Battery management system, method and aerial vehicle

Abstract

The present invention relates to a battery management system, method and aerial vehicle for battery state observation and optionally health parameter observation with two redundant, independent and dissimilar channels, in particular state of charge (SOC) observation. Specifically, the SOC determination of a first of the channels is based on Coulomb counting. The other channel employs a different algorithm than Coulomb counting. In embodiments, battery health observation is further conducted independently by the two channels, with the first employing an aging model and the other channel employing a different (dissimilar) algorithm. Based on the state observation and health observation, the state (functional state) of the battery system can be predicted to determine a flight range according to a predetermined flight profile.

Description

TECHNICAL FIELD

[0001] The present invention relates to a battery management system. More specifically, the present invention relates to a battery management system for health monitoring of an energy storage system of an electric aerial vehicle. BACKGROUND

[0002] In recent years, electricity is becoming increasingly important as an energy form for driving aerial vehicles. This includes in particular electric vertical take-off and landing aircraft (eVTOL).

[0003] A key component of an electric (i.e. electrically driven / electrically propelled) aircraft, including an eVTOL, is a suitable energy storage system (ESS). The energy storage system can be implemented in the form of a battery system of rechargeable batteries, which can be structured as a plurality of individual battery cells. The individual battery cells can be combined together to form one or more battery modules that serve as the aircraft battery system of the energy storage system. An example of a battery type suitable for the framework of the present invention is a lithium (Li) ion battery, to which the present invention is not limited.

[0004] Generally, the function of the energy storage system is to provide sufficient available energy for the electrically driven aircraft for safe flight and landing. Since this is generally the case with aerial traffic, the highest safety standards apply to the components of the aerial vehicle including the ESS. In order to ensure safe operation, in particular to ensure safe landing with a sufficient amount of remaining available energy, parameters of the ESS defining a critical state that limits the amount of available energy must be monitored and communicated to the operator.

[0005] Such parameters include, but are not limited to, for example, cell temperature or state of charge (SOC) of individual battery cells of the ESS. Since some of such parameters are not directly measurable quantities, but are related to internal states, the design of the ESS has to foresee appropriate equipment for the respective state observation. State observation is based on measurements of physical data including, but not limited to, terminal cell voltage, cell surface temperature, or current, etc. Moreover, residual errors that can be present in the state determination have to be considered, as this potentially limits the available energy and thus the range of the air vehicle. Such residual errors generally occur in view of the limited accuracy of an arbitrary model used to describe the ESS and its states. Moreover, the available energy of the battery is highly dependent on the flight profile. Therefore, the energy management system has also to make state predictions on the flight profile of the planning until the electrically driven aircraft achieves a safe landing. Specifically, in the framework of the present disclosure, terms like “monitoring” or “observing” are used to indicate that not only the respective data (functional state, health parameter) are determined at a specific point in time, but also that their determination is repeated in order to gather information on their development over time, in particular before and during a specific flight. The interval at which the individual determinations are updated can be set according to the environment and can in particular be set so small that a quasi-permanent observation is possible.

[0006] More generally, it distinguishes between two types of relevant time-dependent variables characterizing the battery system. On the one hand, the (cell) states are defined by system variables that evolve quickly over time (i.e. in seconds according to the system input). Examples of cell states are the cell state of charge, the cell core temperature, or the cell sheet temperature. On the other hand, the health parameters are system variables that evolve slowly over time (i.e. in days according to the system input).

[0007] The importance of health parameter monitoring lies in the fact that the overall state of the battery depends on additional factors, which can reflect changes such as aging over larger time scales during the battery life cycle, in addition to specific variables reflecting the current state of charge, and are summarized under the term “health parameter”. In particular, the battery health parameters can include, but are not limited to, at least one of a cell capacity and a cell impedance.

[0008] Thus, a key task in the design phase of the ESS is that the maximum error of the state prediction and health prediction of the planned flight is known when determining the current state of the ESS, in particular when determining the remaining available energy defining the remaining flight range, and can be considered each time during the planning of a single flight and during the flight itself. The planned profile defines the power consumed over a period of time and shall strictly comply with the operational requirements of the aircraft. The maximum error measure in the state prediction of the ESS over its lifetime shall be considered as a safety margin. This ensures that the residual error in the state observation and state prediction does not particularly affect the utilization of the ESS within its known physical limits.

[0009] Thus, the state prediction based on the state monitoring and health parameter monitoring can ensure that the planned profile does not violate any safety boundaries until a safe landing is achieved. This allows the operator to confirm the available energy and range for a given mission during the flight according to a specific profile and at any time before landing.

[0010] In the following, reference will be made to Figure 1 a schematic diagram illustrating the tasks for state observation and state prediction to be performed by the energy management system during a flight according to a specific flight profile.

[0011] In the upper part of Figure 1 a graph is shown indicating the power required over time during a flight according to a profile. As can be seen from the graph, the power required is particularly high immediately after take-off and at the last stage of the flight, i.e. before landing. In the example shown in the graph, it is assumed that the current time indicated by the symbol of the aircraft in flight is between the beginning of the take-off and landing phases. Thus, with respect to the current time, the previous flight phases have passed and the upcoming flight phases shall be performed according to the planned flight profile. As further indicated, for safety reasons, by the shaded box at the end of the flight, a certain amount of energy should be available at the destination. Thus, the end point on the time scale is defined by the condition that a predetermined residual energy is still available (“end point condition”). In other words, in order to take into account the uncertainty in the state prediction, the indicated later time point that is still reachable on the basis of the remaining available energy (“physical limit”) shall not be considered as reachable in operation.

[0012] During the flight, the state of the ESS is permanently monitored (“state observation”). This includes, but is not limited to, physical measurements, model-based estimations, observations by means of neural networks, and model-based correction / calibration of measurement data. The state observation specifically observes a plurality of functional states (State of Function SOF). The plurality of functional states can include, but is not limited to, for example, cell state of charge (State of Charge SOC), cell core temperature, cell current connector temperature, cell current, and HV (high voltage) cable temperature.

[0013] The state prediction for a future point in time is based on state observations during a flight phase before (past) the current time. In particular, the state prediction can include a look-up table, a model-based prediction and a prediction using a neural network, but is not limited thereto. By considering a pre-determined arbitrary safety margin and a residual, this enables to predict the state until the end of the planned flight profile (e.g. the SOF listed above and indicated in the lower part of the figure). In particular, once the remaining available energy upon landing according to the planned flight profile is below a pre-defined “remaining energy at destination”, an immediate alert to the operator has to be issued to ensure a safe landing at the nearest reachable airport.

[0014] False determination of the available energy can lead to catastrophic failure cases. This classification stems from the assumption that a false display of the available energy would lead the pilot to perform flight maneuvers (in particular flight distances) that the battery cannot sustain with enough energy to continue flying and landing.

[0015] As mentioned above, determining the state or health of a battery, in particular the state of charge or impedance of a battery cell, is generally not a directly measurable quantity. For this, the question arises how to determine the battery state of charge or health in a reliable way that complies with the highest safety requirements applicable in air traffic, in particular the highest safety requirements applicable for electric air vehicles. SUMMARY

[0016] The present invention aims at providing a battery management system and a corresponding method that enables to reliably determine and monitor the battery state of an ESS of an electric air vehicle, thereby complying with the high safety requirements applicable in air traffic.

[0017] This is achieved by the features of the independent claims.

[0018] According to a first aspect of the present invention, a battery management system for an electric air vehicle, the battery management system being adapted to observe a current state of charge of a battery system forming an energy storage system of the air vehicle, the battery management system comprising: two redundant and dissimilar channels for battery state determination, wherein a first channel of the two channels comprises a device for determining a state of charge of individual battery cells of a plurality of battery cells of the battery system by using a coulomb counting algorithm, and a second channel of the two channels comprises a device for determining a state of charge of individual battery cells of a plurality of battery cells of the battery system using a mechanism different from the coulomb counting algorithm.

[0019] According to a second aspect of the application, a battery management method for observing a current battery state of charge, current battery SOC, of a battery system forming an energy storage system of an electric aerial vehicle, the battery management method comprising the steps of determining the SOC of each individual battery cell of a plurality of battery cells of the battery system by using a mechanism based on a Coulomb counting algorithm, and independently determining the SOC of each individual battery cell of a plurality of battery cells of the battery system by using a mechanism different from the Coulomb counting algorithm.

[0020] The specific method of the application is to combine a Coulomb counting algorithm for determining the state of charge (SOC) of each battery cell of an energy storage system of an electric aircraft, in particular an eVTOL, with another SOC determination algorithm in an independent and redundant manner. In particular, in embodiments, the other SOC determination algorithm is a model-based algorithm. The SOC determination is carried out by two redundant and dissimilar channels, one of which uses the Coulomb counting algorithm. The fact that the two channels are redundant means that each channel is able to observe the ESS at all times with full health parameters without having to rely on any decision made by the other channel. In other words, the measurements of the two channels are completely independent of each other. The fact that the two channels are dissimilar means that the algorithms used by the two channels for determining the SOC are different from each other.

[0021] The basic benefit of the method of the application using Coulomb counting is the fact that Coulomb counting is a simple, easy to implement and highly deterministic SOC monitoring algorithm. This reduces the certification risk of the battery management system. For certification reasons, the eVTOL should rely on two redundant and dissimilar cell state of charge observation methods (channels).

[0022] Another benefit of Coulomb counting is that it is low in complexity and therefore low in computational effort. In fact, Coulomb counting only requires integration. This saves weight and cost.

[0023] According to embodiments, the battery management system is further adapted for battery health observation. Each channel further comprises a device for determining a battery health parameter of each individual battery cell of the plurality of battery cells. The determination of the battery health parameter by the first channel is based on an aging model, e.g. an empirical aging model, which determines the battery health based on observed utilization. The determination of the battery health parameter by the second channel is based on a mechanism different from the aging model. More specifically, the cell impedance and / or the cell capacity of each individual battery cell can be determined and monitored as battery health parameters.

[0024] Electric aircraft, in particular eVTOL applications, rely heavily on low battery impedance to provide high hover power requirements. Increased cell impedance is both a major cell aging mechanism and a major cell failure mechanism. Therefore, an increase in cell impedance significantly impacts the available available energy, i.e. the range of the electric aircraft (eVTOL), and thus the safety. Of course, similar considerations apply to the cell capacity, which determines the amount of energy that can be stored in the ESS.

[0025] Therefore, it is important to consider the current state of the battery system health parameters for the state (available energy prediction). Depending on the algorithm used, the health parameter observations can continue during the flight in order to permanently update the health parameters, or it can be assumed that they remain constant during the flight, which is a reasonable assumption given the larger time scale of health parameter changes compared to battery state.

[0026] In an embodiment, the devices of the first channel for determining the battery health parameters comprise devices for measuring the current, voltage and temperature of the individual cells. Furthermore, the devices of the first channel for determining the battery health parameters can be supported by maintenance procedures that are carried out periodically, wherein the maintenance procedures comprise at least a representative charging procedure or a determination of the pulse power curve. Alternatively, designated maintenance procedures and / or charging procedures, rather than algorithms based on aging models, can also be used for battery health parameter monitoring.

[0027] According to an embodiment, the amount of available energy is determined based on the SOC observations and optionally the health parameter observations in each of the two channels in order to determine the flight range based on a model-based state prediction of the planned flight profile. In particular, this is done in the case that no errors are detected in the determination using the data of the two channels.

[0028] In an embodiment, the second channel uses a model-based SOC estimation algorithm for battery health state observation and optionally a model-based cell parameter estimation algorithm. More specifically, the model-based algorithm used by the second channel is based on the use of a dual Kalman filter.

[0029] However, the algorithm employed by the second channel is not limited to a model-based algorithm. Any other suitable algorithm known or to be known to the skilled person is equally applicable within the framework of the present disclosure. For example, this includes the determination of the cell impedance (health parameter) by electrochemical impedance spectroscopy (EIS).

[0030] According to embodiments employing EIS, the second channel comprises a device for exciting the battery cells of the battery system with a sinusoidal current of variable frequency and a device for measuring the voltage response of the individual cells. More specifically, the second channel further comprises processing circuitry for calculating a system impedance spectrum based on the ratio between the input excitation current and the voltage response. Based on this spectrum, various variables characterizing the state of the battery can be derived, including the state of charge as well as cell core temperature and cell tab temperature.

[0031] According to embodiments, the device for SOC determination by using a coulomb counting algorithm in the first channel comprises a device for determining the charging current for each individual cell of the plurality of battery cells constituting the battery system during ground operation and a device for determining the total load current of the individual cells. More specifically, the device for SOC determination by using a coulomb counting algorithm in the first channel further comprises processing circuitry for calculating the state of charge based on the determined charging current and the determined total load current of the individual cells and integrating the measured current over time. Thus, based on the balance of energy input and energy output with respect to the cells, the state of charge is determined in a deterministic manner by the coulomb counting algorithm.

[0032] In embodiments, the second channel comprises battery cell measurement devices for measuring at least one of voltage (e.g. cell terminal voltage), current (e.g. input or output current) and temperature (e.g. surface temperature or cell tab temperature). These parameters can be obtained by measurement and can form the basis for calculating parameters indicative of the state of charge of the battery and the health of the battery according to well-known algorithms taking into account the specific hardware configuration of the battery system and the cells, including model-based algorithms.

[0033] In embodiments, the aerial vehicle is an electric vertical take-off and landing aircraft, eVTOL.

[0034] According to another particular aspect of the present application, there is provided an aerial vehicle comprising a battery management system according to the above-mentioned aspects or embodiments.

[0035] Further features and advantages of the present application will be set forth in the description that follows, and in part will be apparent from the description, or can be learned by practice of the application.

[0036] The embodiments and features of the present application described herein or illustrated in the accompanying drawings can be combined with each other, unless expressly stated otherwise. BRIEF DESCRIPTION OF DRAWINGS

[0037] Additional features and advantages of the present application will be apparent from the following detailed description, taken in conjunction with the accompanying drawings, which illustrate, by way of example, the principles of the application.

[0038] Figure 1is an overview showing state observation and state prediction for an energy storage system of an aerial vehicle flying according to a predetermined flight profile;

[0039] Figure 2 is a diagram showing operation of a dual-channel battery management system for state observation and health observation according to embodiments of the present invention in phases before and during flight;

[0040] Figure 3 is a diagram showing details of state prediction during flight according to a planned flight profile using state of charge observation and battery health observation according to embodiments of the present invention; and

[0041] Figure 4 is a flow chart showing basic steps of an exemplary battery management method. DETAILED DESCRIPTION

[0042] The present invention relates to a battery management system for electrically driven aerial vehicles, in particular for eVTOLs, for determining the state of charge and optionally the state of health of a battery system forming an energy storage system (ESS). This enables determining the amount of energy available in the ESS in order to determine (predict) the flight range according to a predetermined flight profile, in particular at any time during flight.

[0043] To this end, embodiments of the present invention make use of an innovative two-channel battery management system architecture providing two redundant and dissimilar battery state observation channels, one of which determines the state of charge by coulomb counting. In embodiments, the state of health is additionally observed, wherein an aging model is employed in one of the two channels.

[0044] According to the present invention, a first channel of the two channels operates based on coulomb counting for SOC determination. In a preferred embodiment, an aging model is used in the first channel to determine the state of health (SOH) of the battery. The first channel is also able to determine and monitor (observe) other state variables (functional states), such as cell core temperature, by using appropriate algorithms different and dissimilar from those used by the second channel.

[0045] A second channel of the two channels operates based on a different algorithm. In embodiments, this algorithm is a model-based SOC estimation algorithm for state of charge observation, e.g. using a dual Kalman filter. Similarly, according to embodiments, for battery health parameter observation, a cell parameter estimation algorithm is used, e.g. using a dual Kalman filter. In particular, cell impedance and / or cell capacity are used as battery health parameters.

[0046] In each of the channels, the battery state can be predicted via model-based state prediction in the planned flight profile based on the observed states and optionally health parameters. Specifically, the battery state determines the amount of available energy for the flight according to the planned flight profile. The available energy thus determined including a safety margin based on a predetermined maximum error of the state prediction determines the maximum range of the flight and thus can be confirmed as safely reachable for the planned destination.

[0047] Embodiments of the invention provide both the implementation of SOC observation and battery parameter health observation in a way that meets the certification requirements for excluding single point failures. This is achieved by two respective channels that determine the state in parallel, implementing the SOC observation and the SOH observation in a redundant and dissimilar way.

[0048] According to embodiments, the battery cell measurements by the battery cell measurement devices of the respective independent channels are dissimilar to each other. Thus, the measurement of the physically available parameters itself causes dissimilarity of the channels. This avoids the case of a systematic failure due to any essential flaw in the measurement algorithm or principle. Examples of dissimilar measurement schemes are: a scheme using PTC (positive temperature coefficient) elements for temperature measurement and a scheme using NTC (negative temperature coefficient) elements for temperature measurement, a scheme using shunt for current measurement and a scheme using Hall sensors for current measurement, or a scheme using two different ADC (analog-to-digital converter) suppliers for voltage measurement.

[0049] In the following, a detailed description of the operation of the two redundant and dissimilar channels for battery SOC observation and health parameter observation during all flight phases will be given with reference to Figure 2

[0050] Channel 1 (shown in the bottom row of the figure and corresponding to the second channel introduced in the above summary of the invention) uses a conventional model-based SOC estimation algorithm and health parameter estimation algorithm for state observation and health observation, e.g. using a dual Kalman filter. The inputs of the algorithm are the current, voltage and temperature for each individual cell measured by the respective measurement devices of channel 1. The outputs are the estimates of the SOC and the health parameter for each individual cell. To obtain the SOC output and the health parameter output from the measured parameters, respective evaluations are made based on a model.

[0051] ​Model-based estimation algorithms are well known to the skilled person, so a detailed description thereof will be omitted here. They are used to estimate state variables of a system that cannot be measured directly, such as the state of charge of a cell, the cell core temperature, the cell impedance or the cell capacity. The approach of model-based algorithms is generally based on a comparison of the measured output variable values of a system with known input variable values with the output of a model of the system for the same input values, wherein at least one model parameter that characterizes the state variable to be estimated is periodically updated based on the difference between the measured system output and the model output as feedback.

[0052] As input, the current, voltage and temperature of each individual cell as measured by the battery cell measurement equipment for the channels that employ model-based estimation are used. The estimated system variables are, for example, the SOC and the health parameter for each individual cell. In addition, as shown, if both model-based health estimation and model-based SOC estimation are implemented, an interactive update of the results between the model-based SOC estimation and the model-based health estimation can be made. In the same way, other variables that characterize the state of function of the battery, such as the cell core temperature, can be obtained in a similar way with model-based algorithms, in addition to the SOC.

[0053] Channel 2 (shown in the middle row of the figure and corresponding to the first channel introduced in the above summary of the invention) uses a coulomb counting algorithm for SOC monitoring. The input for the coulomb counting during ground operations is the charging current for each individual cell. To know the initial SOC, prior to the charging operation, the initial SOC is determined by means of an open circuit voltage (OCV) measurement that takes into account in particular the loss of charge (energy) due to self-discharge of the battery system. This preliminary step is shown in the left-hand box labeled "Reset SOC with OCV measurement".

[0054] Alternatively, other methods to reset or recalibrate the SOC can be used, such as resetting the SOC to 100% when the charger determines that the battery is fully charged. In case the coulomb counting algorithm is employed, it is necessary to recalibrate periodically to avoid erroneous results due to long-term drift of the observations.

[0055] During flight, the coulomb counting calculates the battery SOC based on all the load currents for each individual cell. The output from the coulomb counting is the SOC for each individual cell.

[0056] Channel 2 further uses an aging model to determine the battery health based on the observed utilization. Examples of parameters characterizing the utilization of the battery system are the charge throughput (Ampere hours Ah), the average temperature and the depth of discharge. The input to the empirical aging model for the observation of the battery health is the current, voltage and temperature measured by the battery management device of channel 2 for each individual cell. In addition, the empirical model can be supported by a dedicated (predefined) maintenance procedure, where the battery health parameters are determined by a representative charging (and / or discharging) procedure or pulse power profile. The output is the respective health parameter for each individual cell. As mentioned above, important examples of health parameters are the cell capacity and the cell impedance.

[0057] In essence, the aging model receives the ESS utilization data collected by the battery management system (BMS) for each flight. After each flight, the collected data is transmitted to a back-end computer ("back-end") located at a maintenance facility outside the air vehicle, where the respective data of the air vehicle components during their lifetime is maintained ("digital twin"). These data can contain, for example, cell temperature profiles, voltage profiles, current profiles and charge throughput (Ampere hours Ah), but are not limited thereto. The benefit of using data profiles or integrated values instead of individual data points is that the amount of data stored (i.e. during the flight in the BMS) and transmitted (to the back-end) is greatly reduced. The aging model is calibrated by the observed battery aging during laboratory tests and relates the observed utilization of one or more last flights to the expected evolution of the battery health parameters according to the calibrated model.

[0058] Considering the nature of the algorithm for the health observation using the aging model, it is not possible to update the health parameters during the flight in this case. Generally, considering the relatively long time scale of the change of the health parameters compared to the actual battery state, it is sufficient to assume that the health parameters remain constant during the flight, so that a constant health parameter (i.e. the same as the one determined immediately before the flight) is used in the state prediction during the flight. Alternatively, the battery health development during the flight can also be inferred from the previously determined battery health history (worst case).

[0059] As mentioned above, the (empirical) aging algorithm can be complemented by a dedicated charging / discharging procedure for the health parameter determination.

[0060] A dedicated charging or discharging procedure is able to calculate the current total capacity of the battery cells as a health parameter of the ESS when charging / discharging the cells with predefined fixed charging conditions and environmental conditions starting from a fixed initial state of charge to a fixed final state of charge. More specifically, the dedicated charging procedure can comprise one of the following procedures (basic principle):

[0061] 1) When discharged by a predefined constant current and at a predefined steady cell temperature, at least one cell of the ESS shall reach a lower cell voltage limit. The ESS thereby reaches a "fully discharged" state. The current total capacity of the ESS is the amount of charge from this fully discharged state charging at a predefined constant current and at a predefined constant temperature until at least one cell of the ESS reaches an upper cell voltage limit.

[0062] 2) When charged by a predefined constant current and at a predefined steady cell temperature, at least one cell of the ESS shall reach an upper cell voltage limit. The ESS thereby reaches a "fully charged" state. The current total capacity of the ESS is the amount of charge from this fully charged state discharging at a predefined constant current and at a predefined constant temperature until at least one cell of the ESS reaches a lower cell voltage limit.

[0063] These two procedures can then be repeated several times to mitigate hysteresis effects and coulombic efficiency effects, which distort the calculation of the current total capacity.

[0064] The dedicated maintenance procedures can comprise, for example, a pulse power test.

[0065] The pulse power test calculates the cell impedance of the ESS at a predefined cell temperature and cell SOC by measuring the voltage response (system response) of the cells exposed to charging or discharging currents (current pulses) of different sizes (system excitation). The internal resistance of a cell at a given SOC and temperature can be calculated as the difference in voltage response divided by the difference in cell current, which is a general approach to determining system properties.

[0066] Note that in the framework of the present disclosure, all measurements and observations are made at the level of individual battery cells. Given the high safety requirements in air traffic, the cell with the lowest capability is always taken as a basis for the evaluation, such as for the determination of the available energy or the range prediction, and for potential failure prediction.

[0067] As further shown in the figure, in each channel, the observations are used for state prediction. The state prediction thus takes into account the latest cell states as well as the latest health parameters. The state prediction is used for reporting (confirming) whether there is sufficient available energy for a planned flight. In particular, a flight profile determines the required power over time for a planned flight, which can be pre-computed, for example, by a flight management system (FMS) during a flight planning phase based on inputs such as an aircraft model, a weather model, and a path model.

[0068] As further shown, the thus determined available energy (range based on the planned flight profile) is displayed to the operator (pilot) of the aircraft. This is done in each of the independent and dissimilar channels separately. Thus, the operator has displays provided separately for each of the channels and can thus compare the results. The operator can compare the separately displayed state prediction results from both channels to each other. If there is a deviation between the predicted states from both channels and if the size of this deviation exceeds a predetermined threshold, this deviation can alert the operator. The operator should then get as close as possible to the nearest airport (vertical take-off and landing airport in case of an eVTOL) in order to land safely. In any case, the operator can personally compare the state prediction results and decide whether the determination is reliable or whether an emergency landing is required.

[0069] During normal operation, i.e. in the absence of large deviations, the minimum of the displayed ranges (available available energy) of both channels will be used as a basis for any decision. In this framework, the "minimum" refers to the value corresponding to the lowest value of the remaining available energy, i.e. the lowest range (remaining safe flight distance). The same principle is applied to the overall evaluation of the battery system based on the measurements and observations related to the individual cells. The cells are always taken as the basis for the evaluation, the "minimum" is determined in the above-mentioned sense of observing state or health parameters for the cells.

[0070] As further shown in the top line of the figure, before the flight (in particular: during the charging operation of the battery system) and during the flight between take-off and landing, the respective operation of both channels is continuously carried out (in addition to the health monitoring based on the aging model as described above).

[0071] Figure 3 It is shown how the available energy of an air vehicle can be determined using the results of the battery state (in particular SOC) observation and the results of the health parameter observation via model-based state prediction.

[0072] Figure 3 The upper part of the figure repeats the upper part of Figure 1 and shows a graph indicating the power required during the flight according to the flight profile.

[0073] The lower part of the figure shows how the redundant and dissimilar channels allow to calculate the current system functional state, in particular the SOC. Based on the observed battery health parameters, the planned flight profile can be used to predict the SOC evolution of the individual cells until the aircraft reaches a safe landing condition. The planned flight profile is only valid if the state prediction excludes a violation of the limits. Thus, the range of the aircraft, in particular of an eVTOL, can be determined based on the flight profile(s) and a destination outside the range of the aircraft can already be safely excluded before take-off.

[0074] To comply with high safety requirements, the maximum error of the state prediction of a planned flight should be known in advance (e.g. during the planning phase of the ESS) so as to be taken into account each time during the planning of an individual flight and during the flight itself. In Figure 3 The maximum error of the state prediction is shown in the lower part of the diagram by the distance between the dotted line and the dashed line in the diagram showing the state-of-charge observation prediction results over time. This error should be measured during laboratory tests and considered as a safety margin against the uncertainty in the state prediction.

[0075] More specifically, the solid line (upper one of the two lines) labeled "worst case" error state prediction corresponds to the SOC evaluation of the flight according to the planned flight profile, assuming that said maximum error is present. The label "worst case" refers to the fact that this prediction includes a maximum overestimation of the available resources (capabilities), i.e. it corresponds to a "worst case" from the pilot's point of view. The actual available ("physical") capabilities can be lower than the estimate in the worst case by a "maximum error" which corresponds to the difference between the solid and the dashed line. As mentioned above with reference to Figure 1 For safety reasons, there must be some residual available energy at the destination ("end point condition"). This corresponds to Figure 3 a 5% safety margin against the uncertainty in the state prediction, as shown in the lower part.

[0076] The dashed line (lower one of the two lines) labeled "physical" state evolution shows the actual residual functional state (e.g. state-of-charge) in case of the prediction with maximum error according to the solid line, i.e. it proceeds below the solid line by a distance corresponding to the maximum error. It is easily understood by the skilled person that the uncertainty of the prediction increases with increasing flight distance, so that the error of the distance between the two lines increases. As a result, at the end point of the planned flight (destination), the maximum error of the state prediction must not exceed the planned safety margin (in this example: 5% corresponding to the maximum error at the end point). This guarantees the possibility of a safe flight and landing even in case of maximum prediction error.

[0077] Figure 4 is a flowchart of an exemplary method that can be performed by a battery management system (BMS) according to an embodiment of the application.

[0078] In the upper part of the flowchart, on the left-hand side, the operations performed by Figure 2 the lower part of the diagram. Specifically, in step S10, respective measurements are performed at the individual battery cells. This includes in particular the measurement of cell voltage, current and temperature.

[0079] In a subsequent step S12, the state of charge SOC is derived based on the measurements. In embodiments, this is done using a model-based approach, where the SOC is derived from an equivalent circuit model. Optionally, also the cell core temperature can be derived from the cell temperature measurements. In embodiments, this is also done using a model-based approach, where the cell core temperature is also derived from an equivalent circuit model. However, the processing in channel 1 is not limited to this, but channel 1 can also use any other method than Coulomb counting. In parallel, step S15 proceeds with the determination of the cell impedance and / or cell capacity as SOH parameters. In embodiments, this is also done by using a model-based approach. In this case, as Figure 2 indicated, the relevant parameters of the equivalent circuit model are updated online during the estimation and between the state (SOC) and SOH estimation. Again, the processing in channel 1 is not limited to this, but any other method than the aging model can also be used for the SOH estimation in channel 1.

[0080] Then, the processing proceeds to step S17, where the individual determination results obtained (estimated) by channel 1 are used as a basis for predicting the battery state, i.e. the remaining available energy for the remaining range according to the planned flight profile. In a subsequent step S19, a corresponding display to the operator on a first display (or first display section) corresponding to the prediction of channel 1. This display can be implemented in various forms, such as by a graphical representation or a numerical or symbolic indicator, as long as it is suitable for an easy and quick grasp of the situation, in particular of a possible imminent emergency.

[0081] On the right-hand side of the upper part of the flow chart, the corresponding operations by Figure 2 channel 2 as shown in the middle part are shown. In particular, in step S20, the measurement of the Coulomb count (CC) used for the SOC determination of the individual cells of the battery system is performed as described above. This thus includes the counting of the total input current of the individual cells during ground charging after the SOC has been reset by OCV measurement and the counting of the total load current (integrated over time) of the individual cells during flight.

[0082] In step S22, the SOC is derived for the individual cells based on the measurements in step S20.

[0083] In parallel, step S23 proceeds with the necessary measurements for the health parameter determination by employing an (empirical) aging model. As described above, this in particular includes the measurement of the current, voltage and temperature for the individual cells.

[0084] Step S25 determines the health parameters such as cell impedance and / or cell capacity for the individual battery cells based on the measurements in step S23.

[0085] Then, the process proceeds to step S27, in which the individual determination results obtained by channel 2 are used as a basis for predicting the state of the battery, i.e. the remaining available energy for the remaining range is defined in accordance with the planned flight profile. In a following step S29, a corresponding display is made to the operator on a second display (or second display portion) corresponding to the prediction of channel 2. Again, this display can be implemented in various forms, such as by graphical representations or numerical or symbolic indicators.

[0086] In a final step S30, the operator compares the displayed state prediction results. In particular, if the difference between the state prediction results of the specific variables of channel 1 and channel 2 exceeds a predetermined threshold, the operator can conclude that there is an error in at least one of the channels and decides to initiate a landing procedure at the nearest available airport, since in case of a malfunction in one of the channels, a reliable prediction of the remaining available energy is not possible. The system can also make the comparison and alert the operator if an error is detected based on the deviation between the channels being too large.

[0087] In summary, the present application relates to a battery management system and method for battery state observation and optionally health parameter observation, in particular cell state of charge (SOC) observation, with redundancy, independent and dissimilar two channels. In particular, the SOC observation in a first of the channels is based on Coulomb counting. The other channel employs a different algorithm than Coulomb counting. In an embodiment, the battery health observation is further made independently by the two channels, with the first channel employing an aging model and the other channel employing a different (dissimilar) algorithm. Based on the state observation and the health observation, the state (functional state) of the battery system can be predicted to determine the flight range according to a predetermined flight profile.

Claims

1. A battery management system for an electric airborne vehicle, the battery management system being adapted to observe the current state of charge (SOC) of a battery system forming an energy storage system of the airborne vehicle, the battery management system comprising: Two redundant and dissimilar channels are used for battery state determination. The first of the two channels includes a device for determining the State of Charge (SOC) of each individual battery cell among the multiple battery cells of the battery system using a coulomb counting algorithm. The second of the two channels includes a device for determining the state of charge (SOC) of each individual cell among the multiple battery cells of the battery system using a mechanism different from the coulomb counting algorithm. The battery management system is also suitable for monitoring battery health. Each of the channels also includes a device for determining battery health parameters for each individual battery cell among the plurality of battery cells. The first channel determines the battery health parameters based on an aging model, which determines battery health based on observed usage. The second channel determines the battery health parameters based on a mechanism different from the aging model.

2. The system according to claim 1, wherein, The cell impedance or cell capacity of each individual battery cell is used as a battery health parameter.

3. The system according to claim 1 or 2, wherein, The device for determining battery health parameters in the first channel includes a device for measuring the current, voltage, and temperature of each individual cell.

4. The system according to claim 1 or 2, wherein, The device for determining battery health parameters in the first channel is supported by a periodically performed maintenance process, which includes at least a representative charging process or the determination of a pulse power curve.

5. The system according to claim 1 or 2, wherein, The second channel uses a model-based SOC estimation algorithm for battery health status observation.

6. The system according to claim 1 or 2, wherein, The second channel uses a model-based cell parameter estimation algorithm for battery health status observation.

7. The system according to claim 1 or 2, wherein, The device in the first channel for determining SOC using a coulomb counting algorithm includes: a device for determining the charging current of each individual cell among the plurality of battery cells constituting the battery system during ground operation, and a device for determining the total load current of each individual cell.

8. The system according to claim 7, wherein, The device in the first channel for determining SOC using a coulomb counting algorithm also includes processing circuitry for calculating the state of charge based on the determined charging current of each individual cell and the determined total load current.

9. The system according to claim 1 or 2, wherein, The second channel includes a device for measuring at least one of the current, voltage, and temperature of each individual cell in the plurality of battery cells of the battery system.

10. The system according to claim 1 or 2, wherein, The air transport vehicle mentioned is an electric vertical takeoff and landing aircraft, or eVTOL.

11. An air transport vehicle comprising the system according to any one of claims 1 to 10.

12. A battery management method for observing the current state of charge (SOC) and battery health of a battery system forming an energy storage system for an electric airborne vehicle, the battery management method comprising: The state of charge (SOC) of each individual cell in the multiple battery cells of the battery system is determined by using a mechanism based on coulomb counting algorithms. The state of charge (SOC) of each individual cell in the multiple battery cells of the battery system is determined independently using a mechanism different from the coulomb counting algorithm. The battery health parameters of each individual battery cell in the plurality of battery cells are determined based on a mechanism using an aging model, which determines the battery health based on observed usage. The battery health parameters of each individual battery cell in the plurality of battery cells are determined independently based on a mechanism different from the aging model.

13. The method according to claim 12, wherein, The cell impedance or cell capacity of each individual battery cell is used as a battery health parameter.

14. The method according to claim 12 or 13, wherein, Using an aging model to determine battery health parameters involves measuring at least one of the following for each individual cell: current, voltage, and temperature.

Images