in a pulsed manner with defined discharge conditions (current and voltage) with a known pulse duration, wherein each discharge pulse under load is followed by a discharge pulse of known pause time without load (OCV).

[0046] While both a current-loaded voltage curve (cell voltage) and currentless voltage curve (open cell voltage/OCV voltage) over time can be used for evaluation to determine the first and second equivalence points, only the currentless voltage curve is preferably used to determine the third and fourth equivalence points due to excessive overvoltage during recharging of the sensing cell. Therefore, a voltage measurement is preferably recorded at a pulsed current. At the end of a current pulse, the overvoltage is pronounced, which results in the current-loaded curve as the envelope, while at the end of the current pulse period, nearly the open terminal voltage is measured, which results in the currentless measurement curve as the envelope, which can be used to determine the third and fourth equivalence points.

[0047] According to a further embodiment of the invention, Eqs.

or

wherein c_K concentration of the catholyte,

Cd concentration of the anolyte,

V_sK volume of the sensing cell of the catholyte,

V_S,A volume of the sensing cell of the anolyte,

F Faraday constant,

Aq₃ charge difference between 1st and 3rd equivalence points,

Aq₄ charge difference between 2nd and 4th equivalence points, and

Aq_A charge difference between 3rd and 4th equivalence point. are used for evaluation to determine the concentration of the anolyte and the catholyte c_K by comparing the results of several measurements.

[0048] In general, the initial concentrations of the initial intermixed anolyte and catholyte are known. However, during operation, variations in concentration may occur due to membrane effects. In this case, the above measurement method can be used to determine the concentrations of the catholyte and the anolyte. For this purpose, the sensing cell must be separated from the RFB unit in any case. If the sensing cell is not disconnected, shunt currents would prevent a defined reloading. Preferably, pulsed discharging is used here, as explained above. [0049] Further, a complete determination of all independent state variables of the electrolyte system of a flow battery of the electrolyte system can be performed. The independent state variables of the electrolyte system of a flow battery are the con- centrations, volumes, and oxidation states in the anolyte and catholyte: c_K, V_tK, z_K, c_A, VtA, and z_A, and can be uniquely determined in a stepwise procedure.

[0050] According to a further embodiment of the invention, a method for compensating for a shift in the average oxidation state of the entire electrolyte system in a redox flow battery, in particular a VRFB battery, is disclosed, wherein a shift in the average oxidation state of the entire electrolyte system is first determined and then, depending on the determined shift, a reducing agent is added on the catholyte side in the case of an oxidative shift or an oxidizing agent is added on the anolyte side in the case of a reductive shift.

[0051] Here, for example, ethanol, methanol, oxalic acid, formic acid, acetic acid, ascorbic acid or a sugar solution could be added on the catholyte side in the case of an oxidative shift.

[0052] In this way, a shift in the average oxidation state of the entire electrolyte system can first be automatically determined by software through a plurality of measurements, and then either a reducing agent on the catholy side or an oxidizing agent on the anoly side can be automatically added depending on the oxidative shift or reductive shift, thereby increasing the capacity of the working battery by compensating for the shift and eliminating a risk of damage.

[0053] Herein, the amount of reducing agent or oxidizing agent to be added can be calculated from the magnitude of the shift in the average oxidation state of the entire electrolyte system and from the concentration of the reducing agent or oxidizing agent.

[0054] If a reducing agent or oxidizing agent is added to the redox flow battery to compensate for a shift, the battery is preferably charged at least initially to an SOC of at least 70%, preferably at least 80%.

[0055] In this way, an effective capacity increase is ensured. [0056] According to a further embodiment of the invention, in the device according to the invention, the computer is designed to determine whether an oxidative or a reductive shift is present from a comparison of different measurements with differently sized half cells of a sensing cell and/or with different SOCs (states of charge).

[0057] According to a further embodiment of the invention, the apparatus further comprises a dosing device for adding a reducing agent on the catholy side or an oxidizing agent on the anoly side, which is coupled to the computer for automatically metering the reducing agent or the oxidizing agent depending on the calculated shift of the oxidation state.

[0058] In this way, when an RFB unit is in operation, an automatic check of a possible shift of the oxidation state can be performed at regular time intervals and, depending on the result, an automatic addition of a reducing agent or an oxidizing agent can be performed in order to restore the capacity of the battery to the best possible value and to avoid damage.

[0059] It is understood that the above features of the invention, and those to be explained below, can be used not only in the combination indicated in each case, but also in other combinations or on their own, without departing from the scope of the invention. Further features and advantages of the invention will be apparent from the following description of preferred embodiments with reference to the drawings. In the drawings show:

Fig. 1 a schematic diagram of a flow circuit of a redox flow battery;

Fig. 2 the electrical circuit diagram of a monitoring circuit for determining an electrolyte system state, in particular a shift in the oxidation state;

Fig. 3 the OCV voltage and cell voltage of a non-flow-through VFRB cell with a volume of the half-cell in a sensing cell for anolyte and catholyte of 45 ml each, a membrane area of 127 cm² and a charge/discharge current of 6.35 amperes; Fig. 4 Discharge curves of a VRFB sensing cell of different discharge currents with a volume of the half cells for anolyte and catholyte of 10 ml each;

Fig. 5 Experiments on the reproducibility with a small sensing cell (volume of the half cells for anolyte and catholyte of 1.5578 ml each) at a discharge current of 0.2 A and an offset voltage of 3 V;

Fig. 6 different discharge curves of electrolytes with different average oxidation values with a balanced electrolyte (solid curve with an oxidation value of 3.495, an oxidatively shifted electrolyte with an oxidation value of 3.537 and a reductively shifted electrolyte with an oxidation value of 3.459;

Fig. 7 a determination of the equivalence points by forming the first derivative of the discharge curves at a volume of the half cells for anolyte and catholyte of 1.5578 ml each, a discharge current of 300 mA and with an average oxidation value of 3.615;

Fig. 8 a schematic representation of a discharge curve of a VFRB sensing cell, wherein the discharge is continued until a recharge takes place and until two further potential jumps (equivalence points) occur, so that the measurement is only stopped after a total of four equivalence points;

Fig. 9 the voltage curve of a VFRB cell which is not fluxed and for which the discharge is continued until a recharge occurs and until two further potential jumps (equivalence points) occur, so that the measurement is only interrupted after a total of four equivalence points;

Fig. 10 the results for the two average oxidation states z_g_esi,2 of the entire electrolyte system of an industrial VRFB at different SOC starting points for the discharge curve measured on a displacement sensing cell with different half-cell volumes; Fig. 11 a schematic representation of a complete charge curve of a VFRB sensing cell starting from a completely mixed electrolyte.

[0060] A redox flow battery, exemplified by a VRFB battery, is shown in Fig. 1 and designated overall by the numeral 10. The VFRB battery comprises a working battery 12 with two half cells separated by a membrane 13. In series with the working battery 12 is a sensing cell 14 comprising two half-cells 16, 18 separated by a membrane 15.

[0061] Furthermore, an anolyte tank 20 and a catholyte tank 26 are provided. Further, in series with the sensing cell 14 on the anoly side is an optional valve 56 coupled to a pump 22. Thus, on the anoly side, anolyte can be pumped from the anolyte tank 20 via an associated conduit through the associated half-cell 16 of the sensing cell 14 and the optio-nal valve 56 through the associated half-cell of the working battery 12 and flows back into the anolyte tank 20 via a conduit. Similarly, on the catholyte side, catholyte may be pumped from the catholyte tank 26 by a pump 28 through the half-cell 18 via an optional valve 57 through the associated side of the working battery 12 and flows back into the catholyte tank 48 via a conduit. Optionally, bypass valves 54, 55 are further provided to bypass the half-cells 16, 18 of the metering cell 14.

[0062] It is understood that the embodiment and arrangement of a sensing cell shown in Fig. 1 only exemplifies one of several possibilities. In addition to the sequential arrangement and flow through of sensing cell and working cell shown here, a parallel arrangement and flow through is also conceivable, for example, as well as a return of the sensing cell electrolyte circuits directly to the suction side of the pumps in the case of a parallel arrangement of the sensing cell.

[0063] In Fig. 1 , a monitoring circuit 40 according to the invention is also indicated by dashed lines. The electrical circuit diagram of the monitoring circuit 40 is explained below with reference to Fig. 2.

[0064] In Fig. 1 , containers 59 for an oxidizing agent and 60 for a reducing agent with an associated metering pump 61 and 62 are furthermore provided by way of example. Thus, in order to compensate for a measured shift in the average oxidation state of the entire electrolyte system, as will be explained in detail later, an oxidizing agent can be metered in on the anoly side or a reducing agent can be metered in on the catholy side in order to achieve an improvement in capacity. The dosing units may also be different from a pump, especially when gaseous or solid reducing or oxidizing agents are added.

[0065] In the redox flow battery 10, electrical energy is stored in the chemical energy of the electrolyte system consisting of the anolyte and the catholyte. The electrolytes are located in separate tanks 20 and 26, respectively. The electrolytes contain ions of different oxidation states. When charging the battery 10, ions of the anolyte are reduced and ions of the catholyte are oxidized. During discharging, the opposite happens.

[0066] It should be noted that the invention is explained in more detail herein using the example of a VFRB battery with an aqueous sulfuric acid electrolyte, but that the invention is applicable to any type of flow battery.

[0067] In the VFRB battery 10 shown herein, during the charging process, V^v ions are reduced to V^lv ions on the cathode (negative terminal) side and V" ions are oxidized to V¹" ions on the anode (positive terminal) side.

[0068] The solution on the positive pole side contains vanadyl sulfate, VOSO₄, which can be oxidized to the pentavalent ion:

[0069] Positive electrode, positive pole V^lv and V^v:

VO²⁺ + H₂O O VO₂ ⁺ + 2H⁺+ e .

[0070] The solution on the negative pole side contains vanadium¹" sulfate, which can be reduced to the divalent vanadium salt:

[0071] Negative electrode, negative pole:

[0072] Due to undesirable side reactions during operation of a vanadium redox flow battery with sulfuric acid electrolyte, a gradual oxidation of the electrolyte occurs during normal operation. This eventually leads to overloading of the catholyte compared to the anolyte during the charging process. As a result, the battery capacity slowly decreases, and ultimately the ever-increasing overcharge leads to destruction of the stack, which may not be noticed at first. In the process, the cathodic bipolar plate can become severely corroded, ultimately leading to high costs and system downtime.

[0073] According to the invention, in order to determine the electrolyte system state of the redox flow system 10, in particular to determine the concentrations of anolyte and catholyte, as well as the shift in the oxidation state, starting from the charged state of the redox flow battery 10, a half cell 16, 18 of a sensing cell 14 (cf. Fig. 1) is filled with the anolyte and with the catholyte, respectively. For measurement, the pumps 22, 28 are switched off or the sensing cell is disconnected. A voltage source 44 is now connected in series with the sensing cell 14 as shown in Fig. 2 and a circuit 42 of the sensing cell 14 is closed via a resistor 38. Discharging of the sensing cell 14 now takes place, while current and voltage of the sensing cell 14 are monitored over time by means of a current meter 46 and by means of a voltage meter 48.

[0074] In this case, the voltage source 44 has a voltage that can supply a sufficiently large current for a defined discharge of the sensing cell over the entire voltage range occurring during the measurement and whose voltage range is usually larger than the OCV voltage of the sensing cell. This is typically between about 1.2 V and 1.5 V, which corresponds to an SOC of 15 to 85 %. For example, a voltage of 2 V or 3 V is used as the offset voltage. Advantageously, the voltage source 44 is designed as a constant current source. In this case, the current meter 46 can be dispensed with. Furthermore, an electronic evaluation unit 50 is coupled to the sensing cell 14 and to the voltage source 44 and possibly to a current meter 46. Furthermore, a display 52 may be provided which is controlled by the evaluation electrode 50. [0075] For measuring an electrolyte system state according to the invention, in particular a shift in the oxidation state, an unknown state of charge (SOC) is initially assumed.

[0076] Fig. 3 shows the voltage measurements on a non-fluxed sensing cell with current (cell voltage during charging and discharging) and without current (OCV voltage). The following parameters were used: for the charge/discharge current 6.35 A; volume of the half cells of the sensing cell 16, 18 of 45 ml each, and a membrane area of 127 cm².

[0077] There are discontinuities in the voltage curve over time during discharging.

[0078] Fig. 4 shows various discharge curves at different discharge currents. The volume of the sensing cell (anolyte volume = catholyte volume) was 10 ml, the offset voltage was 3 V, and the electrolyte exhibited a positive, i.e., oxidative, shift with an oxidation value of 3.58. The state of charge (SOC) at the beginning of the measurement was about 50%.

[0079] When all V^v+ ions are consumed in the catholyte and only V^lv+ ions are present, or when all V^ll+ ions are consumed in the anolyte and only V^lll+ ions are present, a potential jump occurs. These points are visible in the discharge curve according to Fig. 4 as inflection points and are called equivalence points. As a rule, they do not overlap in a real system with a slightly unbalanced electrolyte. In the ideal case, i.e. if there is a balanced, optimally adjusted electrolyte (with respect to oxidation state and concentration) and the half-cells of the sensing cell are of equal size, then the equivalence points that form for anolyte and catholyte lie exactly one above the other, and only a single potential jump is visible.

[0080] If the ratio of V^v+ and V^ll+ ions in the catholyte and in the anolyte is unbalanced at the start time of the measurement, as is usually the case with a shifted electrolyte, then the equivalence points are shifted by a corresponding time in the diagram. Graphically, this time difference shows up as a shoulder in the voltage curve. The time difference can be determined very precisely via the first derivative of the discharge curve (cf. Fig. 7). Fig. 6 shows the discharge curves of electrolytes with different average oxidation values. For the balanced electrolyte with an oxidation value of 3.495 (approximately 3.5), there is only one equivalence point if the half-cell volumes of the sensing cell are equal. However, if the electrolyte is oxidatively or reductively shifted, divergent curves with two equivalence points result when the half-cell volumes of the sensing cell are the same, as shown in Fig. 6 by the dashed line and the double-dotted line. For Figs. 6 and 7, smaller half-cell volumes of 1.557 ml were used for anolyte and catholyte, respectively.

[0081] For an accurate measurement of the time difference, the first derivative of the discharge curve is determined according to Fig. 7. Additionally, the second derivative can be considered.

[0082] Fig. 5 shows that the reproducibility of different measurements is very good. For this purpose, a total of four measurements with identical parameters were repeated, in each case filling the two half cells 16, 18 of the sensing cell 14 from the same anolyte tank 20 and from the same catholyte tank 26.

[0083] The following procedure is used to evaluate the measurements:

[0084] To calculate the displacement of an electrolyte system, the charge balance that can be balanced between catholyte and anolyte is assumed. The charge Q, the volume V and the amount of substance n are quantity-like quantities which can be balanced. In addition, the charge is always coupled to a charge carrier (Q = n - z - F), and the amount of substance can be given as the product of concentration and volume (n = c - V).

From this, a general displacement formula (1) can be derived:

and

wherein:

^■t,ges, i and Zt,ges,2, are the two possible average oxidation states of the entire electrolyte system of the RFB unit,

ZK,SOC O% and z_A,soc o% are the theoretical oxidation states of the catholyte and anolyte ions, respectively, in a discharged battery (state of charge (SOC): 0%),

V is the volume,

K refers to the catholyte,

A refers to the anolyte, c is the concentration of redox ions, t refers to the amount of electrolyte in the tanks and electrolyte circuits of the RFB unit, s refers to the sensing cell, q1, q2 are the charge differences from the start of discharge to the first equivalence point and to the second equivalence point, respectively, and

F is the Faraday constant.

[0085] Equation (1) contains all influencing variables responsible for a shift of the electrolyte system of a flow battery under operating conditions (Ct_K, c^, V_tK, VtA, z_tK (determined by Aq1) and ZtA (determined by Aq2)), as well as all parameters characterizing the sensing cell (C_SK, C_SA, V_SK, V_SA).

[0086] Starting from equation (1), depending on which of the parameters are already known, the remaining parameters can be determined by different measurements with variation of the other parameters, e.g., with different SOCs, different volumes, and different measurement times.

[0087] In this context, the invention relates to the totality of measurement applications that have the above formula (1) as a basis. In general, the more quantities of the general formula (1) are known, the fewer measurements under conditions defined differently from each other have to be carried out. Or the other way around: The less quantities of the general formula (1) are known, the more measurements under conditions defined differently from each other have to be performed to determine the displacement.

[0088] For the special case of a VRFB battery where ZK,SOC O% = 4 and z_A,soco%,= 3 and where the ion concentration in the tanks (20, 26) of the working battery (12) are the same for catholytes and anolytes and in the half cells (16, 18) of the sensing cell (14) are the same for catholytes and anolytes (C_SK = C_SA = Ct_K = c^ = c), the equations (1. 1) and (1.2) to equations (2.1) and (2.2):

(2.2)

[0089] For the special case of being a VRFB battery, where the volumes of the half-cells (16, 18) of the sensing cell are the same for the anolyte and the catholyte (V_sK = V_sA = VHZ), where the volumes of the tanks (20, 26) of the working battery (12) are the same for the anolyte and the catholyte (V_M = V«), and wherein the ion concentrations in the tanks (20, 26) of the working battery (12) are the same for catholyte and anolyte and in the half cells (16, 18) of the sensing cell (14) are the same for catholyte and anolyte (CSK = CSA = Ct_K = c^ = c), equation (1) simplifies to the following equation (3):

and

[0090] Herein, z_t,_ges,i and z_t,_ges,2, hereinafter also z_ges,i and z_ges,2, are the two possible oxidation states of the entire electrolyte system, where Aq1 and Aq2 are the charge differences from the onset of discharge to the first equivalence point and from the onset of discharge to the second equivalence point, respectively, and F is the Faraday constant.

[0091] This is explained in more detail below with examples.

Example 1

[0092] For a VFRB battery, a measurement is made with the following parameters: F = 96485.45 C/mol (Faraday constant in Coulomb/mol) c = 1.6 l/mol (with c_sk = c_sA = c_tK = c_tA = c) (concentration) V_tK = 431.3 I und VtA (catholyte and anolyte volumes, respectively) V_SK = V_SA = 1.5572 ml (volumes of the sensing cell)

Aq1 = 0.02744 Ah and Aq2 = 0.04200 Ah (measured values for charge differences). From this, the two possible oxidation states are calculated according to equation (2):

ZKO,I = 4.411 and z_Ao,i = 2.371 , resulting in z_ges,i = 3.388 (from 2.1); Z_KO,2 = 4.629 and z_Ao,2= 2.589, resulting in z_ges,2 = 3.606 (from 2.2). In both cases, SOC of the electrolyte system is about 50%. [0093] Whether now the first case (z_ges,i = 3.388) is present or the second case (z_ges,2 = 3.606) cannot be determined with this measurement alone. At least one further measurement is required for this. First, in examples 2 and 3, measurements with other half-cell volumes of the sensing cell are shown.

Example 2

[0094] F = 96485.45 C/mol (Faraday constant in Coulomb/mol) c = 1.6 l/Mol (with c_sk = c_sA = c_tK = c_tA = c) (concentration) V_tK = 431.3 I and VtA = 433.6 I catholyte and anolyte volumes, respectively) V_SK = V_SA = 0.7786 ml volumes of the sensing cell)

Aq1 = 0.01221 Ah and Aq2 = 0.01994 Ah (measured values for charge differences) From this, the two possible oxidation states are calculated according to equation (2): ZKO,I = 4.366 and z_Ao,i = 2.403, resulting in z_ges,i = 3.382 (from 2.1);

ZKO,2 = 4.597 and z_Ao,2 = 2.634, resulting in z_ges,2 = 3.613 (from 2.2). In both cases, SOC of the electrolyte system is about 50%.

[0095] In Example 2, measurements were made using a sensing cell with half volumes in each case. Within the limits of the measurement accuracy, the results are the same as in example 1 . However, it is also not possible to distinguish here whether the first case (z_ges,i = 3.382) or the second case (z_ges,2 = 3.613) is present.

Example 3

[0096] F = 96485.45 C/Mol (Faraday-Konstante in Coulomb/mol) c = 1.6 l/Mol (mit c_sk = c_sA = c_tK = c_tA = c) (concentration) V_tK = 431 .3 I und V_tA = 433.6 I (catholyte and anolyte volumes) SK = 0.7786 ml und V_sA = 1.5572 ml (volumes of the sensing cell) Aq1 = 0.01717 Ah und Aq2 = 0.02458 Ah (measured values for charge differences) From this, the two possible oxidation states are calculated according to equation (2): ZKO,I = 4.514 and z_A0,i = 2.632, resulting in z_ges,i = 3.571 (from 2.1);

ZKO,2 = 4.736 and z_A0,2 = 2.743, resulting in z_ges,2 = 3.737 (from 2.2). In both cases, SOC of the electrolyte system is about 50%.

[0097] In Example 3, measurements were made using a sensing cell with two different halfvolumes. At first glance, the results z_ges,i = 3.571 and z_ges,2 = 3.737 do not appear to be congruent with those in Examples 1 and 2.

[0098] However, within the limits of measurement accuracy, z_ges,2 = 3.737 of Example 3 corresponds to z_ges,2 = 3.613 of Example 2 and z_ges,2 = 3.606 of Example 1. Thus, there is a positive shift in the average oxidation state of the entire electrolyte system. However, this determination requires at least two sensing cells with different half-cell volumes. An improved methodology using only one sensing cell with different half-cell volumes is shown in Example 4-7.

Measurement accuracy

[0099] Since the electrolyte is measured in the same state (from the large tank volume) in each of Examples 1 to 3, the same oxidation value z should actually result in all Examples 1 to 3. Within the limits of measurement accuracy, this is also approximately the case when the results z_ges,2 = 3.737 of Example 3, z_ges,2 = 3.613 of Example 2 and z_ges,2 = 3.606 of Example 1 are compared with the (exact) value of 3.58 determined by cerimetry.

[00100] It has been shown that the electrolyte in the inlet and outlet tubes systematically influences the measurement result because of diffusion and shunt currents. This influence can be calculated out. Furthermore, measurements with interrupted contact to the electrolyte-carrying hoses are conceivable, for example through shut-off or check valves in the inlet and outlet hoses.

Determination of the correct displacement value

[00101] For this purpose, the following procedure is used, which exploits the different half-cell volumes of the sensing cell of example 3. The method will first be described using Exam- pies 4 through 7, wherein an electrolyte of the same oxidative shift is measured at different initial states of charge SOC:

Example 4

[00102] SOC = 9.24%

F = 96485.45 C/mol c = 1.6 ± 0.02 mol/l

V_sK = 0.7786 ± 0.0100 ml

V_sA = 1.5572 ± 0.0100 ml

V_tK = 451.3 ± 2.0 I

VtA = 453.6 ± 2.0 I

I = 300 ± 2 mA t1 = 30.2 ± 0.5 s t2 = 89.7 ± 0.5 s

Aq1 = 7.13 ± 0.16 As

Aq2 = 24.98 ±0.23 As

Zges.i ⁼ 3.4762 ± 0.0056

Zges,2 = 3.5875 ± 0.0062 (correct)

Example 5

[00103] SOC 25.45%

F = 96485.45 C/mol c = 1.6 ± 0.02 mol/l

V_sK = 0.7786 ± 0.0100 ml

V_sA = 1.5572 ± 0.0100 ml

V_tK = 451.3 ± 2.0 I

VtA = 453.6 ± 2.0 I

I = 300 ± 2 mA t1 = 133.6 ± 0.5 s t2= 137.7 ± 0.5 s

Aq1 = 37.65 ± 0.31 As

Aq2 = 38.85 ± 0.31 As

Zges.i = 3.5739 ± 0.0077 (correct)

Zges,2 = 3.5814 ± 0.0077 (correct)

Example 6

[00104] F = 96485.45 C/mol c= 1.6 ±0.02 mol/l

V_sK = 0.7786 ±0.0100 ml

V_sA = 1.5572 ±0.0100 ml

V_tK = 451.3 ± 2.0 I

VtA = 453.6 ± 2.0 I

I = 300 ± 2 mA t1 = 138.0 ± 0.5 s t2= 138.1 ± 0.5 s

Aq1 =40.50 ±0.31 As

Aq2 = 40.50 ± 0.31 As

Zges.i = 3.5823 ± 0.0079 (correct)

Zges,2 = 3.5823 ± 0.0079 (correct)

Example 7

[00105] F = 96485.45 C/mol c= 1.6 ±0.02 mol/l

V_sK = 0.7786 ±0.0100 ml

V_sA = 1.5572 ±0.0100 ml

V_tK = 451.3 ± 2.0 I

VtA = 453.6 ± 2.0 I

I = 300 ± 2 mA t1 = 133.6 ± 0.5 s t2 = 137.7 ± 0.5 s

Aq1 = 61.95 ± 0.45 As

Aq2 = 88.95 ± 0.62 As

Zges.i ⁼ 3.5703 ± 0.0105 (correct)

Zges,2 ⁼ 3.7386 ± 0.0119

Evaluation of examples 4 to 7 to determine the correct displacement value. a) SOC 9.24% (Example 4): As Example 4 shows, two results are obtained from Equation (2), and it is not yet possible to say which is the correct one. b) SOC 25.45% (example 5): Here, both equivalence points are close to each other. This leads to the fact that both results are almost the same within the measurement accuracy. c) SOC 26.81% (Example 6): Here the equivalence points overlap. Thus, both equations of formula (2) give the same result within the measurement accuracy. If such a case is measured at random, the average oxidation level of the entire electrolyte system has been determined unambiguously. d) SOC 49.19% (Example 7): Here, both equivalence points are again clearly distinguishable from each other. There are again two results for the mean oxidation state, of which one cannot tell from the measurement alone which result is the correct one.

[00106] As examples 4 to 7 show, without random overlap of the equivalence points, the mean oxidation state of the electrolyte cannot be unambiguously determined from one measurement. However, it is uniquely determined by two measurements, namely by a combination of Example 4 and Example 7. [00107] Comparing the results of Example 4 and Example 7, it can be seen that z,_ges,2 of Example 4 deviates the least from z,_ges,i of Example 7, and is even the same within the limits of measurement inaccuracy.

[00108] Since this is not so for all other combinations z,_ges,i Example 4/ z,_ges,i Example 7, z,_ges,2 Example 4/ _z,_ges,2 Example 7 and z,_ges,i Example 4/ z,_ges,2 Example 7, it must be the case that z_ges,2 Example 4/ z,_ges,i Example 7 is the correct result. In this case, the average oxidation state of the electrolyte system can be determined by two measurements at different SOCs if there is equality within the measurement accuracy between two results of different indices. In case of doubt, measurements must be made at a further SOC.

[00109] The applicable result in example 4 occurs at index 2, and in example 7 at index 1. This is due to the fact that in a sensing cell with different half-cell vo-lumes, the potential jumps "overtake" each other as a function of the initial state of charge. Physically, this can be explained by an example: An electrolyte is charged (SOC 50%), in the sensing cell and slightly positively shifted. The half cell with the smaller volume has fewer charge carriers available to react. In this example, the catholyte volume is smaller. When the charge carriers in the smaller volume are used up, the first potential jump occurs in this cell. Thus, equation (1.1) is valid and the correct result is the one with index 1. This is true if the catholyte undergoes the potential jump first (Example 7).

[00110] If the initial state of charge is now smaller, the positive shift of the electrolyte relative to the state of charge becomes more important. A positive shift means that too many charge carriers are available on the catholyte side. These compensate for the smaller volume from a certain reduced state of charge. If the compensation of volumetric and oxidative shift is complete, both half cells pass through the potential jump at the same time during discharge, and the same result is obtained from both equations (1.1) and (1.2) of equation (1) or (2.1) and (2.2) of equation (2). If the initial state of charge is reduced further, then the smaller volume of the catholyte is overcompensated by the positive shift. Thus, during the discharge process, the charge carriers available for reaction in the anolyte are consumed first, so that the first potential jump can be assigned to the anolyte (Example 4). This means that the second equation of equation (1) or (2) holds and the result with index 2 is the correct one. [00111] Thus, in the methodology described here, for an unambiguous determination of the shift, two measurements with different initial charge states are needed, between which an overtaking of the potential jumps has taken place. This can be seen in the result for the mean oxidation state by the fact that the result of the two measurements, which is the same within the limits of measurement accuracy, has changed index.

[00112] In practice, this procedure can be easily applied by first performing these two measurements at different SOC for an indeterminate system. If one knows by the described procedure whether the electrolyte system is positively or negatively shifted, then subsequent measurements at regular intervals are usually sufficient, since the shift usually changes only slowly and steadily in one direction.

[00113] Further, it is preferred to measure only at SOCs less than or equal to 50% when the sensing cell is not disconnected. The higher the state of charge, the longer the measurement and the influence of the shunt currents as well as the membrane effects.

[00114] Fig. 10 shows the results for the two mean oxidation states z_t,_ges,i and z_t,g_es,2 of the entire electrolyte system of an industrial VRFB at different SOC starting points for voltage curves measured on a sensing cell with different half-cell volumes. While one value of the two average oxidation levels of the whole electrolyte system changes strongly with the SOC starting point, the other remains the same within the measurement error and a superimposed systematic error due to a non-separated sensing cell. The value of 3.45, which is more constant over the SOC, is the real average oxidation state of the entire electrolyte system. The electrolyte system state of the VRFB system is less than 3.5 and thus negatively shifted. One can see in Fig. 10 a linear dependence of z_t,_ges,i and z_t,g_es,2 on the SOC. An extrapolated value of the trend line with the smaller slope at low SOC is most accurate. Thus, it is usually sufficient to perform two measurements using a sensing cell with different half-cell volumes at different SOC to determine the linear trend lines of the SOC dependence of z_t,_ges,i and z_t,_ges,2 and the shift in the mean oxidation state of the entire electrolyte system.

Determination of the concentrations of anolyte and catholyte. [00115] In general, the initial concentration of the initial intermixed electrolyte is known. During operation, variations in concentration may occur due to membrane effects.

[00116] This difference in concentration between anolyte and catholyte can be determined by recharging the sensing cell. Reloading of the sensing cell is only possible when the sensing cell is separated. When the sensing cell is not disconnected, however, the shunt currents prevent a defined recharging. Recharging is achieved by not stopping the measurement after two equivalence points, but by continuing to discharge. Then, with a vanadium electrolyte, the catholyte becomes an anolyte and the anolyte becomes a catholyte. The discharging process is therefore simultaneously the charging process during recharging. During this process, two equivalence points will generally occur again, so that overall the measurement is only stopped after four equivalence points. This is shown as an example in Fig. 8.

[00117] In practice, however, the third and fourth equivalence points are generally not visible in the current-loaded case when measuring the cell voltage, but only in the current-unloaded case when measuring the OCV voltage. To measure the OCV voltage during recharging, a measurement wherein the current is always applied for only a short period of time, i.e. pulsed, is expedient. While discharging takes place during the current pulse, the OCV voltage is adjusted during the pauses between the current pulses. Fig. 9 shows such a pulsed measurement. Shown is the voltage curve of a non-fluxed VFRB cell with a volume of the half-cell for anolyte and catholyte of 45 ml each, a membrane area of 127 cm2 at a pulsed discharge current of 6.35 amperes with a pulse duration of 10 s and a pause time of 10 s, during which the discharge continues until a recharge occurs and until two further potential jumps (equivalence points) occur, so that the measurement is only stopped after a total of four equivalence points. It can be seen that the third and fourth equiva-lence points are visible for the enveloping currentless OCV voltage curve, but not for the enveloping current-loaded cell voltage curve.

[00118] The special feature of these additional equivalence points or potential jumps is that they occur at a distance of z=1 from the previous ones of the respective half cell. This is because the catholyte with oxidation state 4 at the first potential jump of the catholyte halfcell becomes oxidation state 3 at the second potential jump of the catholyte half-cell at the beginning of the recharge. In the anolyte, the change from 3 to 4 takes place, i.e., exactly the reverse.

[00119] Thus, two additional charge differences, namely Aq3 and Aq4 can be determined, which can be assigned differently. With z=1 holds:

wherein c_K concentration of the catholyte, c_A concentration of the anolyte,

V_sK volume of the sensing cell of the catholyte,

V_sA volume of the sensing cell of the anolyte,

F Faraday constant,

Aq₃ charge difference between 1^st and 3^rd equivalence points,

Aq₄ charge difference between 2^nd and 4^th equivalence points, and

Aq_A charge difference between 3^rd and 4^th equivalence point.

[00120] The designation of the charge differences corresponds to that shown in Fig. 8. There are four different cases for the result of CK and c_A. If it is known whether the electrolyte is positively or negatively shifted (see the procedure described above for a system state after mixing), then only two of the four cases can apply. These can be further narrowed down by several measurements similar to the procedure described for distinguishing between positive and negative displacement. It is helpful to know the initial concentration of the initial mixed electrolyte and to start from the most probable cases, i.e. rather small concentration differences. [00121] Fig. 9 shows the voltage curve of a non-flow-through VFRB cell with a volume of the halfcell for anolyte and catholyte of 45 ml each, a membrane area of 127 cm2 at a pulsed discharge current of 6.35 amperes with a pulse duration of 10 s and a pause time of 10 s, during which the discharge is continued until a recharge occurs and until two further potential jumps (equivalence points) occur, so that the measurement is stopped altogether only after four equivalence points. It can be seen that the third and fourth equivalence points are visible for the enveloping currentless OCV voltage curve, but not for the enveloping current-loaded cell voltage curve.

[00122] Fig. 10 shows the results for the two average oxidation states z_gesi,2 of the entire electrolyte system of an industrial VRFB at different SOC starting points for the discharge curve measured on a displacement measurement cell with different half-cell volumes. It can be seen that one value of the two average oxidation states of the whole electrolyte system changes strongly with the SOC starting point, while the other remains the same within the measurement error and a superimposed systematic error. The value of 3.45, which is more constant over the SOC, is the real mean oxidation state of the entire electrolyte system. The electrolyte system state of the VRFB system is less than 3.5 and thus negatively shifted.

[00123] Fig. 11 shows a schematic representation of a complete charge curve of a VFRB sensing cell starting from a completely mixed electrolyte, where in general two potential jumps (equivalence points) occur until a state of charge (SOC) of 0% is reached.

Method for complete determination of all independent state variables of the electrolyte system

[00124] The independent state variables of the electrolyte system of a flow battery are the concentrations, volumes, and oxidation states in the anolyte and catholyte: c_K, V_tK, z_K, c_A, VtA, and z_A and are to be determined in the step-by-step procedure described herein.

[00125] Substeps of the procedure described herein are based on the previously described procedures for determining the average oxidation state of the entire electrolyte system and for determining the concentrations of anolyte and catholyte. The procedure differs slightly. It can be assumed that similar procedures will also work. Thus, this is an exemplary presentation of a procedure for the complete determination of all independent state variables of an electrolyte system.

[00126] The procedure includes several steps:

1^st step:

[00127] A final charge curve including recharge is recorded by a displacement sensor according to the invention by means of a preferably pulsed measurement at least during recharge (Fig. 9).

[00128] Via q3, q4 and q (Fig. 8) as well as the half-cell volumes of the sensing cell (= sensor cell), it is possible to determine the concentration of the sensor via Eqs. 4.1 and 4.2 to determine concentration values for the anolyte and catholyte. There are initially four different possibilities for the assignment.

2^nd step:

[00129] As already implicitly described in the determination of the average oxidation state for the entire electrolyte system, the oxidation states of catholyte and anolyte zko and z_Ko in the sensor cell can be determined via q1 and q2 for known concentrations and half-cell volumes of the sensor cell:

(5.2)

[00130] There are two possibilities here, a negative and positive shift. Together with the four different possibilities of assignment for the concentrations (step 1), there are theoretically 8 possibilities from step 1 and step 2. However, due to the fact that q1 can only occur with q3 and q2 only with q4, which means nothing else, that a catholyte oxidation state belongs to a catholyte concentration and an anolyte oxidation state to an anolyte concentration, there are only four possibilities of combining the calculated concentrations and oxidation states. It is not yet possible to say which one is the one present in the system.

3^rd step:

[00131] Here, a final charge curve including recharge (Fig.9) is additionally recorded by a displacement sensor according to the invention, at a second SOC. If the half cells have different volumes and the procedure described above is carried out to distinguish between positive and negative displacement (Fig.10), then the four possible cases of zko, c_K, z_Ao and c_A from steps 1 and 2 can be narrowed down to two.

4^th step:

[00132] While previously the flow cell was used disconnected without flow, it is now used as the OCV cell of the flow battery. Certainly, another sensing cell can be used as an OCV cell. Here, the state 0 of the electrolyte is the one at which measurements have been made so far. It is marked by an OCV value and can also be restarted after a change in the state of charge of a flow battery. In the VRFB electrolyte system, when the battery is discharged below the state of charge of 0%, the OCV voltage will reduce. In a shifted system, a shoulder exists between the anolyte and catholyte potential jumps. This is difficult to measure in practice when discharging, because when the stack voltage collapses (=current loaded cell voltage*number of cells), the inverter enters a voltage range at which it no longer functions. Thus, to measure the curve shown in Fig. 11 , in practice the RFB system must first be mixed at least to the point where the OCV voltage collapses and then charged back up to the 0 state. In general, this does not require complete mixing, but can be done.

[00133] Via the determination of q5 and q6 (Fig.11), wherein q5 >q6, there are two possible solutions for determining the volumes of anolyte and catholyte in the RFB system:

[00134] Thus, together with two solutions for ZKO, C_K, Z_AO, and c_A from steps 1-3, there are four possible solutions for the complete state of the electrolyte system ZKO, C_K, V_tK, z_A0, c_A, and V_tA after this step.

Using

the average oxidation state of the entire electrolyte system can be determined.

5^th step:

[00135] From the four solutions for the complete state of the electrolyte system zko, c_K,,V_tK, z_A0, c_A, and V_tA, the one that correctly characterizes the system can be found. For this purpose, the fact is used that, as a rule, the initial concentration of the intermixed electrolyte and its total volume are fixed, since this is part of the specification of the flux bath. By means of the volume balance

and the mass balance

[00136] halve the number of possibilities in each case, so that only one solution remains at the end of the procedure. Thus, the state of the electrolyte system is determined unambiguously and completely via the multi-stage procedure described.

[00137] This method for complete determination of the electrolyte system state can be used if, for example, no information is available on the current electrolyte volumes, which are often monitored by level sensors, or concentrations. In general, of course, the results of other measurements or determination methods for the concentrations and volumes can be taken into account and simplify the procedure.

Correction of shifts of the electrolyte system state

[00138] In the operation of a vanadium redox flow battery with sulfuric acid electrolyte, positive (oxidative) shifts generally occur, which can be automatically measured and evaluated by the method described above. For other electrolyte systems, there is also usually a trend toward increasingly positive or negative shifts.

[00139] Depending on the relevant result, which can be indicated visually by the monitoring circuit 40 according to Fig. 2, for example via a display 52 or a display, an automatic correction of a shift in the average oxidation state of the entire electrolyte system can be carried out by metering a reducing agent or an oxidizing agent from a reservoir 59 or 60 according to Fig. 1 via a corresponding metering unit, for example a metering pump 61 or 62, in an amount calculated from the shift and the concentration of the added agent.

[00140] Herein, a reducing agent is added on the catholy side in the case of an oxidative (positive) shift or an oxidizing agent is added on the anoly side in the case of a reductive (negative) shift.

Claims

38

Claims A method of determining an electrolyte system state, in particular a shift of an oxidation state of the electrolyte in a redox flow battery (10), in particular a VRFB battery, comprising the following steps:

(a) Filling in each case a half cell (16, 18) of a sensing cell (14) with an A- nolyte and a catholyte;

(b) connecting a voltage source (44) in series with the sensing cell (14);

(c) closing a circuit (42) across the sensing cell (14);

(d) discharging the sensing cell (14) while monitoring the current and voltage of the sensing cell (14) as a function of time; and

(e) determining the electrolyte system state, in particular a shift in oxidation state, from the current and voltage values during discharge, from the concentrations of anolyte and catholyte, from the volumes of the half cells (16, 18) and from the volumes of anolyte and catholyte. The method of claim 1, wherein the discharge of the sensing cell (14) is carried out up to a termination criterion which ensures that the occurrence of a first and - if present - second inflection point of the voltage over time (equivalence point) is part of the measurement, the electrolyte system state being calculated from the charge transferred between the equivalence points, the concentrations of anolyte and catholyte, the volumes of the half cells and from the volumes of anolyte and catholyte; wherein, in the case where no second equivalence point should occur for equal half-cell volumes, it is determined that there is no shift in oxidation state; wherein, in the case where no second equivalence point should occur for unequal half-cell volumes, it is determined that a shift in oxidation state has been unambiguously determined by a measurement. The method of claim 1 or 2, wherein the termination criterion is selected on the basis of a crossing of a minimum voltage difference during discharge, in such a way that it is ensured that not only the first equivalence point, but - if present - also a second equivalence point is detected. The method of any one of the preceding claims, wherein the discharge of the sensing cell (14) is terminated when no further voltage change occurs over a period of time which is greater than a discharge shoulder, i.e. no further voltage change occurs over time for a certain period of time, and/or a minimum discharge voltage difference has been passed through or when a period of time has elapsed wherein the discharge shoulder(s) occur(s) according to experience. The method of any one the preceding claims, wherein the discharge of the sensing cell (14) is terminated when a period of time has elapsed which is greater than discharge shoulders to be detected and their time intervals until they occur are long. The method of any one the preceding claims, wherein in step (b) a constant current source is used as voltage source (44), which is capable of supplying a cir- rent sufficiently large for a defined discharge of the sensing cell over the entire voltage range occurring during the measurement (14) and whose voltage is preferably greater than the OCV voltage of the sensing cell (14). The method of any one of claims 1 to 6, wherein the oxidation state shift is determined according to the equations:

wherein:

Zt,ges,i and z_t,ges,2 are the two possible average oxidation states of the entire electrolyte system of the RFB unit, z_{K S}oc 0% and z soc o%, are the theoretical oxidation states of the catholyte or anolyt ions in a discharged battery (state of charge (SOC)= 0%),

V is the volume,

K refers to the catholyte,

A refers to the anolyte, c is the concentration of redox ions, t refers to the amount of electrolyte in the tanks and electrolyte circuits of the

RFB unit, s refers to the sensing cell, q1 und Aq2 are the charge differences from the start of discharge to the first equivalence point and from the start of discharge to the second equivalence point, respectively, and

F is the Faraday constant. The method according to any one of claims 1 to 7, wherein the oxidation state shift in the case of a VRFB battery is determined according to the equations:

and

wherein z_t,_ges,i and z_t,g_es,2 are the two possible oxidation states and provided that the volumes of the half-cells (16, 18) of the sensing cell are equal for the anolyte and the catholyte (V_sK =

= V_Hz), that the two possible oxidation states and provided that the volumes of the half-cells (16, 18) of the sensing cell are equal for the anolyte and the catholyte (V_M = V«), and that the ion concentration in the tanks (20, 26) of the working battery (12) are the same for catholyte and anolyte and in the half cells (16, 18) of the sensing cell (14) are the same for catholyte and anolyte (Csk = C_SA = Ct_K = c^ = c), wherein Aq1 and Aq2 are the charge differences from the start of discharge to the first equivalence point and from the start of discharge to the second equivalence point, and F is the Faraday constant. The method according to any one of the preceding claims, wherein the method steps (a) to (e) are repeated at least with changed parameters of the sensing cell, and/or with a changed SOC, and further parameters are determined from a comparison of the different measurements. The method according to any one of claims 1 to 9, wherein the sensing cell (14) is disconnected from the working battery (12), and the step (d) of discharging is continued until the sensing cell (14) is recharged by means of the voltage source (44) and - if present - a third and a fourth equivalence point are obtained, and the further charge differences measured during this process are used to determine further parameters in conjunction with further measurements. The method according to any one of the preceding claims, wherein steps (c) and (d) are carried out in a pulsed manner, in that the discharging or recharging according to step (c) is carried out in a pulsed manner with defined discharge conditions (current and voltage) with a known pulse duration, each discharge pulse under load being followed by a discharge pulse of known pause time without load (OCV). The method according to any one of claims 9, 10 or 11 , wherein the method steps (a) to (e) are repeated with a changed volume of the half cells (16, 18) and/or with a changed SOC (state of charge) and it is determined from a comparison with the value calculated in at least one other measurement for the shift of the average oxidation state of the entire electrolyte system whether it is an oxidative or a reductive shift.

The method according to any one of claims 11 or 12, wherein the charge differences (Aq3) and (Aq4) associated with the third and fourth equivalence points from the start of discharge to the third equivalence point and from the start of discharge to the fourth equivalence point are used to determine the concentrations of the anolyte (c_A) and the catholyte (CK).

The method of claim 13, wherein, in the case of a VFRB battery, the equations

or

wherein c_K concentration of the catholyte, c_A concentration of the anolyte,

V_sK volume of the sensing cell of the catholyte,,

V_sA volume of the sensing cell of the anolyte,

F Faraday constant,

Aq₃ charge difference between 1st and 3rd equivalence points

Aq₄ charge difference between 2nd and 4th equivalence points, and

Aq_A charge difference between 3rd and 4th equivalence point are used for evaluation to determine the concentrations (CA, C_K) of the anolyte and the catholyte by comparing the results of multiple measurements. 43 The method according to any one of claims 9 to 14, wherein a complete determination of all independent state variables, i.e. the concentrations, volumes and oxidation states in the anolyte and catholyte (z_Ko, c_K,,V_tK, z_Ao, c_A, V_tA) of the electrolyte system is performed, wherein the state variables (z_Ko, c_K,,V_tK, z_A0, c_A, V_tA) are determined stepwise by repeating the measurements with changed parameters or changed SOC and comparing with the results of the previous measurements. A method for compensating for a shift in an average oxidation state of the entire electrolyte system in a redox flow battery (10), in particular a VRFB battery, wherein a shift in the average oxidation state of the entire electrolyte is first determined according to one of the preceding claims, and then, as a function of the determined shift, a reducing agent is added on the catholyte side in the case of an oxidative (positive) shift or an oxidizing agent is added on the anolyte side in the case of a reductive (negative) shift. The method according to claim 16, wherein ethanol, methanol or ascorbic acid is added in the case of an oxidative shift on the catholyte side. The method according to claim 16 or 17, wherein the amount of reducing agent or oxidizing agent to be added is calculated from the magnitude of the shift of the oxidation state and from the concentration of the reducing agent or oxidizing agent. The method according to any one of claims 16 to 18, wherein prior to the addition of the reducing agent or oxidizing agent, the redox flow battery (10) is charged to an SOC of at least 70%, preferably at least 80%. A device for determining an electrolyte system state, in particular a shift in an average oxidation state of the entire electrolyte system, in a redox flow battery (working battery) (10), in particular a VRFB battery, comprising: 44 a sensing cell (14) with two half-cells (16, 18) separated by a membrane (15), one of which can be filled with the anolyte and one with the ca-tholyte from the working battery (12); a voltage source (44) which can be connected in series with the sensing cell (14) and can be connected to a closed circuit (42) for discharging and subsequently recharging the sensing cell (14); a detector (40) for detecting a voltage change and for detecting the charge of the sensing cell (14) during discharging; and a computer (50) for calculating an electrolyte system state, in particular a shift in the oxidation state, from the current and voltage values during discharge, from the concentrations of catholytes and anolytes in the half-cells of the sensing cell, from the volumes of the half-cells and from the volumes of anolyte and catholyte, in particular according to a method according to one of claims 1 to 19. The device according to claim 20, wherein the computer (50) is designed to determine whether an oxidative or a reductive shift is present from a comparison of the different measurements with differently sized half-cells of a sensing cell and/or with different SOCs (charge states). The device according to claim 20 or 21 , comprising dosing device (59, 61 ; 60, 62) for adding a reducing agent on the catholy side or an oxidizing agent on the anoly side, which are coupled to the computer (50) for automatically metering the reducing agent or the oxidizing agent in dependence on the calculated shift of the oxidation state.