Method and system for determining charging current limit for rechargeable battery charging process
By measuring temperature, voltage, and current on a lithium-ion battery, and using a data-driven model to predict lithium deposition time, the charging current and heating control are dynamically adjusted, solving the lithium deposition problem during rapid charging of lithium-ion batteries at low temperatures, and achieving precise control and improved energy efficiency.
Patent Information
- Application Number
- CN202480046004.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2023-07-07
- Filing Date
- 2024-07-05
- Publication Date
- 2026-02-13
AI Technical Summary
When lithium-ion batteries are rapidly charged at low temperatures, lithium deposition easily occurs, leading to battery degradation. Furthermore, existing methods such as parameter monitoring and external heating schemes are either inaccurate or energy-intensive.
By measuring temperature, voltage, and current on the battery device, and combining this with physics-based and data-driven models to predict the timing of lithium deposition, the charging current limit and heating control parameters are dynamically adjusted to avoid lithium deposition and optimize the charging process.
It achieves precise control of the fast charging process under low-temperature conditions, avoids lithium deposition, saves energy and shortens charging time.
Smart Images
Figure CN121532925A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present invention relates to a method for determining a charge current limit value for a charging process of a rechargeable battery device, in particular at low operating temperatures. The present invention further relates to a computer program product for executing the steps of the method of the present invention based on a computer, and to a control system for controlling a charging process of a rechargeable battery device. Furthermore, the present invention relates to a battery charging system having the control system of the present invention. BACKGROUND
[0002] Fast charging of rechargeable batteries is increasingly important for a variety of different application fields. In particular in the automotive industry, fast charging of vehicle batteries is intended to enable long-distance travel of purely electrically driven vehicles as well. For this purpose, the charging process should be completed within an acceptable time frame.
[0003] The solutions known from the prior art allow fast charging of batteries under certain conditions. A major challenge in fast charging of batteries is, among other things, that a large amount of electrical charge must be transported within the battery by the course of electrochemical processes within a short time. These processes have a temperature dependency, which at temperatures below 15°C and high charging current intensities can lead to a degradation of the battery to be charged.
[0004] The principle of action leading to the degradation in lithium-ion batteries is the so-called "lithium deposition".
[0005] Lithium deposition refers to the formation of metallic lithium at the anode of a lithium-ion battery during the charging process. By the deposition on the anode, on the one hand, the number of lithium ions available in the electrolyte of the battery is reduced. At the same time, a barrier layer is formed on the anode, which hinders the free diffusion of lithium ions into the anode. As a result, fewer lithium ions can diffuse into the anode and intercalate therein. This intercalation can also be referred to as "intercalation".
[0006] One reason for lithium deposition is that under high charging current conditions, a lithium ion enrichment can occur on the anode surface, i.e. there are more lithium ions present at the anode than can intercalate therein. This can lead to a lithium ion reaction to form metallic lithium and deposit on the anode. This effect is further exacerbated by the presence of low charging temperatures, as the speed of diffusion of lithium ions into the anode is slowed down at such times.
[0007] In the prior art, as shown in US 2017 / 203667 A1, attempts are made to solve the problem of lithium deposition by means of parameter monitoring. Thereby, specific parameters of the battery, such as its discharge voltage, are monitored and, when a set limit value is reached, it is determined that lithium deposition has occurred. However, such a method is less accurate and only attempts to intervene in the charging process when lithium deposition has already occurred.
[0008] Other devices known from the prior art solve the problem of lithium deposition by switching on an external heating device for the battery at low temperatures to thus prevent or at least delay the occurrence of lithium deposition. The disadvantage of such a solution is that the battery heating is a relatively energy-intensive process, thus prolonging the charging process. Furthermore, due to the complexity of the internal processes of the battery, a charging process control process that is adapted to the situation cannot be achieved by switching on the heating device in general, thus preventing or at least inhibiting the occurrence of lithium deposition. Furthermore, this process requires the battery to have a certain amount of residual charge to achieve heating. Thus, precisely for batteries that are deeply discharged, this process cannot be achieved, thus either having to be carried out under disadvantageous conditions or not being able to be carried out at all. SUMMARY
[0009] The task of the present invention is to at least partially solve the above-mentioned disadvantages. The task of the present invention is in particular to provide a method and a system by means of which a charging process, in particular a fast charging process, can be carried out as quickly as possible at low operating temperatures in an energy-efficient manner and with avoidance of metal deposition.
[0010] The above-mentioned task is solved by a method having the features of claim 1, a computer program product having the features of claim 13, a control system having the features of claim 14 and a battery charging system having the features of claim 17.
[0011] Further advantages and features of the present invention result from the dependent claims, the description and the figures. Here, the features and details described for the method of the present invention apply equally to the computer program product, the control system and the battery charging system of the present invention and vice versa, so that the disclosure of the respective invention aspects always refers to each other and can be cross-referenced.
[0012] The first aspect of the present invention relates to a method for determining a charging current limit value for a charging process of a chargeable battery device. The method comprises a step in which a measurement parameter is determined on the battery device. Here, the measurement parameter comprises at least an operating temperature of the battery device, a battery voltage and a battery current. Furthermore, a battery parameter is determined from a physics-based battery model for characterizing physical processes occurring within the battery device. Based on the acquired measurement parameters as input parameters for the battery model, at least an expected temperature evolution of the operating temperature and an electrode voltage are determined as the battery parameter. In a further step, a prediction parameter for the occurrence of metal deposition at a certain electrode of the battery device is determined from a prediction model, in particular a data-driven prediction model. Here, at least a predicted occurrence time of the metal deposition is determined as the prediction parameter based on the determined battery parameter as input parameter for the prediction model. Furthermore, at least one control parameter for controlling the charging process is determined from a control model, in particular a data-driven control model. Here, at least the charging current limit value is determined as the at least one control parameter based on the measurement parameter, the battery parameter and the prediction parameter as input parameters for the control model, and the determined charging current limit value is output.
[0013] Herein, "output" can mean, inter alia, providing at least one value. Herein, this can mean, inter alia, simply providing at least one value, but also outputting at least one value for further representation and / or further processing, in particular conditioning.
[0014] Herein, the outputted charge current limit is the upper charge current limit.
[0015] In other words, the application can provide a method by means of which a control parameter for a charging process of a rechargeable battery device can be determined. The control parameter can be, inter alia, a parameter which allows a battery charging process to be controlled and / or conditioned. The rechargeable battery device can be, inter alia, an electrochemical energy store which can be charged.
[0016] In the method of the application, a measurement parameter is determined on the battery device. The measurement parameter is acquired, inter alia, by means of a physical measurement on the battery device. Thus, for example, an operating temperature can be determined as an internal temperature of the battery device. The operating temperature can also be, for example, an average value of a plurality of cell temperatures, a housing temperature or a similar value. A battery voltage can be picked up, inter alia, as a total voltage at the connection contacts of the battery device and / or acquired by determining individual cell voltages. A battery current can mean, inter alia, the total current strength which the battery device can currently provide.
[0017] Furthermore, a battery parameter is determined from a physics-based battery model for characterizing a physical process which occurs in the battery device. The physics-based battery model can characterize a physical process which occurs in the battery device by means of, for example, equations, constant presets or other relationships. By means of the acquired measurement parameters, the physics-based battery model can determine a battery parameter which characterizes the state of the battery device or the state of a battery device component from a physical point of view.
[0018] In the sense of the application, the determined operating parameters comprise at least an expected temperature evolution of the operating temperature and an electrode voltage. The expected temperature evolution should mean, inter alia, a prediction of the temperature change over a future time period. For example, the operating temperature and its change over the next 5 seconds can be determined as an operating parameter by means of the physics-based battery model. Preferably, the time period for the prediction begins immediately after the acquisition time point of the measurement parameter acquisition.
[0019] The electrode voltage can mean, inter alia, an estimate of the current value. It is also conceivable for the electrode voltage to mean a predicted value for a future time period.
[0020] The battery parameters are used, in particular in a data-driven prediction model, to determine a predicted occurrence time of the occurrence of metal deposition at at least one electrode of the battery device. The occurrence time can be designed, for example, as an absolute date in the future, a remaining operating time until the occurrence of metal deposition, a remaining number of charging cycles or similar. Here, the predicted occurrence time of metal deposition can be determined by the prediction model based on at least the determined battery parameters and the acquired measurement parameters as prediction model input parameters.
[0021] Here, "metal deposition" can mean, in particular, the formation of metal deposition at one of the electrodes of a metal-ion battery. For example, sodium deposition can occur in a sodium-ion battery or calcium deposition can occur in a calcium-ion battery.
[0022] In contrast to the battery parameters, the determination of the prediction parameters takes place by means of a model, in particular a data-driven model. Here, a "data-driven model" can mean, in particular, a model in which the relationship between input quantities and output quantities is determined and / or modeled by means of data.
[0023] In addition, at least one control parameter for controlling the charging process is determined from a control model, in particular a data-driven control model. Here, based on the measurement parameters, the battery parameters and the prediction parameters as control model input parameters, at least a charging current limit is determined as the at least one control parameter.
[0024] The weighting in the sense of the invention can mean, in particular, a combination of the measurement parameters, the battery parameters and the prediction parameters in which a weight is assigned to each parameter. The weight of each parameter can be different from one another. The weight can reflect, for example, the influence of each parameter on the occurrence of metal deposition.
[0025] The battery charging process can thus be controlled in such a way that metal deposition at one of the electrodes of the battery device can be reliably suppressed while at the same time avoiding a slowing down of the charging process. This is based, in particular, on the fact that in the method the battery parameters, which are good indicators of the occurrence of metal deposition, can be determined relatively precisely by means of measurement values and by means of analytical methods. According to the invention, the occurrence time point can be predicted by means of the battery parameters despite the high complexity of the processes occurring inside the battery. This is possible because this complexity is mastered by means of a model, in particular a data-driven model. The determination of the appropriate control parameters in the case of the introduction of the occurrence time point can likewise be carried out reliably and precisely. This is possible despite the high complexity of the relationship between the control operations, such as the charging current limiting, and the resulting changes in the electrical and chemical state inside the battery. According to the invention, the existing complexity is also mastered in this case by means of a model, in particular a data-driven model. As a result, unnecessary charging current limiting is no longer required during the rapid charging phase. A further advantage is that the method can be implemented by means of conventional battery infrastructure.
[0026] Preferably, the at least one control parameter for controlling the charging process, in particular the fast charging process, can be determined at an operating temperature of less than 15°C, 10°C, 5°C, 0°C or -5°C.
[0027] Herein, the "fast charging process" can in particular refer to a charging process with a charging power of more than 50 kW.
[0028] Thereby, a fast charging process can be performed even under low temperature conditions. Herein, there is no need to worry about an increased risk of metal deposition occurring, since the method can provide a correspondingly suitable charging current limiting.
[0029] According to a preferred design, a plurality of control parameters for controlling the charging process can be determined, in particular including a heating control parameter for controlling the heating of the battery device. The control parameters can in particular include a heating control parameter for controlling the preheating of the battery device. The heating control parameter can preferably be set for an internal pulse rate heating of the battery device. Preferably, the heating control parameter can at least include an activation and / or deactivation of an external or internal heating device, a heating activation, a heating deactivation, a preheating time, a preheating amplitude, a charging current frequency, a heating charging current pulse width, a discharging current frequency, a discharging current pulse width, a battery target temperature, an operating temperature for activating the preheating of the battery device or an operating temperature limit for deactivating the preheating of the battery device.
[0030] Thereby, to prevent the occurrence of metal deposition, not only the charging current limit can be adjusted, but also the charging process and / or the battery device can be actively intervened. For this purpose, an internal or external heating device of the battery device can be activated or deactivated. Herein, by determining the point in time at which the occurrence of metal deposition is possible, the heating device can only be switched on when needed. Thereby, energy can be saved and the duration of the charging process can be shortened. In particular, by the present application it is also possible to weigh up whether a generally higher charging current can be achieved after heating, which can compensate for the time loss in the initial preheating phase or pre-warming phase.
[0031] According to a further preferred design, the battery model can be designed as an analytical model, a Kalman filter, a Doyle-Fuller-Newman model and / or a single particle model for determining the state of charge of the battery device. Herein, the "analytical model" can in particular refer to a model in which the relationship between the input quantities and the output quantities is determined and / or modeled by means of a mathematical relationship.
[0032] The above-described configuration of the battery model allows an analytically precise description of the current state and possible future states of the battery device. In particular, further indicators for the occurrence of metal deposition can also be determined more precisely analytically. As examples of such indicators, a gradually decreasing discharge voltage, an increase in electrode resistance, an increase in electrode overpotential or a change in electrolyte polarization can in particular be considered.
[0033] According to a preferred design, either or both of the control model and the prediction model can be based on a machine learning method, preferably on a reinforcement learning method.
[0034] By the above configuration, in particular highly complex or unknown relationships can be represented by data and simulated based on the model, so that an exact determination of the control parameters is enabled.
[0035] According to another preferred design, a charge current limit as one of the control parameters can be determined based on a weighting of the measurement parameters, the battery parameters and the prediction parameters as input parameters of the control model, in particular the weighting of the control model input parameters can be determined in at least two steps by means of machine learning. Thus, as a weighting, an initial weighting can be determined in a learning step. In a reevaluation step, the weighting can be determined during the charging process, wherein in this step the weighting is adjusted based on at least the measurement parameters. Preferably, the weighting can be continuously adjusted in the reevaluation step.
[0036] Alternatively or additionally, the weighting of the prediction model input parameters can be determined in at least two steps by means of machine learning. Thus, as a weighting, an initial weighting can be determined in a learning step and the weighting can be determined during the charging process in a reevaluation step, in which the weighting can be adjusted based on at least the measurement parameters. Preferably, the weighting can be continuously adjusted in the reevaluation step.
[0037] Hereby, the respective, in particular data-driven, models can be initially trained by data filling and the thus determined weightings can be verified. Thus, the method can be provided with high precision and adapted to the respective process or the respective battery model. Since the weightings can be continuously adjusted according to the configuration, a continuous learning of the respective, in particular data-driven, models is enabled. Hereby, changes due to an unavoidable aging process or damage of the battery device can also be taken into account.
[0038] According to a preferred design, the weighting of the control model input parameters can also be determined from a time-varying priority ranking of the input parameters and a time-varying priority ranking of the output control parameters. Preferably, the priority ranking of the control model input parameters can be made to assess a risk of metal deposition occurring. The priority ranking of the output control parameters can preferably be used to assess a controllability of metal deposition occurring by the respective control parameter. Here, the priority ranking can be determined based on a comparison of at least current values of the measurement parameters and / or the battery parameters with relative historical values of the output control parameters, respectively.
[0039] Hereby, the models can not only take into account the relationship of the output values to the input values, but also the importance of the one to the other. Hereby, the control performance can be optimized and determined as a compromise between mutually contradictory requirements as far as possible.
[0040] According to a preferred design, the measurement parameters can further comprise at least a single cell voltage of one or more single cells of the battery device, an occurring charging current or a plurality of locally different operating temperatures as battery voltage.
[0041] By acquiring further physical parameters of the battery device with measurement techniques, the accuracy of the model results and thus of the determined control parameters can be further improved. Also the influence of measurement errors or measurement data outliers can be reduced.
[0042] According to another preferred design, the battery parameters can further comprise at least a state of charge of the battery device, an electrode overpotential or a state of health of the battery device.
[0043] The expected temperature evolution of the operating temperature can preferably comprise a future time course of the operating temperature at a certain section or at different sections of the battery device.
[0044] Thereby further indicators for the occurrence of metal deposition can be determined.
[0045] According to a preferred design, the prediction parameters can further comprise a predicted remaining charging time.
[0046] Thereby further information for controlling the charging process can be determined, which can be used by a user of the control model and / or method.
[0047] According to another preferred design, the control parameters can further comprise at least a start of the charging process, a deactivation of the charging process, a pulse width of the charging current or a duration of the charging current.
[0048] By additionally determining further control parameters, the charging process can additionally be influenced in other ways. Thereby a possibility is provided to suppress the occurrence of metal deposition by measures other than current limiting or heating of the battery device.
[0049] According to a preferred design, the battery model can comprise a battery temperature model for characterizing a time evolution and / or a local spatial distribution of at least one operating temperature in the battery device. Alternatively or additionally, the battery model can comprise a battery state model for characterizing chemical and / or electrical processes in the battery device, preferably numerically. Herein, the battery state model can preferably comprise at least an initial ion concentration of an electrolyte of the battery device, a diffusion rate of ions into a certain electrode, a reaction coefficient of chemical reactions occurring in the battery device or an electrical conductivity of an electrolyte provided in the battery device as input parameters.
[0050] Thereby the accuracy of the analytically determined battery parameters can be improved, so that an improvement of the determination of the control parameters can be achieved in general.
[0051] According to another preferred design, the battery device can be a lithium-ion battery having preferably one or more single cells. Thereby, for example, lithium deposition can occur as metal deposition. The electrode can be an anode. The electrode voltage can be an anode voltage. Preferably, the anode can comprise graphite and / or metal.
[0052] Thereby, a determination of a charge current limit for a lithium-ion battery can be achieved.
[0053] Another aspect of the present application relates to a computer program product comprising instructions which, when the program is executed by a computer, cause the computer to carry out one or more steps of the above-described method.
[0054] Another aspect of the present application relates to a control system for determining a charge current limit of a chargeable battery device. The control system has a measurement module for acquiring measurement parameters of the battery device. Therein, the measurement parameters comprise at least an operating temperature, a battery voltage and a battery current of the battery device. The control system further has a battery state module for determining battery parameters from a physics-based battery model for characterizing physical processes occurring within the battery device. Therein, the battery model has at least the measurement parameters as input parameters. The battery parameters comprise at least an expected temperature evolution of the operating temperature and an electrode voltage. The control system further has a prediction module for determining a prediction parameter of a metal deposition occurrence at a certain electrode of the battery device from a prediction model, in particular a data-driven prediction model. Therein, the prediction model has at least the battery parameters as input parameters. The prediction parameter comprises at least a predicted metal deposition occurrence time. The control system further has a control determination module for determining at least one control parameter for controlling a charging process from a control model, in particular a data-driven control model. Therein, the control model has the measurement parameters, the battery parameters and the prediction parameter as input parameters, wherein the at least one control parameter is based on a weighting of these input parameters, in particular. The at least one control parameter comprises at least the charge current limit. The charge current limit can be determined as a time-varying curve. The control system further has an output module for outputting the determined charge current limit for controlling the charging process based on the at least one control parameter, in particular.
[0055] Preferably, the measurement module can have at least a current sensor, a voltage sensor and / or a temperature sensor.
[0056] The same technical effects and advantages as already described for the above-described method can be achieved by the control system and the computer program product. In particular, a precise determination of control parameters for a fast charging process of a battery device under low temperature conditions can be achieved. A precise control of the charging process can also be achieved.
[0057] According to a preferred design, the measurement module and / or the battery status module can be arranged on the first computing unit. Thereby, for example, the measurement module and / or the battery status module can be arranged on a battery control unit of the battery device. Alternatively or additionally, the prediction module and / or the control determination module can be arranged on the second computing unit. Preferably, the second computing unit can be arranged different from the first computing unit here. Thereby, the prediction module and / or the control determination module can be arranged, for example, on an external server or a cloud server.
[0058] Thereby, for example, the measurement data acquisition and its processing can be implemented directly on the battery control device of the battery module. Thereby, for example, the measurement parameters can be processed. In contrast, in particular data-driven models can be run on an external computer, which can have higher computing power. Thereby, the speed of determining the charge current limit value can be increased. Thereby, for example, the control system can be set up as a real-time-capable system.
[0059] A further aspect of the present application relates to a battery charging system. The battery charging system has a chargeable battery device equipped with at least one single cell. The battery charging system further has a charging terminal for coupling an electrode of the battery device with a charging device to charge the battery device with a charging current. The battery charging system further has the above-described control system. Here, the output charge current limit value is pre-set by the control system as an upper limit for the charging current.
[0060] The same technical effects and advantages as already described for the above-described method, the above-described computer program product and the above-described control system can be achieved by the above-described battery charging system. In particular, a fast charging process of the battery device can be precisely controlled under low-temperature conditions and the risk of metal deposition occurring can be reduced. BRIEF DESCRIPTION OF DRAWINGS
[0061] Further advantages, features and details of the present application result from the following detailed description of embodiments of the present application with reference to the drawings. In which:
[0062] Figure 1 schematically illustrating one embodiment of the method of the present application,
[0063] Figure 2 schematically illustrating one embodiment of the control system of the present application,
[0064] Figure 3 schematically illustrating a further embodiment of the control system of the present application,
[0065] Figure 4 schematically illustrating a further embodiment of the control system of the present application,
[0066] Figure 5 schematically illustrating one embodiment of the battery charging system of the present application. DETAILED DESCRIPTION
[0067] Figure 1 The steps of the inventive method 100 are exemplarily shown from which, for example, the sequence of instructions of the inventive computer program product can also be derived. Figures 2 to 5 Different embodiments of the inventive control system 10 are exemplarily shown, respectively. Figure 5 One embodiment of the inventive battery charging system 90 is shown.
[0068] Figure 1 The method 100 in the battery charging system 90 is designed for determining a charging current limit value influencing the charging process of the chargeable battery device 1000 and for this has a series of steps which can be iteratively executed during the charging process. Preferably, these steps can also be repeatedly and continuously executed. Thus, the value of the charging current limit value can change or remain unchanged in the next iteration.
[0069] For this, a measurement step S20 is executed in which a measurement parameter MP is determined on the battery device 1000. The measurement parameter MP comprises at least the operating temperature of the battery device 1000, the battery voltage and the battery current.
[0070] Further, a battery state determination step S30 is executed in which a battery parameter BP is determined using a physics-based battery model. Here, the physics-based battery model is designed to characterize the physical processes occurring within the battery device 1000. The battery parameter BP comprises the expected evolution of the operating temperature of the battery device 1000 and the electrode voltage.
[0071] In a prediction step S40, a prediction parameter VP is determined using a prediction model, in particular a data-driven prediction model, of the occurrence of metal deposition at the electrodes 1001, 1002 of the battery device 1000. Here, the prediction parameter VP comprises at least a predicted occurrence time of metal deposition at the electrodes of the battery device 1000 and is determined based on at least the battery parameter BP as a prediction model input parameter.
[0072] In a control determination step S50, a control parameter KP is determined using a control model, in particular a data-driven control model, required for controlling the charging process. The control parameter KP comprises at least the charging current limit value. The control parameter is determined by the control model based on a weighting of the measurement parameter MP, the battery parameter BP and the prediction parameter VP as control model input parameters.
[0073] In an output step S60, the charging current limit value determined in the manner described above is output for controlling the charging process.
[0074] In contrast to the battery model, the prediction model and the control model are, in particular, data-driven models, but not analytical models. The prediction model can be designed, for example, as an artificial neural network, a Markov model, a logistic regression or an algorithm based on decision trees. An input layer can be provided here, in which individual neurons represent input features. An intermediate layer and an output layer with output features can also be provided. The individual layers can be connected to one another by weight and input matrices.
[0075] The input features of the prediction model can be, for example, the anode overvoltage, the state of health of the battery device 1000, the cell voltage or the temperature corresponding to the individual zones of the battery cell 1003. The output features of the prediction model can be the estimated charging duration or the predicted time to the occurrence of a metal deposition event. Thus, the prediction model can output, for example, the following occurrence time prediction: a metal deposition is expected to occur within the next 30 seconds.
[0076] The control model can preferably simulate a controller for the operating temperature and the charging current of the battery device 1000. From the control model, an appropriate heating configuration can thus be determined, for example, for the warm-up phase of the battery device 1000. From the control model, it is thus possible to determine, for example, the charge-discharge pulses to be set and their pulse frequency.
[0077] Preferably, the control model can be based on reinforcement learning. The constraints that arise during charging can be defined, for example, as a Markov decision process. In this case, the components of the Markov decision process include the environment, the environment state and the executable actions from which a selection can be made. In the present embodiment, the warm-up and the short charging time of the battery device 1000 can be defined, for example, as components of the Markov environment. The charging time and the prediction of the occurrence of metal deposition can be defined as the environment state. The warm-up regulation, the charge-discharge pulse configuration and the charging current limit can be defined as executable actions. In this case, the actions can be, for example, to keep the last applied value unchanged, to reduce it or to increase it. Preferably, the control model is designed in such a way that the selection of the actions is not based on randomness, so that the selection is deterministic.
[0078] In the training steps S41, S51, the weights or other parameters of the, in particular, data-driven models can be determined, for example, by a sufficient number of training data sets. The weights or parameters determined and trained in this way can be verified by a validation data set. Preferably, the training and validation data sets come from a real experimental environment. When the accuracy of the weights or other parameters no longer improves significantly, the training step can be considered complete. Alternatively, the training step can be considered complete when the absolute error between the validation data set and the data set calculated by the, in particular, data-driven model is sufficiently reduced.
[0079] The continuous learning and dynamic adjustment of the weights can be implemented, for example, in the reevaluation steps S42, S52. Here, historical data and parameters are analyzed, for example, and compared to the output values of the prediction model or the control model.
[0080] Figure 2 An example of a control system 10 is shown. The control system 10 is suitable and designed for controlling a fast charging process of a rechargeable battery device, such as the above-mentioned battery device 1000.
[0081] The battery device 1000 can be, inter alia, a lithium-ion battery. The battery device 1000 is exemplarily shown in Figure 4 .
[0082] The battery device 1000 can have one or more battery cells 1003. The voltage potential of the battery cells 1003 can be reversibly adjustable. The individual battery cells 1003 can be connected together as a voltage source and / or a current source. The battery device 1000 can have two electrodes 1001, 1002 for power output or power input. It is further conceivable that the battery device 1000 has further electrodes for potential and current measurement. It is further conceivable that the battery device 1000 can be a component of the control system 10.
[0083] As shown in Figure 2 , the control system 10 has a measurement module 20 for acquiring measurement parameters MP of the battery device 1000. To this end, the measurement module 20 or the battery device 1000 can have sensors suitable for acquiring the measurement parameters. Further details on the sensors will be explained in the following description of Figure 4 .
[0084] As shown in Figure 2 , the control system 10 further has a battery state module 30 for determining battery parameters BP. To this end, the battery state module 30 has a physics-based battery model. By means of the physics-based battery model, for example, physical processes occurring within the battery device 1000 can be formulated and numerically expressed. It is clear from Figure 2 that the battery model has at least the measurement parameters MP as input parameters and has the battery parameters BP as output parameters.
[0085] Both the battery state module 30 and the measurement module 20 can be arranged on a first computing unit 11, respectively. The first computing unit 11 can be, for example, a battery control unit.
[0086] As shown in Figure 2 , the control system 10 further has a prediction module 40 for determining, from an inter alia data-driven prediction model, prediction parameters VP of the occurrence of metal deposition at the electrodes 1001, 1002 of the battery device 1000. It is clear from Figure 2 that the prediction model has, in addition to the battery parameters BP, the measurement parameters MP as input parameters. It is further conceivable that only the measurement parameters MP are used as prediction model input parameters.
[0087] As Figure 2 As further shown, the control system 10 has a control determination module 50 for determining control parameters KP for controlling the charging process by means of a data-driven control model, in particular. Figure 2 By way of example, the control model has as input parameters a measurement parameter MP, a battery parameter BP and a prediction parameter VP. The control determination module 50 weights these input parameters in order to determine at least one control parameter KP.
[0088] As Figure 2 As further shown, the control system 10 has an output module 60 for outputting at least one control parameter KP in order to control the charging process on the basis of the at least one control parameter KP. The output module 60 can be designed as a data interface or data bus for transmitting data to other devices outside the control system. Such other devices can be, for example, the battery device 1000 or an external battery charging device.
[0089] The prediction module 40, the control determination module 50 and / or the output module 60 can be arranged on the second computing unit 12, respectively. Preferably, the second computing unit 12 can be a cloud server.
[0090] Figure 3 A preferred design of the control system 10 is shown. In this case, Figure 2 The control system 10 is shown in a similar configuration.
[0091] Figure 3 A preferred design of the battery status module 30 is disclosed. Herein, Figure 3 The battery status module 30 is shown exemplarily as having a battery temperature model 31 and a battery status model 32 as battery models.
[0092] By means of the battery temperature model 31, the expected temporal evolution of the temperature of the battery device 1000 can be determined as a temperature parameter TP. Furthermore, the local temperature distribution of the battery device 1000 can be determined in a resolved manner as a further component of the temperature parameter TP.
[0093] By means of the battery status model 32, chemical and / or electrical processes within the battery device 1000 can be simulated and output as a battery status parameter BSP. Herein, the battery status model 32 can take into account further parameters in addition to the measurement parameter MP, for example the initial ion concentration or the ion diffusion rate of the battery device 1000.
[0094] The battery status parameter BSP and the temperature parameter TP can be output from the battery status module 30 as a battery parameter BP by means of a signal connector 34.
[0095] Figure 4 A preferred design of the control system 10 is shown. In this case, Figure 2 and Figure 3The same control system 10. The same implementation features are not described again below.
[0096] However, Figure 4 A preferred design of the measurement module 20 is disclosed.
[0097] Figure 4 The control system 10 is shown with a measurement module 20, which has a plurality of sensors for acquiring measurement parameters MP of the battery device 1000. It is also conceivable that the battery device 1000 has such sensors, in particular. Figure 4 The cell voltage of each battery cell 1003 is shown by way of example to be determined by a voltage sensor 21. The local temperature distribution of the operating temperature of the battery device 1000 can be acquired by means of a plurality of temperature sensors 22 arranged in a distributed manner. The battery current output from the battery device 1000 or input for charging can be acquired by the measurement module 20 by means of a current sensor 23.
[0098] The battery device 1000, the sensors 21, 22, 23 and the first computing unit 11 can be provided as an integral unit and in this way, for example, constitute a battery module 1100, which can have the measurement module 20 and the battery state module 30.
[0099] Figure 4 A preferred design is also disclosed in which the prediction module 40, the control determination module 50 and the output module 60 are completely contained in the second computing unit 12. Accordingly, the prediction module 40, the control determination module 50 and the output module 60 are not shown separately in Figure 4 the figures.
[0100] Figure 5 A battery charging system 90 is shown, which has a battery device 1000 comprising at least one battery cell 1003 described above and a control system 10. Figure 5 A charging device 1200 is also shown by way of example, which is connected to the electrodes of the battery device 1000 via electrical leads through the charging terminals of the battery charging system 90 in order to charge the battery device 1000 using a charging current. The charging device 1200 can be an integral part of the battery charging system 90. The charging current limit determined and output by the control system 10 is set by the control system 10 as at least one control parameter KP as an upper limit for the charging current of the battery device 1000 or the charging device 1200.
[0101] The description of the embodiments above describes the application only in example ranges. It is obvious that the individual features of the embodiments can be freely combined with one another as far as this is technically reasonable without departing from the scope of the application.
[0102] List of reference signs
[0103] 10 control system
[0104] 11 first computing unit, battery control unit
[0105] 12 second computing unit, cloud server
[0106] 20 measurement module
[0107] 21 voltage sensor
[0108] 22 temperature sensor
[0109] 23 current sensor
[0110] 30 battery state module
[0111] 31 battery temperature model
[0112] 32 battery state model
[0113] 34 signal connector
[0114] 40 prediction module
[0115] 50 control determination module
[0116] 60 output module
[0117] 90 battery charging system
[0118] 100 method
[0119] 1000 battery device
[0120] 1001, 1002 electrode
[0121] 1003 battery cell
[0122] 1100 battery module
[0123] 1200 charging device
[0124] KP control parameter
[0125] MP measurement parameter
[0126] BP battery parameter
[0127] TP temperature parameter
[0128] BSP battery state parameter
[0129] VP prediction parameter
[0130] S20 measurement step
[0131] S30 battery state determination step
[0132] S40 prediction step
[0133] S50 control determination step
[0134] S60 output step
[0135] S41, S51 learning step
[0136] S42, S52 reevaluation step
Claims
1. A method (100) for determining a charging current limit during the charging process of a rechargeable battery device (1000), comprising the following steps: The measurement parameters (MP) of the battery device (1000) are collected, wherein the measurement parameters (MP) include at least the operating temperature, battery voltage and battery current of the battery device (1000); Its characteristics are, Battery parameters (BP) are determined from a physics-based battery model used to characterize the physical processes occurring within the battery device (1000), wherein at least the expected temperature evolution of the operating temperature and the electrode voltage are determined as battery parameters (BP) based on the acquired measurement parameters (MP) as input parameters to the battery model. The prediction parameter (VP) for metal deposition at the electrodes (1001, 1002) of the battery device (1000) is determined from the prediction model, wherein the prediction parameter (VP) is determined based at least on the determined battery parameters (BP) as input parameters of the prediction model. At least one control parameter (KP) for controlling the charging process is determined from the control model, wherein the charging current limit is determined as at least one control parameter (KP) based on the measured parameter (MP), the battery parameter (BP), and the predicted parameter (VP) as input parameters to the control model; and Output a defined charging current limit.
2. The method (100) according to claim 1, wherein, The predicted timing of the metal deposition is determined by a prediction model based at least on determined cell parameters (BP) and acquired measurement parameters (MP) as input parameters.
3. The method (100) according to claim 1 or 2, wherein, The charging current limit was determined as a time-varying curve.
4. The method (100) according to any one of the preceding claims, wherein, The at least one control parameter (KP) for controlling the charging process, especially the fast charging process, is determined at operating temperatures below 15°C, 10°C, 5°C, 0°C, or -5°C.
5. The method (100) according to any one of the preceding claims, wherein, Several control parameters (KP) are determined for controlling the charging process, and in particular these control parameters (KP) also include heating control parameters for controlling the heating, preferably preheating, of the battery device (1000), preferably by means of internal pulse rate heating.
6. The method (100) according to claim 5, wherein, The heating control parameters include at least: starting the external or internal heating device, stopping the external or internal heating device, starting heating, stopping heating, preheating time, preheating amplitude, charging current frequency, heating charging current pulse width, discharging current frequency, discharging current pulse width, battery target temperature, operating temperature limit for starting the preheating of the battery device (1000), or operating temperature limit for stopping the preheating of the battery device (1000).
7. The method (100) according to any one of the preceding claims, wherein, The control model and / or the prediction model are based on machine learning methods, preferably reinforcement learning methods, and / or have artificial neural networks, and / or wherein the battery model is configured with at least one analytical model, Kalman filter, Doyle-Fuller-Newman model or single-particle model for determining the electrochemical state of the battery device (1000).
8. The method (100) according to any one of the preceding claims, wherein, Based on a weighted average of the measured parameter (MP), the battery parameter (BP), and the predicted parameter (VP) as input parameters to the control model, the charging current limit, as one of the control parameters (KP), is determined, and in particular, by means of machine learning, the weighting of the control model input parameters and / or another weighting of the predicted model input parameters is determined in at least two steps, the at least two steps including: Learning steps (S41, S51), in which the initial weighting is determined as the weighting, and The re-evaluation step (S42, S52) during the charging process, in which the weighting is preferably continuously adjusted based at least on the measurement parameter (MP).
9. The method (100) according to any one of the preceding claims, wherein, The weighting of the input parameters of the control model is determined from the following aspects: The priority ranking of the time-varying input parameters of the control model is preferably used to assess the risk of metal deposition. The priority ranking of the output control parameters (KP) over time is preferably used to evaluate the resistance to metal deposition through each control parameter (KP). The priority ranking is preferably determined by comparing the current values of at least the measured parameter (MP) and / or the battery parameter (BP) with the relative historical values of the output control parameter (KP).
10. The method (100) according to any one of the preceding claims, wherein, The measured parameters (MP) also include at least the single-cell voltage of one or more individual cells of the battery device (1000) as the battery voltage, the charging current, or multiple locally different operating temperatures, and / or The battery parameters (BP) further include at least the electrode overpotential, the health status of the battery device (1000), or the state of charge of the battery device (1000), wherein the expected temperature evolution process preferably includes the future time variation curve of the operating temperature of the battery device (1000) in a certain segment or different segments, and / or The prediction parameter (VP) also includes the predicted remaining charging time, and / or The control parameters (KP) also include at least the initiation of the charging process, the deactivation of the charging process, the charging current pulse width, or the charging current duration.
11. The method (100) according to any one of the preceding claims, wherein, The battery model includes: A battery temperature model (31) is used to characterize the time evolution and / or local distribution of at least one operating temperature of the battery device (1000), and A battery state model (32) is used to numerically characterize the chemical and / or electrical processes in the battery device (1000), wherein the battery state model (32) preferably further includes at least the conductivity of the electrolyte provided in the battery device (1000), the initial ion concentration of 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 the chemical reaction occurring in the battery device (1000) as input parameters.
12. The method (100) according to any one of the preceding claims, wherein, The battery device (1000) is a lithium-ion battery having preferably one or more single cells (1003), wherein the metal deposition is a lithium deposition, and wherein the electrodes (1001, 1002) are anodes, and the electrode voltage is an anode voltage, wherein the anode preferably comprises graphite and / or metal.
13. A computer program product comprising instructions that, when executed by a computer, cause the computer to perform the steps of the method (100) according to any one of claims 1 to 12.
14. A control system (10) for determining a charging current limit value for a rechargeable battery device (1000), comprising: A measurement module (20) for acquiring measurement parameters (MP) on the battery device (1000), wherein, The measurement parameters (MP) include at least the operating temperature, battery voltage, and battery current of the battery device (1000); Its characteristic is that it also has A battery state module (30) is used to determine battery parameters (BP) from a physics-based battery model for characterizing the physical processes occurring within the battery device (1000), wherein the battery model has at least the measured parameters (MP) as input parameters, and wherein the battery parameters (BP) include at least the expected temperature evolution of the operating temperature and the electrode voltage. The prediction module (40) is used to determine from the prediction model a prediction parameter (VP) for the occurrence of metal deposition at the electrodes (1001, 1002) of the battery device (1000), wherein the prediction model has at least the battery parameter (BP) as an input parameter, and wherein the prediction parameter (VP) includes at least the predicted time of metal deposition. A control determination module (50) is configured to determine at least one control parameter (KP) for controlling the charging process from a control model, wherein the control model has the measured parameter (MP), the battery parameter (BP), and the predicted parameter (VP) as input parameters, wherein the at least one control parameter (KP) is based, in particular, on a weighted average of these input parameters, and wherein the at least one control parameter (KP) includes at least a charging current limit; and The output module (60) is used to output the determined charging current limit so as to control the charging process in particular based on the at least one control parameter (KP).
15. The control system (10) according to claim 14. in, The measurement module (20) and the battery status module (30) are mounted on the first computing unit (11), preferably the battery control unit of the battery device (1000), and The prediction module (40) and the control determination module (50) are located in the second computing unit (12), preferably an external server or a cloud server, wherein the second computing unit (12) is preferably 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 a voltage sensor (21), a temperature sensor (22), and / or a current sensor (23).
17. A battery charging system (90) comprising: A rechargeable battery device (1000) equipped with at least one battery cell (1003). Charging terminals are used to couple the electrodes (1001, 1002) of the battery device (1000) to the charging device (1200) so as to charge the battery device (1000) by means of charging current. Its characteristics are, It also includes a control system (10) according to any one of claims 14 to 16. The output charging current limit, which serves as the upper limit of the charging current, is preset by the control system (10).
Citation Information
Patent Citations
Charging strategies to mitigate lithium plating in electrified vehicle battery
US20170203667A1