Method and system for determining the charging current limit for the charging process of a rechargeable battery
The method uses physics-based and data-driven models to predict lithium plating and adjust charging current limits, addressing inefficiencies and safety issues in rapid charging of lithium-ion batteries at low temperatures by preventing lithium plating and optimizing charging processes.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- AVL LIST GMBH
- Filing Date
- 2024-07-05
- Publication Date
- 2026-07-02
AI Technical Summary
Existing methods for rapid charging of lithium-ion batteries at low temperatures are inefficient and energy-intensive, leading to lithium plating, which reduces the battery's capacity and safety, and lack precise control to prevent this degradation.
A method using physics-based and data-driven models to predict lithium plating and adjust the charging current limit, incorporating machine learning for real-time adaptation to battery conditions, thereby preventing lithium plating while maintaining rapid charging.
Enables rapid charging at low temperatures with reduced energy consumption and enhanced safety by accurately predicting and preventing lithium plating, ensuring battery performance and longevity.
Smart Images
Figure 2026521940000001_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a method for determining a charging current limit for a charging process of a rechargeable battery device, particularly under a low operating temperature. Further, the present invention relates to a computer program product for performing each step of the method according to the present invention in a computer-based manner, and further to a control system for controlling a charging process of a rechargeable battery device. Further, the present invention relates to a battery charging system having the control system according to the present invention.
Background Art
[0002] The rapid charging of rechargeable batteries is becoming increasingly significant for many different application fields. Particularly in the vehicle industry, the rapid charging of vehicle batteries should enable vehicles that operate purely electrically to travel even longer distances. Therefore, it is preferable to be able to perform the charging process within an acceptable time.
[0003] The solution known in the prior art enables the rapid charging of batteries under specific conditions. One challenge in the rapid charging of batteries is, in particular, that a high charge amount has to be carried in the battery within a short time due to the progress of the electro-chemical process. Such a process has a temperature dependence that can lead to the degradation of the battery to be charged under temperatures below 15°C and a high charging current intensity.
[0004] The underlying principle of degradation is so-called lithium plating in the case of lithium-ion batteries.
[0005] Lithium plating is the formation of metallic lithium at the anode of a lithium-ion battery during the charging process. This deposition at the anode reduces the amount of lithium ions available in the battery's electrolyte. Simultaneously, it creates a shielding layer at the anode against the free diffusion of lithium ions. Therefore, fewer lithium ions can diffuse into and be inserted into the anode. Such insertion can also be called intercalation.
[0006] One cause of lithium plating is that under high charging currents, lithium ions can accumulate on the surface of the anode, meaning that many lithium ions that could be inserted into it are present on the anode. This can result in lithium ions reacting to form metallic lithium and depositing on the anode. This effect is further amplified by the presence of low charging temperatures, because in that case, the diffusion rate of lithium ions into the anode slows down.
[0007] In prior art, attempts have been made to address the lithium plating problem, for example, by parameter monitoring, as demonstrated in U.S. Patent Application Publication 2017 / 203667A1. In this way, specific parameters, such as the battery discharge voltage, are monitored, and lithium plating is determined to have occurred when a specified limit is reached. However, this type of method is relatively inaccurate and only attempts to affect the charging process after lithium plating has occurred.
[0008] Another device known from prior art addresses the lithium plating problem by switching on external heating for the battery at low temperatures, thereby preventing or at least delaying the occurrence of lithium plating. The drawback of such a solution is that heating the battery is a relatively energy-intensive process, thus prolonging the charging process. Furthermore, due to the complexity of the process carried out in the battery by comprehensively switching on the heating, it is not possible to enable a situation-appropriate control process for the charging process in order to prevent or at least counteract lithium plating. Moreover, such a method requires a specific remaining charge level of the battery in order to enable heating. Therefore, such a method cannot be used with batteries that are particularly heavily discharged, resulting in either having to perform rapid charging despite unfavorable circumstances, or being unable to perform rapid charging at all. [Prior art documents] [Patent Documents]
[0009] [Patent Document 1] U.S. Patent Application Publication No. 2017 / 203667A1 [Overview of the Initiative] [Problems that the invention aims to solve]
[0010] The object of the present invention is to eliminate, at least partially, the drawbacks described above. In particular, the object of the present invention is to provide a method and system that can control the charging process, especially the rapid charging process, to be carried out as quickly as possible, while being energy-efficient at low operating temperatures and avoiding metal plating. [Means for solving the problem]
[0011] The above problems are solved by a method having the constituent elements of claim 1, a computer program product having the constituent elements of claim 13, a control system having the constituent elements of claim 14, and a battery charging system having the constituent elements of claim 17.
[0012] Other advantages and constituent elements of the present invention will become apparent from the dependent claims, the detailed description of the invention, and the drawings. In this regard, constituent elements and details described in relation to the method according to the present invention will naturally also apply in relation to the computer program product, the control system, and the battery charging system according to the present invention, and vice versa, so that there is always mutual reference, or can be made, with respect to the disclosure of individual embodiments of the invention.
[0013] A first aspect of the present invention relates to a method for determining a charging current limit for the charging process of a rechargeable battery device. The method includes the step of determining measurement parameters in the battery device, where the measurement parameters include at least one operating temperature, battery voltage, and battery current of the battery device. Furthermore, battery parameters are determined from a physics-based battery model to reflect the physical processes underway in the battery device. At least one expected temperature progression of the operating temperature and electrode voltage are determined as battery parameters, based on the detected measurement parameters as input parameters to the battery model. In the next step, prediction parameters for the occurrence of metal plating at the electrodes of the battery device are determined, in particular from a data-driven prediction model. At least one predicted time for the occurrence of metal plating is determined, based on at least one determined battery parameter as an input parameter to the prediction model as a prediction parameter. Furthermore, at least one control parameter for controlling the charging process is determined, in particular from a data-driven control model. At least a charging current limit is determined as at least one control parameter, based on the measurement parameters, battery parameters, and prediction parameters as input parameters to the control model, and the determined charging current limit is output.
[0014] To output here can be understood as providing at least one value. This can be understood as simply providing at least one value, for example, but it can also be understood as outputting at least one value, especially for adjustment, for subsequent display and / or reprocessing.
[0015] At this time, the output charging current limit forms the upper limit of the charging current.
[0016] In other words, the present invention can provide a method for determining control parameters for the charging process of a rechargeable battery device. These control parameters may, in particular, be parameters that enable control and / or adjustment of the battery charging process. A rechargeable battery device can be understood, in particular, as a rechargeable electrochemical current storage device.
[0017] In the method according to the present invention, measurement parameters are determined in the battery device. The measurement parameters are detected, in particular, by physical measurements in the battery device. For example, the operating temperature can be determined as the internal temperature of the battery device. The operating temperature may be, for example, the average value of multiple cell temperatures, the housing temperature, etc. The battery voltage can be picked up, for example, as the total voltage at the connection terminals of the battery device and / or detected by determining the individual cell voltages. The battery current can be understood, for example, as the total current strength currently available from the battery device.
[0018] Furthermore, battery parameters are determined from a physics-based battery model that reflects the physical processes taking place within the battery device. A physics-based battery model can reflect the physical processes occurring within the battery device, for example, through equations, specific settings, or other relationships. Based on the detected measurement parameters, such a physics-based battery model can determine battery parameters that characterize the state of the battery device, or the state of its components, from a physical perspective.
[0019] The battery parameters to be determined, in the sense of the present invention, include at least one expected temperature progression of the operating temperature and the electrode voltage. The expected temperature progression is understood, in particular, to be a prediction of the temperature progression over a future time period. For example, the operating temperature and its progression over the next 5 seconds can be determined by a physics-based battery model. The time period for the prediction preferably begins immediately following the detection of the measured parameters.
[0020] The electrode voltage may, in particular, be an estimate of its current value. However, it is also conceivable to understand the electrode voltage as a predicted value for a future time period.
[0021] Battery parameters are used, particularly in data-driven predictive models, to determine the predicted time for metal plating to occur on at least one of the electrodes of the battery device. The time for occurrence may be expressed as, for example, an absolute future date, remaining operating time until metal plating occurs, or remaining charge cycles. The predicted time for metal plating can then be determined from the predictive model, based at least on the determined battery parameters and detected measurement parameters as input parameters to the predictive model.
[0022] Here, metal plating can be understood as the formation of metal deposits, particularly on one of the electrodes of a metal-ion battery. For example, sodium deposition can occur in sodium-ion batteries, and potassium deposition can occur in potassium-ion batteries.
[0023] Unlike battery parameters, the determination of predictive parameters is performed, in particular, by data-driven models. Here, a data-driven model can be understood as a model in which the relationship between input and output quantities is determined and / or modeled using data.
[0024] Furthermore, at least one control parameter for controlling the charging process is determined, particularly from a data-driven control model. At this time, at least the charging current limit is determined as at least one control parameter on the basis of measurement parameters, battery parameters, and prediction parameters as input parameters of the control model.
[0025] Weighting in the sense of the present invention can be understood as a combination of measurement parameters, battery parameters, and prediction parameters, in particular, to which weights are assigned to respective parameters. The weights of respective parameters may be different from each other. The weights can reflect, for example, the influence of individual parameters on the occurrence of metal plating.
[0026] In this way, it is possible to reliably prevent metal plating on one of the electrodes of the battery device, and at the same time, it is possible to control the battery charging process so that the charging process does not slow down. This depends particularly on the fact that battery parameters, which are good indicators for the occurrence of metal plating, can be determined relatively accurately by this method from measured values by an analysis method. According to the present invention, despite the high complexity of the processes occurring in the battery, predictions regarding the occurrence time can be made using battery parameters. The reason this is possible is particularly that the complexity is managed by a data-driven model. Appropriate control parameters can be determined with certainty and accuracy by making use of the occurrence time. This is possible despite the fact that there is also a highly complex relationship between control actions such as the limitation of the charging current and the resulting changes in the electrical and chemical states in the battery. In this case as well, according to the present invention, the existing complexity is managed particularly by a data-driven model. Unnecessary limitation of the charging current during the rapid charging stage is thus eliminated. Another advantage is that the method can be implemented by ordinary battery infrastructure.
[0027] At least one control parameter can preferably be determined to control the charging process, particularly the rapid charging process, under an operating temperature below 15°C, or below 10°C, or below 5°C, or below 0°C, or below -5°C.
[0028] Here, the rapid charging process can be understood as a charging process with a charging power exceeding 50 kW.
[0029] In this way, the rapid charging process can be performed even at a low temperature. At this time, there is no need to worry about an increase in the risk of the occurrence of metal plating. This is because the method can provide an appropriately suitable limitation of the charging current.
[0030] In a preferred embodiment, a plurality of control parameters for controlling the charging process, particularly heating control parameters for controlling the heating of the battery device, can be determined. In particular, the control parameter can have a heating control parameter for controlling the preheating of the battery device. The heating control parameter is preferably intended for pulse rate heating inside the battery device. The heating control parameter preferably has at least one activation and / or deactivation of external or internal heating, activation of heating, deactivation of heating, preheating time, preheating amplitude, charging current frequency, pulse width of the heating charging current, discharge current frequency, pulse width of the discharge current, battery target temperature, operating temperature limit value for activating the preheating of the battery device, or operating temperature limit value for deactivating the preheating of the battery device.
[0031] In this way, the charging current limit is not only adapted to prevent metal plating, but it also becomes possible to actively influence the charging process and / or the battery device. To this end, the internal or external heating of the battery device can be activated or deactivated. At this time, based on a determination of the likely timing of metal plating, the heating can be switched on only when heating is necessary. In this way, energy can be saved and the duration of the charging process can be shortened. In particular, the present invention also makes it possible to consider whether a higher overall charging current can be achieved after heating, which can compensate for the time loss in the initial preheating stage or the preheating stage.
[0032] In a more preferred embodiment, the battery model may be a Kalman filter, a Doyle-Fuller-Newman model, and / or a single-particle model as an analytical model for determining the electrochemical state of the battery device. Here, an analytical model can be understood as a model in which the relationship between input and output quantities is determined and / or modeled using mathematical relationships.
[0033] The battery model configurations described above allow for an analytically accurate description of the current and, potentially, future state of the battery device. In particular, other indicators of metal plating occurrence can also be analytically determined. Examples of such indicators include a gradually decreasing discharge voltage, an increase in electrode resistance, an increase in electrode overpotential, or a change in electrolytic polarization.
[0034] In a preferred embodiment, the control model, the predictive model, or the control model and the predictive model may be based on machine learning techniques, preferably reinforcement learning methods.
[0035] The configurations described above allow for the representation, modeling, and simulation of particularly complex or unknown relationships using data, thereby enabling accurate determination of control parameters.
[0036] In a more preferred embodiment, the charging current limit, as one of the control parameters, can be determined based on the weighting of the measured parameters, battery parameters, and predicted parameters as input parameters of the control model, and in particular, the weighting of the input parameters of the control model can be determined by machine learning in at least two steps. For example, in the learning step, the initial weighting can be determined as the weighting. In the re-evaluation step, the weighting can be determined during the learning process, and in this step, the weighting is fitted based on at least the measured parameters. Preferably, the weighting can be fitted continuously in the re-evaluation step.
[0037] Alternatively or as an addition, the weighting of the input parameters of a predictive model can be determined by machine learning in at least two steps. For example, in the learning step, initial weights can be determined as weights, and in the re-evaluation step, the weights can be determined during the learning process and fitted in this learning process, at least based on the measured parameters. Preferably, the weights can be fitted continuously in the re-evaluation step.
[0038] In this way, each model, especially data-driven models, can be initially trained with data, and the weights determined in this manner can be validated. Thus, the method can be provided with high accuracy and adapted to each process or each battery model. Continuous learning of data-driven models is possible based on a configuration that allows for continuous weighting adaptation. This allows for taking into account the inevitable aging process of battery devices and changes due to damage.
[0039] In a preferred embodiment, the weighting of the input parameters of the control model can be further determined from the time-varying prioritization of the input parameters and the time-varying prioritization of the output control parameters. The prioritization of the input parameters of the control model is preferably performed to assess the risk of metal plating occurrence. The prioritization of the output control parameters is preferably performed to assess the controllability of each control parameter in responding to metal plating occurrence. In this case, the prioritization can be determined on a case-by-case basis, at least by comparing the current values of the measurement parameters and / or battery parameters with the historically relative values of the output control parameters.
[0040] For example, the model can be trained to consider not only the relationship between output and input values, but also their relative importance. In this way, the control behavior can be optimized and, whenever possible, judged as a compromise between conflicting requirements.
[0041] In a preferred embodiment, the measurement parameters may further include at least one cell voltage as the operating voltage of one or more cells in the battery device, the resulting charging current, or a plurality of locally different operating temperatures.
[0042] By using measurement engineering techniques to detect other physical parameters of the battery device, the accuracy of the model results and the control parameters determined accordingly can be further improved. Furthermore, the impact of measurement inaccuracies and outliers in the measurement data can be reduced.
[0043] In a more preferred embodiment, the battery parameters may further include at least one charge state of the battery device, electrode overpotential, or health state of the battery device.
[0044] Preferably, the expected temperature progression of the operating temperature can include the future temporal progression of the operating temperature in one or different areas of the battery device.
[0045] In this way, another indicator of the occurrence of metal plating can be determined.
[0046] In a preferred embodiment, the prediction parameter may further include the predicted remaining charge time.
[0047] In this way, the control model and / or the user of the method can determine further information for controlling the charging process, which can be applied.
[0048] In a more preferred embodiment, the control parameters may further include at least one activation of the charging process, deactivation of the charging process, pulse width of the charging current, or duration of the charging current.
[0049] By determining additional control parameters, it becomes possible to influence the charging process in a different way. In this way, the possibility of addressing metal plating through measures other than limiting the charging current or overheating the battery device is provided.
[0050] In a preferred embodiment, the battery model may have a battery temperature model to reflect the temporal progression and / or local profile of the operating temperature of at least one of the battery devices. Alternatively or in addition, the battery model may have a battery state model to preferably numerically reflect the chemical and / or electrical processes in the battery device. In this case, the battery state model may preferably have as input parameters at least one initial ion concentration in the electrolyte of the battery device, the diffusion rate of ions to one of the electrodes, the reaction coefficient of a chemical reaction proceeding in the battery device, or the conductivity of the electrolyte provided in the battery device.
[0051] In this way, the accuracy of the analytically determined battery parameters can be improved, thereby resulting in an overall improvement in the determination of control parameters.
[0052] In a more preferred embodiment, the battery device may preferably be a lithium-ion battery having one or more battery cells. Thus, lithium plating may occur as metal plating. The electrode may be an anode. The electrode voltage may be an anode voltage. The anode may preferably have graphite and / or metal.
[0053] In this way, it is possible to provide a determination of the charging current limit for lithium-ion batteries.
[0054] A further aspect of the present invention relates to a computer program product having commands that, when the program is executed by a computer, instruct the computer to perform one or more of the steps of the method described above.
[0055] A further aspect of the present invention relates to a control system for determining the charging current limit of a rechargeable battery device. The control system has a measurement module for detecting measurement parameters in the battery device, where the measurement parameters include at least one operating temperature of the battery device, battery voltage, and battery current. The control system further has a battery state module for determining battery parameters from a physics-based battery model to reflect the physical processes progressing in the battery device, where the battery model has at least measurement parameters as input parameters, where the battery parameters include at least one expected temperature progression of the operating temperature and electrode voltage. The control system further has a prediction module for determining prediction parameters from a data-driven prediction model in particular regarding the occurrence of metal plating at the electrodes of the battery device, where the prediction model has at least battery parameters as input parameters, where the prediction parameter has at least one predicted occurrence time of metal plating. The control system further has a control determination module for determining at least one control parameter for controlling the charging process from a data-driven control model in particular, where the control model has measurement parameters, battery parameters, and prediction parameters as input parameters, where at least one control parameter is based in particular on a weighting of these input parameters. At least one control parameter has at least one charging current limit. The charging current limit can be determined as a temporal progression. The control system may further have an output module for outputting the determined charging current limit in order to control the charging process, particularly based on at least one control parameter.
[0056] The measurement module may preferably have at least one current sensor, a voltage sensor, and / or a temperature sensor.
[0057] The control system and computer program products can achieve the same technical effects and benefits already described in the methods described above. In particular, they can accurately determine the control parameters for the rapid charging process of battery devices at low temperatures. Furthermore, they enable precise control of the charging process.
[0058] In a preferred embodiment, the measurement module and / or battery state module may be located in the first computing unit. For example, the measurement module and / or battery state module may be located in the battery control unit of the battery device. Alternatively or additionally, the prediction module and / or control decision module may be located in the second computing unit. In this case, the second computing unit may be located differently from the first computing unit, which is preferable. For example, the prediction module and / or control decision module may be located on an external server or cloud server.
[0059] In this way, for example, the collection and processing of measurement data can be directly implemented in the battery control device of the battery module. This allows for the processing of measurement parameters, for example. In contrast, data-driven models, in particular, can be operated on an external computer with even higher computing power. In this way, the speed of determining the charging current limit can be improved. In this way, for example, the control system can be intended to be a system with real-time capabilities.
[0060] A further aspect of the present invention relates to a battery charging system. The battery charging system has a rechargeable battery device having at least one battery cell. The battery charging system further has a charging connection for linking the electrodes of the battery device to an electric charging device in order to electrically charge the battery device with a charging current. Furthermore, the battery charging system has the control system described above. The output charging current limit is set by the control system as an upper limit of the charging current.
[0061] The battery charging system described above can achieve the same technical effects and advantages already described for the method, the computer program product, and the control system described above. In particular, it allows for precise control of the rapid charging process of the battery device at low temperatures, thereby reducing the risk of metal plating. [Brief explanation of the drawing]
[0062] Other advantages, constituent elements, and specific details of the present invention will become apparent from the following description, which describes embodiments of the invention in detail with reference to the drawings. [Figure 1] This is one embodiment of the method according to the present invention. [Figure 2] This is one embodiment of a control system based on the present invention. [Figure 3] This is another embodiment of the control system based on the present invention. [Figure 4] This is another embodiment of the control system based on the present invention. [Figure 5] This is one embodiment of a battery charging system based on the present invention. [Modes for carrying out the invention]
[0063] Figure 1 illustrates each step of method 100 according to the present invention, and for example, the sequence of commands in a computer program product according to the present invention can also be derived therefrom. Figures 2 to 5 illustrate different embodiments of the control system 10 according to the present invention. Figure 5 shows an embodiment of the battery charging system 90 according to the present invention.
[0064] Method 100 in Figure 1 is designed to determine a charging current limit to affect the charging process of a rechargeable battery device 1000, and for this purpose has a series of steps that can be performed repeatedly during the charging process. Furthermore, it is preferable that each step can be performed repeatedly and consecutively. In this way, the value of the charging current limit can be changed or kept the same in the next iteration.
[0065] A measurement step S20 is performed in which the measurement parameter MP is determined in the battery device 1000. The measurement parameter MP includes at least one operating temperature of the battery device 1000, battery voltage, and battery current.
[0066] Furthermore, a battery state determination step S30 is performed in which battery parameters BP are determined from a physics-based battery model. At this time, the physics-based battery model is configured to reflect the physical processes taking place in the battery device 1000. The battery parameters BP include the expected temperature progress of the operating temperature and the electrode voltage of the battery device 1000.
[0067] In prediction step S40, a prediction parameter VP for the occurrence of metal plating at electrodes 1001 and 1002 of the battery device 1000 is determined, particularly from a data-driven prediction model. At this time, the prediction parameter VP includes at least one predicted occurrence time for metal plating at the electrodes of the battery device 1000 and is determined based on at least one battery parameter BP as an input parameter of the prediction model.
[0068] In the control determination step S50, control parameters KP are determined, particularly from a data-driven control model, to be explored for controlling the charging process. The control parameters KP include at least one charging current limit. The control parameters are determined by the control model based on the weighting of the measured parameter MP, the battery parameter BP, and the predicted parameter VP as input parameters of the control model.
[0069] In output step S60, the determined charging current limit is output for control of the charging process.
[0070] Unlike battery models, predictive and control models are specifically data-driven models, not analytical models. A predictive model may be constructed as, for example, an artificial neural network, a Markov model, logistic regression, or a decision tree-based algorithm. In this case, an input layer may be provided, where each neuron represents an input metric. Furthermore, a hidden layer and an output layer with an output metric may be provided. Each layer may be connected to one another via weighting and an input matrix.
[0071] The input indicators for the prediction model may be, for example, anode overvoltage, the health status of the battery unit 1000, cell voltage, or temperature corresponding to different segments of the battery cell 1003. The output indicators for the prediction model may be the estimated charging duration or the predicted time to a metal plating event. For example, the prediction model may output that the start of metal plating is expected within the next 30 seconds as the predicted time to occur.
[0072] Preferably, the control model can simulate the controller for the operating temperature of the battery device 1000 and the charging current. For example, the control model can determine the appropriate heating configuration for the preheating stage of the battery device 1000. For instance, it can determine the charging and discharging pulses that should be adjusted, and their pulse frequencies.
[0073] The control model can be based on reinforcement learning. The requirements imposed on the charging process can be defined, for example, as a Markov decision process. In this case, the Markov decision process includes an environment, the state of the environment, and feasible actions as components, and selections can be made between these. For example, in this example, preheating of the battery device 1000 and a short charging time can be defined as components of the Markov environment. As the state of the environment, the charging time and predictions about the occurrence of metal plating can be defined. As feasible actions, the preheating operation, the configuration of charging and discharging pulses, and the charging current limit can be defined. In this case, as an action, for example, one can maintain the most recently assumed value, increase it, or decrease it. Preferably, the control model is configured such that the selection of actions is not based on randomness, that is, the selection is deterministic.
[0074] In training steps S41 and S51, weights or other parameters, particularly for data-driven models, can be determined using a sufficient amount of training data. These determined and pre-trained weights or parameters can then be validated using a validation dataset. Preferably, both the training and validation datasets originate from real-world experimental environments. The training step cannot be considered complete while the accuracy of the weights or other parameters is still improving. Alternatively, the training step can be considered complete when the absolute error between the validation dataset and the dataset calculated from the data-driven model becomes sufficiently small.
[0075] Continuous weighting learning and dynamic fitting can be implemented, for example, in re-evaluation steps S42 and S52. At this time, for example, past data and parameters are analyzed and associated with the output values of the predictive or control model.
[0076] Figure 2 shows an example of the control system 10. The control system 10 is suitable for and designed to control the rapid charging process of a rechargeable battery device, such as the battery device 1000 described above.
[0077] The battery device 1000 may be a lithium-ion battery in particular. An example of the battery device 1000 is shown in Figure 4.
[0078] The battery device 1000 may have one or more battery cells 1003. The potential difference of the battery cells 1003 may be reversibly adaptable. Each battery cell 1003 may be integrated as a single voltage source and / or current source. The battery device 1000 may have two electrodes 1001, 1002 for power discharge or power consumption. Furthermore, the battery device 1000 may have additional electrodes for potential measurement and current measurement. Furthermore, the battery device 1000 may be part of a control system 10.
[0079] Furthermore, as shown in Figure 2, the control system 10 has a measurement module 20 for detecting the measurement parameter MP in the battery device 1000. For this purpose, the measurement module 20 or the battery device 1000 may have a sensor suitable for detecting the measurement parameter. Further details about the sensor will be described in the following description relating to Figure 4.
[0080] Furthermore, as shown in Figure 2, the control system 10 further includes a battery state module 30 for determining the battery parameter BP. For this purpose, the battery state module 30 has a physics-based battery model. The physics-based battery model allows, for example, the physical processes occurring in the battery device 1000 to be described as chemical formulas and expressed numerically. As is clear from Figure 2, the battery model has at least a measurement parameter MP as an input parameter and a battery parameter BP as an output parameter.
[0081] The battery status module 30 and the measurement module 20 may both be provided in the first calculation unit 11. The first calculation unit 11 may be, for example, a battery control unit.
[0082] As further shown in Figure 2, the control system 10 further includes a prediction module 40 for determining the prediction parameter VP for the occurrence of metal plating at electrodes 1001 and 1002 of the battery device 1000, particularly from a data-driven prediction model. As is clear from Figure 2, the prediction model can also have measurement parameter MP in addition to the battery parameter BP as an input parameter. It is also conceivable to use only the measurement parameter MP as an input parameter for the prediction model.
[0083] As further shown in Figure 2, the control system 10 has a control determination module 50 for determining a control parameter KP for controlling the charging process, particularly by a data-driven control model. Figure 2 illustrates that the control model has a measurement parameter MP, a battery parameter BP, and a prediction parameter VP as input parameters. The control determination module 50 weights these input parameters and uses this to determine at least one control parameter KP.
[0084] As further shown in Figure 2, the control system 10 has an output module 60 for outputting at least one control parameter KP in order to control the charging process based on at least one control parameter KP. The output module 60 may be configured as a data interface or data bus for transmitting data to other devices outside the control system. Such other devices may be, for example, a battery device 1000 or an external battery charger.
[0085] The prediction module 40, the control decision module 50, and / or the output module 60 may each be provided in the second computing unit 12. The second computing unit 12 may be a cloud server.
[0086] Figure 3 shows a control system 10 with a configuration similar to that of Figure 2. Therefore, the commonalities between each embodiment will not be described below.
[0087] Figure 3 discloses a preferred embodiment of the battery state module 30. The battery state module 30 shown as an example in Figure 3 includes a battery temperature model 31 and a battery state model 32 as battery models.
[0088] For example, using the battery temperature model 31, the expected temporal progression of the temperature of the battery device 1000 can be determined as the temperature parameter TP. Furthermore, the local temperature profile of the battery device 1000 can be analytically determined as another component of the temperature parameter TP.
[0089] The battery state model 32 can be used to simulate chemical and / or electrical processes in the battery device 1000 and output them as battery state parameters BSP. In this case, the battery state model 32 can also take into account parameters different from the measured parameters MP, such as the initial ion concentration and ion diffusion rate of the battery device 1000.
[0090] The battery state parameter BSP and temperature parameter TP can be output as battery parameter BP from the battery state module 30 via the signal connector 34.
[0091] Figure 4 shows the control system 10 having the same configuration as in Figures 2 and 3. Therefore, the commonalities between each embodiment will not be described below.
[0092] However, Figure 4 discloses a preferred embodiment of the measurement module 20.
[0093] Figure 4 shows a control system 10 having a measurement module 20 with multiple sensors to detect a measurement parameter MP in the battery device 1000. Alternatively, the battery device 1000 itself may have such sensors. Figure 4 illustrates that the cell voltage of each battery cell 1003 is determined by a voltage sensor 21. Multiple distributed temperature sensors 22 can detect the local temperature profile of the operating temperature of the battery device 1000. The battery current discharged from or consumed for charging the battery device 1000 can be detected by a current sensor 23 of the measurement module 20.
[0094] A battery device 1000, sensors 21, 22, 23, and a first calculation unit 11 which may have a measurement module 20 and a battery state module 30 may be provided as a design unit, so that, for example, a battery module 1100 can be formed.
[0095] Furthermore, Figure 4 discloses a preferred embodiment in which the prediction module 40, the control decision module 50, and the output module 60 are all included in the second computing unit 12. Accordingly, the prediction module 40, the control decision module 50, and the output module 60 are not shown separately in Figure 4.
[0096] Figure 5 shows a battery charging system 90. This system comprises a battery device 1000 having at least one battery cell 1003 as described above, and a control system 10. Furthermore, Figure 5 illustrates an electric charger 1200 connected to the electrodes of the battery device 1000 via an electrical circuit having a charging connection part of the battery charging system 90, in order to electrically charge the battery device 1000 with a charging current. The charger 1200 may be a component of the battery charging system 90. A charging current limit, determined and output by the control system 10 as at least one control parameter KP, is set by the control system 10 as an upper limit of the charging current for the battery device 1000 or the charger 1200.
[0097] The above descriptions of each embodiment are illustrative only. Naturally, the individual components of each embodiment can be freely combined with one another without departing from the framework of the present invention, as long as it is technically meaningful. [Explanation of Symbols]
[0098] 10 Control Systems 11. First computing unit, battery control unit 12. Second computing unit, cloud server 20 Measurement Modules 21 Voltage Sensor 22 Temperature Sensor 23 Current Sensor 30 Battery Status Modules 31 Battery Temperature Models 32. Battery State Model 34 Signal Connectors 40 Prediction Modules 50 Control Judgment Modules 60 Output Modules 90 Battery Charging System 100 ways 1000 Battery Unit 1001,1002 electrode 1003 Battery Cell 1100 Battery Module 1200 Charging device KP Control Parameters MP measurement parameters BP Battery Parameters TP temperature parameters BSP Battery Status Parameters VP prediction parameters S20 Measurement Step S30 Battery status determination step S40 Prediction Step S50 Control Determination Step S60 Output Step S41, S51 Learning Steps S42, S52 Re-evaluation Step
Claims
1. A method (100) for determining the charging current limit for the charging process of a rechargeable battery device (1000), - A step in which a measurement parameter (MP) is detected in the battery device (1000), wherein the measurement parameter (MP) has at least one operating temperature, battery voltage, and battery current of the battery device (1000). Includes, - Battery parameters (BP) are determined from a physics-based battery model to reflect the physical processes taking place in the battery device (1000), and based on the detected measurement parameters (MP) as input parameters of the battery model, at least one expected temperature progression of the operating temperature and the electrode voltage are determined as the battery parameters (BP), - Prediction parameters (VP) for the occurrence of metal plating at the electrodes (1001, 1002) of the battery device (1000) are determined from the prediction model, and based on at least the determined battery parameters (BP) as input parameters of the prediction model, at least one predicted occurrence time of metal plating is determined as the prediction parameter (VP), - At least one control parameter (KP) for controlling the charging process is determined from the control model, and at least a charging current limit is determined as at least one of the control parameters (KP) based on the measurement parameter (MP), the battery parameter (BP), and the prediction parameter (VP) as input parameters of the control model. - The determined charging current limit is output. A method characterized by the following features.
2. The method according to claim 1 (100), wherein the predicted occurrence time of metal plating is determined from the prediction model, based at least on the determined battery parameter (BP) and the detected measurement parameter (MP) as input parameters of the prediction model.
3. The method according to claim 1 or 2 (100), wherein the charging current limit is determined as a change over time.
4. The method according to any one of the prior claims (100), wherein at least one of the control parameters (KP) is determined to control the charging process, particularly the fast charging process, under operating temperatures below 15°C, below 10°C, below 5°C, below 0°C, or below -5°C.
5. The method (100) according to any one of the prior claims, wherein a plurality of control parameters (KP) are determined for controlling the charging process, and in particular the control parameters (KP) further have heating control parameters for controlling heating of the battery device (1000) by internal pulse rate heating, preferably preheating.
6. The method (100) according to claim 5, wherein the heating control parameters include activation of at least one external or internal heater, deactivation of an external or internal heater, activation of heating, deactivation of heating, preheating time, preheating amplitude, charging current frequency, pulse width of heating charging current, discharge current frequency, pulse width of discharge current, battery target temperature, operating temperature limit for activating preheating of the battery device (1000), or operating temperature limit for deactivating preheating of the battery device (1000).
7. The control model and / or the prediction model comprises a machine learning technique, preferably based on reinforcement learning, and / or an artificial neural network, and / or the battery model is intended to be at least one analytical model, a Kalman filter, a Doyle-Fuller-Newman model, or a single-particle model for determining the electrochemical state of the battery device (1000), according to any one of the prior claims (100).
8. The charging current limit, as one of the control parameters (KP), is determined based on the weighting of the measurement parameter (MP), the battery parameter (BP), and the prediction parameter (VP) as input parameters of the control model, in particular the weighting of the input parameters of the control model and / or another weighting of the input parameters of the prediction model is determined by machine learning in at least two steps, the steps being: - A learning step in which the weighting is determined as the initial weighting (S41, S51), and - Re-evaluation steps (S42, S52) in which weighting is preferably continuously adapted based on at least the measurement parameters (MP). A method (100) according to any one of the prior claims, having the following characteristics.
9. The weighting of the input parameters of the aforementioned control model is as follows: - Preferably, a time-varying prioritization of the input parameters of the control model for evaluating the risk of metal plating, and - Preferably, a time-varying prioritization of the output control parameters (KP) to evaluate the controllability of responding to the occurrence of metal plating using each control parameter (KP). From there, it was determined, Preferably, the prioritization is determined each time by comparing the current values of at least the measurement parameter (MP) and / or the battery parameter (BP) with a historically relative output value of the control parameter (KP) according to any one of the prior claims (100).
10. - The measurement parameter (MP) further comprises at least one cell voltage as the operating voltage of one or more cells of the battery device (1000), the charging current being generated, or a plurality of locally different operating temperatures, and / or - The battery parameter (BP) further comprises at least one electrode overpotential, the health status of the battery device (1000), or the charge status of the battery device (1000), preferably the expected temperature progression comprising the future temporal progression of the operating temperature in one or different areas of the battery device (1000), and / or - The prediction parameter (VP) further includes the predicted remaining charging time, and / or, - The method according to any one of the prior claims (100), wherein the control parameter (KP) further comprises at least one activation of the charging process, deactivation of the charging process, pulse width of the charging current, or duration of the charging current.
11. The aforementioned battery model, - A battery temperature model (31) for reflecting the temporal progression and / or local profile of at least one operating temperature of the battery device (1000), and, - The method (100) of any one of the prior claims, comprising a battery state model (32) for numerically reflecting the chemical and / or electrical processes in the battery device (1000), wherein the battery state model (32) preferably further has as input parameters the conductivity of at least one electrolyte provided in the battery device (1000), the initial ion concentration in the electrolyte provided in the battery device (1000), the diffusion rate of ions to one of the electrodes (1001, 1002), or the reaction coefficient of a chemical reaction proceeding in the battery device (1000).
12. The battery device (1000) is preferably a lithium-ion battery having one or more battery cells (1003), the metal plating is lithium plating, the electrodes (1001, 1002) are anodes, the electrode voltage is the anode electrode, and the anode is preferably graphite and / or metal, according to the method (100) of any one of the prior claims.
13. A computer program product having commands that, when the program is executed by a computer, instruct the computer to perform each step (100) of the method according to any one of claims 1 to 12.
14. A control system (10) for determining the charging current limit of a rechargeable battery device (1000), - The battery device (1000) has a measurement module (20) for detecting measurement parameters (MP), the measurement parameters (MP) include at least one operating temperature of the battery device (1000), battery voltage, and battery current. In the control system, - The battery device (1000) has a battery state module (30) for determining battery parameters (BP) from a physics-based battery model to reflect the physical processes taking place in the battery device, the battery model having at least the measurement parameters (MP) as input parameters, and the battery parameters (BP) having at least one expected temperature progression of the operating temperature and electrode voltages. - The battery device (1000) has a prediction module (40) for determining prediction parameters (VP) for the occurrence of metal plating on the electrodes (1001, 1002) from a prediction model, wherein the prediction model has at least the battery parameters (BP) as input parameters, and the prediction parameters (VP) have at least one predicted occurrence time for metal plating. - A control determination module (50) for determining at least one control parameter (KP) for controlling the charging process from a control model, wherein the control model has the measurement parameter (MP), the battery parameter (BP), and the prediction parameter (VP) as input parameters, at least one of the control parameters (KP) is based in particular on the weighting of the input parameters, and at least one of the control parameters (KP) has at least one charging current limit, and - In particular, to control the charging process based on at least one of the control parameters (KP), the system has an output module (60) for outputting a determined charging current limit. A control system characterized by the following features.
15. The measurement module (20) and / or the battery status module (30) are provided in the first calculation unit (11), preferably in the battery control unit of the battery device (1000). The control system (10) according to claim 14, wherein the prediction module (40) and the control determination module (50) are provided in a second computing unit (12), preferably on an external server or cloud server, and preferably the second computing unit (12) is different from the first computing unit (11).
16. The control system (10) according to claim 14 or 15, wherein the measurement module (20) has at least one voltage sensor (21), a temperature sensor (22), and / or a current sensor (23).
17. A battery charging system (90), - A rechargeable battery device (1000) having at least one battery cell (1003), -In order to electrically charge the battery device (1000) with a charging current, a charging connection part is provided to link the electrodes (1001, 1002) of the battery device (1000) with an electric charging device (1200). It has, Having a control system (10) according to any one of prior claims 14 to 16, The output charging current limit is set by the control system (10) as the upper limit of the charging current. A battery charging system characterized by the following features.