Control of an electrochemical device with integrated diagnostics, prognostics and lifetime management

Abstract

A method for controlling the operation of an electrochemical device having at least one operating organ, comprising the steps of: receiving measurements related to the operation of the electrochemical device, and estimating at least diagnostics data based on said measurements, estimating prognostics data based on said diagnostics data and providing operation instructions to control said operating organ of the electrochemical device, said operation instructions being optimized with respect to said estimated diagnostics and prognostics data.

Claims

1. A method for controlling an operation of an electrochemical device comprising at least one internal regulatory system, the method comprising the steps of: receiving measurements related to the operation of the electrochemical device, and estimating at least diagnostics data based on said measurements, estimating prognostics data based on said estimated diagnostics data and providing operation instructions to control said internal regulatory system of the electrochemical device, said operation instructions being optimized with respect to said estimated diagnostics and said estimated prognostics data, wherein a feedback of operation instructions data is provided to estimate said prognostics data, and said prognostics data are estimated further on the basis of said operation instructions, and wherein said diagnostics data includes reversible disruption data, and said operating instructions are based at least on said reversible disruption data so as to counteract current reversible disruption, and each occurrence of determining said operating instructions on the basis of reversible disruption data being stored in a memory, and said prognostics data being estimated further on the basis of a number of occurrences of counteracted reversible disruptions, wherein said diagnostics data includes at least data of a current state of health of the electrochemical device, including an assessment of a possible degradation of the electrochemical device, and the prognostics data include at least an estimation of a remaining useful life time of the electrochemical device, and where said reversible disruption data are related to faulty operation of the electrochemical device and are detected by variation of at least externally measured variables including at least a voltage response of the electrochemical device.

2. The method of claim 1, wherein said internal regulatory system is controlled so as to lengthen remaining useful life time of the electrochemical device.

3. The method of claim 1, comprising the steps of determining the end of the remaining useful life time of the electrochemical device, and generating an alarm signal if said remaining useful life time is below a predetermined threshold.

4. The method of claim 1, comprising the steps of determining the end of the remaining useful life time of the electrochemical device, and instructing said internal regulatory system with modified nominal operation parameters so as to use the electrochemical device with said modified nominal operation parameters, in view of lengthening said remaining useful life time.

5. The method of claim 1, wherein said prognostics data are estimated further on the basis of expected future environment data.

6. The method of claim 1, wherein, said electrochemical device comprising at least one fuel cell having a fuel canal, said measurements include a monitoring of variations of voltage provided by the fuel cell, and said internal regulatory system comprises an air bleed inlet in the fuel canal so as to evacuate contaminant from the fuel canal if a decrease of an average voltage value is observed from said measurements.

7. The method of claim 1, wherein, said electrochemical device comprising at least one fuel cell having an oxidant canal, said measurements include a pressure loss measurement in said oxidant canal, sensed by one or several sensors provided in said oxidant canal, and said internal regulatory system comprises a moisture controller in said oxidant canal so as to: dry the oxidant canal if pressure loss is detected as being higher than a first threshold, and humidify the oxidant canal if pressure loss is detected as being lower than a second threshold.

8. The method of claim 7, wherein, said electrochemical device comprising at least one fuel cell having the fuel canal, said measurements include further a monitoring of variations of voltage provided by the fuel cell, and said internal regulatory system comprises an air bleed inlet in the fuel canal so as to evacuate a contaminant from the fuel canal if a decrease of an average voltage value is observed from said measurements, while said pressure loss is detected as being between said first and second thresholds.

9. A non-transitory computer storage medium storing a computer program comprising instructions to implement the method according to claim 1, when said instructions are run by a computer processor.

10. A computer control unit comprising at least an input interface to receive measurements, a processor and at least a memory unit to perform the method according to claim 1, and an output interface to provide operation instructions to control at least one internal regulatory system of an electrochemical device.

11. An electrochemical device including at least one sensor connected to the computer control unit according to claim 10, and at least one internal regulatory system connected to the computer control unit to operate the electrochemical device on the basis of operating instructions provided by the computer control unit.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 illustrates the flow of information and actions in an embodiment of the invention, in the considered example of PEMFC for μ-CHP application;

(2) FIG. 2A illustrates a flowchart of an algorithmic method according to an embodiment of the invention;

(3) FIG. 2B illustrates a internal device provided in an electrochemical system (such as a fuel cell) to perform the method of FIG. 2A in an exemplary embodiment;

(4) FIG. 3 illustrates the time dependence of the single cell voltage on CO concentration in the hydrogen canal of the fuel cell, and air bleed concentration;

(5) FIG. 4 illustrates the time dependence of the single cell voltage on CO concentration and air bleed concentration, under the regime of one of the proposed control strategies according to the invention;

(6) FIG. 5 illustrates the sensitivity of different types of membrane electrode assembly (MEA) on CO contamination (carbon monoxide) in a fixed set of operating conditions 2.5 hours after CO concentration set (the different sensitivity level illustrates the usefulness of an active air bleed control);

(7) FIG. 6 shows the influence of the presence of water in the oxygen canal, on the fuel cell voltage and on the pressure loss of oxygen in the canal.

DETAILED DESCRIPTION

(8) Referring to FIG. 1, an electrochemical device 10 (for example a fuel cell such as a micro Combined Heat and Power system μCHP) includes one or several sensors so as to provide measurements M to a component 11 of a computer circuit unit, for: assessing diagnostics, estimating thus a State of Health SoH of the fuel cells, and identifying any reversible disruption to counteract.

(9) Data of Reversible Disruption RD are provided further to a control component 13 of the computer circuit unit.

(10) The State of Health SoH is used by a prognostics component 12 of the computer circuit unit, together with information about the Expected Future Environment EFE of the system. Data of the Expected Future Environment EFE are therefore data which can be input in the computer circuit unit.

(11) The control component 13 estimates a Control Action (CA). The prognostics component 12 receives control action data CA and calculates the Remaining Useful Life (RUL).

(12) More particularly, the control component 13 can generate a control action command CA that in part compensates the Reversible Disruption identified by the diagnostics component 11, and in part can be optimised (dashed arrow OPT), interacting with the prognostics component 12, to maximise the Remaining Useful Life (RUL) or a function thereof. Control Action data CA are then fed to the μCHP system to implement with its low-level internal regulatory system (LLIRS as referenced in FIG. 2B commented below).

(13) Referring now to FIG. 2A, in a first step S10, measurements M are acquired and sent to the diagnostics component at step S11. Then, a state of health SoH of the electrochemical device can be estimated in step S111. The state of health SoH can be transmitted then to a prognostics component PROG at step S12. The prognostics assessment can be based on the state of health SoH, but also on an expected future environment EFE (such as for example a nominal voltage or current to be provided). On that basis, the remaining useful life RUL of the electrochemical device can be assessed in step S121, and further data related to reversible disruption RD can be transmitted to a control component in step S13. The aforesaid control component CTRL can elaborate then control commands CA in step S131 so as to operate the electrochemical device with respect to these control commands CA. To that end, the control commands CA can be provided in step S14 to a low-level internal regulatory system LLRIS which is usually provided in electrochemical devices.

(14) Furthermore, a feedback loop is provided so as to optimize in step S132 the estimation of the prognostics based on possible new control commands.

(15) Referring now to FIG. 2B, the electrochemical device 10 comprises an anode AN, a cathode CAT, and further a low-level regulatory system LLRIS, so as to control current and/or voltage provided by the electrochemical device. More particularly, the electrochemical device 10 according to the invention further comprises a computer circuit unit CCU, connected to a sensor unit SEN.

(16) The sensor unit SEN can include sensors, for example for: measuring the voltage which is provided by the electrochemical device (voltage value and possibly voltage variations), measuring preferably pressure at the input and the output of a canal (such as for example the oxygen canal of a fuel cell, so as to measure a pressure loss of oxygen due to moisture in the canal), possibly but not mandatorily, measuring concentration of carbon monoxide CO (usually in the hydrogen canal of a fuel cell), and for any other possible measurements.

(17) The computer circuit unit CCU includes: an input interface INT to receive measurements M from the sensor unit SEN; a processor PROC for running instructions of a computer program according to the present invention, and including more particularly computer components as described above: diagnostics component 11, prognostics component 12, and control component 13; a memory unit MU2 including for example a working memory WM which stores the instructions of the computer program according to the invention, and possibly any other data (temporarily, for example for the program run's sake, or permanently, for example for storing data related to incidents during the operation of the electrochemical device: dryness or flooding of the oxygen canal, CO contamination of the hydrogen canal, etc.); possibly a memory unit MU1 (which can be the same as memory unit MU2) having for example an input to store data such as the aforesaid expected future environment EFE; at least one output OUT to send control signals so as to command the low-level regulatory system LLRIS, the low-level regulatory system LLRIS being further connected to one or several inlet valves 21 (so as to inject an air bleed for example in the hydrogen canal to evacuate CO), or bubbler 22 (so as to dry or humidify the oxygen canal of a fuel cell for example), or any other element for controlling the operation of the electrochemical device; a communication port COM (such as a connection to a display unit, or an antenna to send data) so as to inform a user of a near end of lifetime of the electrochemical device for example, or of any other future failure.

(18) FIG. 5 shows the influence of CO contamination on different types of MEA (assembled stack of proton exchange membranes for example, or alkali anion exchange membrane, catalyst and/or flat plate electrode). There appears an anode over-potential, in any MEA case, as the CO content grows in the anode gas inlet.

(19) Therefore, a parameter to monitor is the concentration of CO, usually in the hydrogen canal (typically in a case where a first canal of the fuel cell is provided with hydrogen as a first reactant, and, for example, a second canal is provided with oxygen as a second reactant).

(20) As shown on FIG. 3, the presence of CO can be detected through voltage variations CFV (with a diminution of the average voltage as shown between 40 and 45 hours of operation in the example of FIG. 3). These unwanted problems of voltage variations can be however overcome thanks to an air bleed ABC which can be injected by steps in the hydrogen canal so as to retrieve progressively a stable voltage CFV having an expected value. In practice, the injected air reacts with CO to produce CO.sub.2 which is then easily eliminated.

(21) Therefore, the voltage variations can be monitored or sensed so as to detect CO contamination and the control of air bleed concentration ABC in the hydrogen canal (for example through an inlet valve or the like) is a way to solve the problem of such voltage variations.

(22) More particularly, FIG. 3 illustrates the time dependence of a single cell voltage on CO concentration and air bleed concentration. Moreover, a possible control strategy is shown. Starting after 55 hours of operation, CO flow is started (COF), which is kept constant for the rest of the shown test. First a short air bleed flow (ABC) at medium concentration (1.4%) is applied, showing no observable increase in cell voltage (CFV). Then, air bleed is set to 0 (resulting in significant voltage decay) and then increased in 0.2% steps with approximatively 3 hours as hold time. At 1.4% air bleed concentration, an almost stable cell voltage is obtained. To validate this operating point, the concentration is further stepwise increased up to 2%, and then maintained.

(23) FIG. 4 shows an example of constant regulation of the air bleed in an embodiment of the invention. In regular intervals (1 hour in the shown example), the air bleed concentration ABC is increased for e.g. 5 minutes (test phase). If no voltage increase is resulting, the air bleed will be reduced e.g. by 0.1%. The term “voltage increase” in this context may include additional derived data (e.g. “voltage noise level decrease”) or other derived functions, including functional combinations of them.

(24) If the cell voltage CFV increases during test phase, a second air bleed concentration increase will be provided for another e.g. 5 min, if the cell voltage increases, a third increase will be performed. If no significant effect (voltage increase) is achieved, the air bleed will be set from the last period (e.g. the last 1 hour) holding value upwards to a level corresponding to the number of air bleed concentration increases in the test phase.

(25) The resulting values may be limited to upper and lower boundary values e.g. between 0.1 and 2%. Preferred test times are 0.1 min to 60 min, specially preferred times are from 1 to 15 min. Preferred hold times are from 5 min to 240 min, specially preferred hold times are from 15 to 120 min. Additional test times may be triggered by voltage decay or other derived data. The controller strategy is shown in two examples.

(26) The control strategies are not limited to the test/hold strategies described above, but is covering any analogue strategy, where air bleed is applied, and a cell voltage (and/or derived data) are analysed, and a resulting measure (air bleed increase/decrease/no change) strategy applies.

(27) The air bleed (e.g. hold) values adjusted by such an algorithm may be used estimate CO concentrations e.g., but not limited to a use of a calibration curve family or a mathematical function for trace back to CO concentrations and may be used in direct or converted form for system control and/or stack SoH estimation.

(28) FIG. 6 shows the influence of the presence of water (liquid in the form of droplets) in the oxygen canal, on the fuel cell voltage and on the pressure loss of oxygen in the canal. More particularly, in the given example, the general stack temperature Tsta is maintained at a given level (for example 50° C.), while the temperature Tsat can be controlled for regulating the saturated vapour pressure of water in the oxygen canal (using for example to that end a bubbler or the like). Therefore, in the given example, the oxygen canal is dried, at first (left part of FIG. 6), leading to an observed diminution of the cell voltage CFV, and then, the temperature Tsat is increased so as to introduce moisture in the oxygen canal. It appears, at first, that the fuel cell voltage increases, but decreases then owing to a too large number of water drops in the oxygen canal (at the abscissa of 75).

(29) Therefore, the voltage diminution parameter here is not sufficient to detect either dryness or flooding of the oxygen canal. Advantageously, another parameter related to pressure loss PLO in the oxygen canal can be detected. Therefore, this parameter can be sensed by using pressure sensors at the input and the output of the oxygen (and/or hydrogen) canal(s) and the sensor unit SEN of FIG. 2B can include as for an example at least such pressure sensors.

(30) More particularly, in a possible embodiment, if voltage CFV is being detected as lower than a threshold: while pressure loss PLO is detected as being higher than a first threshold, then temperature Tsat is decreased in the regulation so as to dry the oxygen canal and observe a voltage increase, while pressure loss PLO is detected as being lower than a second threshold, then temperature Tsat is increased in the regulation so as to humidify the oxygen canal and observe a voltage increase, while pressure loss PLO is detected as being between the first and second thresholds, then CO contamination can be suspected and air bleeding flow is increased so as to stabilize voltage to an upper average nominal value.

(31) Moreover, it has been found that dryness and flooding in the fuel cell canal can lead to mechanical dilatations and stress which can finally damage the fuel cell, involving its end of life. In an embodiment then, each occurrence of a Tsat regulation due to dryness or flooding of the canal is stored in a memory unit (MU2 for example as shown on FIG. 2B) so as to estimate the remaining useful life of the fuel cell. As for an example, if the occurrences' number of Tsat regulation becomes higher than a predetermined threshold, then an alarm signal can be generated and sent through a communication link (a wireless link for example, as shown on FIG. 2B) by a communication interface COM, or to a display unit provided in the fuel cell armature (not shown on FIG. 2B) so as to warn a user for example of a near end of the fuel cell lifetime. In an alternative or complementary embodiment, the operating point of the fuel cell can be modified by the computer circuit unit CCU so as to operate with adapted nominal parameters (for example lower (or higher, depending on diagnostics data) hydrogen and/or oxygen flows, lower (or higher) stack temperature, etc.).

(32) Of course, the invention is not limited to the embodiment described above as an example; it extends to other variants.

(33) The invention can be used with fuel cells as disclosed in the detailed specification above, but more generally with any electrochemical device having parameters which can be sensed to perform diagnostics and possibly also prognostics, so as to estimate for example the remaining useful life (RUL) of an electrochemical device, and to employ regulation to maximise the RUL or a closely related function thereof.

(34) More particularly, the invention enables to diagnose and/or detect drifts towards disruptive conditions in the electrochemical device and appropriately counteract these.

(35) The electrochemical device can be of the type providing heat and/or power to an end-user and/or a distribution grid. It can be also of the type providing power to at least one electric motor in a vehicle and/or a robot. It can be also of the type storing energy, chemically or electrochemically. The electrochemical device can be used further to decontaminate and/or clean wastewaters or flue gases, or to concentrate and/or extract a compound from a gas and/or liquid phase.

(36) Furthermore, in a general embodiment, sensor data history can be stored over time and employed in the calculations of the estimated lifetime (such as the parameters related to oxygen pressure loss in the example given above, but other parameters can be used alternatively or complementarily). Faulty operation can be detected by variation of externally measured variables and observation of the device voltage response, but also externally measured variables can be adjusted to a level that guarantees non faulty operation (such as the temperature Tsat).

(37) A given degradation level can be estimated then by measuring output variables of the electrochemical device by means for example of cyclic measurements and possible changes around an operating point. That embodiment enables detection of a faulty condition, at least by statistical analysis of the device voltage measurements.

(38) Furthermore, imminent failure/irreversible damage of the electrochemical device can be detected by changes in one or more output variables as explained above.