System and method for real-time estimation of rechargeable battery capacity
Through the battery control module and Kalman filter technology, the problem of difficulty in accurately estimating the real-time capacity of rechargeable batteries in the prior art is solved, and accurate monitoring and management of the real-time status of the battery is achieved, and battery life and vehicle efficiency are improved.
Patent Information
- Application Number
- CN202110163423.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2016-02-19
- Filing Date
- 2016-08-02
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2036-08-02
AI Technical Summary
The prior art is difficult to accurately estimate the real-time capacity and state of charge of rechargeable batteries, especially in vehicle applications, where traditional methods such as complete discharge measurements are impractical.
Using a battery control module, by monitoring the monitoring parameters of the battery, using two linear regression models and a Kalman filter, the real-time capacity of the battery is recursively calculated based on the equivalent circuit model, and the capacity is calculated based on the relaxation open circuit voltage and Coulomb count.
Accurate estimation of the real-time capacity and charging status of rechargeable batteries is achieved, real-time and accuracy of battery management is improved, battery life is extended, and fuel efficiency of hybrid electric vehicles is improved.
Smart Images

Figure CN113030735B_ABST
Abstract
Description
[0001] This application is a divisional application of an invention patent application with an application date of August 2, 2016, an international application number of PCT / US2016 / 045084, a national application number of 201680082627.3, and an invention name of “System and method for real-time estimation of rechargeable battery capacity”. Technical Field
[0002] The present disclosure relates generally to the field of batteries and battery modules. In particular, the present disclosure relates to estimating the real-time capacity of a rechargeable battery. Background Art
[0003] This section is intended to introduce the reader to various aspects of the art that may be related to the various aspects of the present disclosure described below. It is believed that this discussion helps to provide background information for the reader to better understand the various aspects of the present invention. Therefore, it should be understood that these statements should be read from this perspective, rather than as an admission of the prior art.
[0004] A vehicle that uses one or more battery systems to provide all or part of the power for the vehicle can be referred to as an xEV, wherein the term "xEV" is defined herein to include all of the following vehicles or any variant or combination thereof: it uses electricity as all or part of its vehicle power. For example, xEV includes an electric vehicle (EV) that uses electricity as all of its power. As those skilled in the art will appreciate, a hybrid electric vehicle (HEV), which is also considered an xEV, combines an internal combustion engine propulsion system with a battery-powered electric propulsion system (e.g., a 48 volt (V) or 130V system). The term HEV may include any variant of a hybrid electric vehicle. For example, a full hybrid system (FHEV) may use one or more electric motors, only an internal combustion engine, or both to provide power and other electricity to the vehicle. In contrast, a mild hybrid system (HMEV) deactivates the internal combustion engine when the vehicle is idling, and uses a battery system to continue to power an air conditioning unit, a radio, or other electronic equipment, and restart the engine when propulsion is required. A mild hybrid system may also apply a certain degree of power assistance, such as during acceleration, to supplement the internal combustion engine. Mild hybrids are typically between 96V and 130V, and recover braking energy through a belt or crank integrated starter generator. In addition, micro hybrid electric vehicles (mHEVs) also use a "start-stop" system similar to mild hybrids, but the micro hybrid system can provide power assistance or no power assistance to the internal combustion engine and operate at voltages below 60V. For the purposes of this discussion, it should be recognized that mHEVs generally do not technically use electricity directly provided to the crankshaft or drivetrain as power for any part of the vehicle, but mHEVs can still be considered xEVs because they do use electricity to supplement the vehicle's power needs when the vehicle is idling (where the internal combustion engine is disabled) and recover braking energy through an integrated starter generator. In addition, a plug-in electric vehicle (PEV) is any vehicle that can be charged from an external power source, such as a wall socket, and the energy stored in the rechargeable battery pack drives or helps drive the wheels. PEV is a subcategory of EV that includes all-electric or battery electric vehicles (BEV), plug-in hybrid electric vehicles (PHEV), and retrofit electric vehicles of hybrid electric vehicles and conventional internal combustion engine vehicles.
[0005] Such xEVs may offer several advantages over more traditional gas-powered vehicles that use only an internal combustion engine and a traditional electrical system (typically a 12V system powered by a lead-acid battery.) For example, xEVs may produce fewer undesirable emissions products and may exhibit greater fuel efficiency than traditional internal combustion vehicles, and in some cases, such xEVs may not use gasoline at all, as some types of EVs or PEVs do.
[0006] As technology continues to evolve, there is a need to provide improved status indicators for battery modules of such vehicles. For example, the power used by an xEV may be provided by a rechargeable battery. While the rechargeable battery is operating, it may be difficult to accurately depict the state of charge or capacity of the rechargeable battery. The present disclosure generally relates to estimating real-time parameters of the rechargeable battery during operation of the rechargeable battery and / or xEV. Summary of the invention
[0007] The following summarizes the specific embodiments disclosed herein. It should be understood that these aspects only provide the reader with an overview of these specific embodiments, and these aspects are not intended to limit the scope of the present disclosure. In fact, the present disclosure can cover a number of aspects not described below.
[0008] The present disclosure relates to a battery system, comprising a battery coupled to an electrical system. The battery system further comprises a battery control module electrically coupled to the battery. The battery control module monitors at least one monitoring parameter of the battery, and the battery control module recursively calculates a first capacity estimate of the battery using two linear regression models. The two linear regression models are based on at least one equivalent circuit model, at least one monitoring parameter, and a Kalman filter.
[0009] The present disclosure also relates to a method for determining a real-time capacity of a rechargeable battery coupled to an electrical system. The method includes monitoring at least one monitoring parameter of the rechargeable battery during battery operation using a battery control module through one or more sensors coupled to the rechargeable battery. The method also includes recursively calculating a first real-time capacity of the rechargeable battery through the battery control module using two linear regression models. The two linear regression models are based on at least one equivalent circuit model, at least one monitoring parameter, and a Kalman filter.
[0010] The present disclosure also relates to an energy storage component for a vehicle. The energy storage component includes a housing, a first terminal and a second terminal, and a rechargeable battery disposed in the housing. The rechargeable battery is coupled to the first terminal and the second terminal. The energy storage component also includes a battery control module that monitors at least one monitoring parameter of the energy storage component. The battery control module also recursively calculates a first capacity estimate of the energy storage component using two linear regression models based on at least one equivalent circuit model, at least one monitoring parameter, and a Kalman filter. In addition, the battery control module recursively calculates a second capacity estimate of the energy storage component based at least in part on two relaxed open circuit voltage measurements of the energy storage component and a coulomb count of the energy storage component. BRIEF DESCRIPTION OF THE DRAWINGS
[0011] Various aspects of the present disclosure may be better understood by reading the following detailed description and referring to the accompanying drawings, in which:
[0012] Figure 1 is a perspective view of a vehicle (xEV) having a battery system that contributes to all or part of the vehicle's electrical power according to one embodiment of the present method;
[0013] Figure 2 A hybrid electric vehicle (HEV) according to one embodiment of the method Figure 1 Schematic cross-section of the xEV;
[0014] Figure 3 According to one embodiment of the method Figure 1 Schematic diagram of the xEV battery system;
[0015] Figure 4 According to one embodiment of the method Figure 1 1-RC equivalent circuit model of the energy storage component of the xEV;
[0016] Figure 5 According to one embodiment of the method Figure 1 A graph showing the relationship between the open circuit voltage (OCV) and state of charge (SOC) of the energy storage components of an xEV;
[0017] Fig. 6A and 6B is a process flow chart describing a method for calculating energy storage component parameters and determining parameter convergence of an energy storage component according to an embodiment of the present method;
[0018] Fig. 7A and 7B is a circular buffer for storing data according to one embodiment of the method, the data being stored using Fig. 6A and 6B The process flow chart is calculated;
[0019] Fig. 8A and 8B is a process flow diagram describing a method for calculating the capacity of an energy storage component using two linear regression models according to one embodiment of the present method;
[0020] Fig. 9A and 9B is a process flow diagram describing a method for calculating the capacity of an energy storage component using two relaxed open circuit voltage measurements and coulomb counting according to one embodiment of the present method;
[0021] Fig.10 is a process flow diagram describing a method for directional verification of a capacity estimate of an energy storage component according to an embodiment; and
[0022] Fig.11 According to one embodiment, Fig.10A diagram of an embodiment of a process flow diagram. DETAILED DESCRIPTION
[0023] One or more specific embodiments are described below. In order to provide a brief description of these embodiments, not all features of the actual implementation are described in this specification. It should be understood that in the development of any such actual implementation, as in any engineering or design project, many implementation-specific decisions must be made to achieve the developer's specific goals, such as meeting system-related and business-related constraints, which may vary from implementation to implementation. In addition, it should be understood that such development work may be complex and time consuming, but it should nevertheless be a routine task in design, fabrication, and manufacturing for those of ordinary skill in the art who benefit from the present invention.
[0024] The battery system described herein can be used to provide power to various types of electric vehicles (xEVs) and other high voltage energy storage / consumption applications (e.g., grid power storage systems). Such a battery system may include one or more battery modules, each having a plurality of battery cells (e.g., lithium ion (Li-ion) electrochemical battery cells) arranged and electrically interconnected to provide a specific voltage and / or current that is beneficial for powering one or more components of, for example, an xEV. As another example, a battery module according to the present embodiment can be combined with or provide power to a fixed power system (e.g., a non-automotive system).
[0025] Manufacturers that typically produce conventional gas-powered vehicles may wish to use improved vehicle technologies (e.g., regenerative braking technologies) within their vehicle lines based on the advantages over conventional gas-powered vehicles. Typically, these manufacturers may utilize one of their conventional vehicle platforms as a starting point. Since conventional gas-powered vehicles are designed to utilize a 12-volt battery system, a 12-volt lithium-ion battery may be used to supplement the 12-volt lead-acid battery. More specifically, the 12-volt lithium-ion battery may be used to more efficiently capture the electrical energy generated during regenerative braking and subsequently supply the electrical energy to power the vehicle's electrical systems.
[0026] As vehicle technology advances, high voltage electrical devices may also be included in the electrical system of the vehicle. For example, a lithium-ion battery may supply electrical energy to an electric motor in a mild hybrid electric vehicle. Typically, these high voltage electrical devices use voltages greater than 12 volts, such as up to 48 volts. Therefore, in some embodiments, a DC-DC converter may be used to boost the output voltage of a 12 volt lithium-ion battery to power the high voltage device. Additionally or alternatively, a 48 volt lithium-ion battery may be used to supplement a 12 volt lead-acid battery. More specifically, a 48 volt lithium-ion battery may be used to more efficiently capture the electrical energy generated during regenerative braking and subsequently provide electrical energy to power the high voltage device.
[0027] Therefore, the design choice of whether to use a 12 volt lithium-ion battery or a 48 volt lithium-ion battery may depend directly on the electrical devices included in a particular vehicle. However, although the voltage characteristics may be different, the operating principles of a 12 volt lithium-ion battery and a 48 volt lithium-ion battery are generally similar. More specifically, as described above, both can be used to capture electrical energy during regenerative braking and subsequently supply electrical energy to power electrical devices in the vehicle.
[0028] Therefore, in order to simplify the following discussion, the present technology will be described in conjunction with a battery system having a 12V lithium-ion battery and a 12V lead-acid battery. However, a person of ordinary skill in the art will be able to adapt the present technology to other battery systems, such as a battery system having a 48V lithium-ion battery and a 12V lead-acid battery.
[0029] The present disclosure relates to batteries and battery modules. More specifically, the present disclosure relates to estimating real-time parameters of rechargeable batteries. Specific embodiments relate to lithium-ion battery cells that can be used in vehicle environments (e.g., hybrid electric vehicles) and other energy storage / consumption applications (e.g., energy storage for power grids).
[0030] In view of the foregoing, the present disclosure describes techniques for estimating real-time parameters of a rechargeable battery. Traditionally, the rated capacity of a rechargeable battery can be determined by completely discharging the rechargeable battery from a fully charged state using a constant discharge rate at room temperature. This approach may be impractical for rechargeable batteries used in a vehicle environment. Therefore, a partial discharge of the rechargeable battery can be used to estimate the capacity of the rechargeable battery. Therefore, the battery control unit described in the present disclosure can estimate the capacity and other parameters of the rechargeable battery in real time or near real time using the systems and methods described in detail below.
[0031] To help illustrate, Figure 1 is a perspective view of one embodiment of a vehicle 10 that may utilize a regenerative braking system. Although the following discussion relates to vehicles having regenerative braking systems, the techniques described herein may be applicable to other vehicles that utilize batteries to capture / store electrical energy, which may include electric and gas powered vehicles.
[0032] As described above, it is expected that the battery system 12 is largely compatible with conventional vehicle designs. Therefore, the battery system 12 can be arranged in the vehicle 10 where a conventional battery system would be accommodated. For example, as shown, the vehicle 10 may include a battery system 12 in a location similar to the location of a lead-acid battery of a typical internal combustion engine vehicle (e.g., under the hood of the vehicle 10). In addition, as will be described in more detail below, the battery system 12 can be positioned to facilitate management of the temperature of the battery system 12. For example, in some embodiments, positioning the battery system 12 under the hood of the vehicle 10 can enable air ducts to enable air flow to flow over the battery system 12 and cool the battery system 12.
[0033] Figure 2 1. As shown, the battery system 12 includes an energy storage component 14 that is coupled to an ignition system 16, an alternator 18, a vehicle console 20, and optionally to an electric motor 22. In general, the energy storage component 14 can capture / store electrical energy generated in the vehicle 10 and output electrical energy to power electrical devices in the vehicle 10.
[0034] In other words, the battery system 12 can provide power to components of the vehicle electrical system, which can include a radiator cooling fan, a climate control system, an electric power steering system, an active suspension system, an automatic parking system, an electric oil pump, an electric supercharger / turbocharger, an electric water pump, a heated windshield / defogger, a window lift motor, a vanity light, a tire pressure monitoring system, a sunroof motor controller, power seats, an alarm system, an infotainment system, a navigation feature, a lane departure warning system, an electric parking brake, exterior lights, or any combination thereof. Illustratively, in the illustrated embodiment, the energy storage component 14 powers a vehicle console 20, a display 21 within the vehicle, and an ignition system 16, which can be used to start (e.g., via cranking) an internal combustion engine 24.
[0035] In addition, the energy storage component 14 can capture electrical energy generated by the AC generator 18 and / or the motor 22. In some embodiments, the AC generator 18 can generate electrical energy when the internal combustion engine 24 is running. Specifically, the AC generator 18 can convert the mechanical energy generated by the rotation of the internal combustion engine 24 into electrical energy. In addition or alternatively, when the vehicle 10 includes the motor 22, the motor 22 can generate electrical energy by converting the mechanical energy generated by the movement of the vehicle 10 (e.g., wheel rotation) into electrical energy. Therefore, in some embodiments, the energy storage component 14 can capture the electrical energy generated by the AC generator 18 and / or the motor 22 during regenerative braking. Therefore, the AC generator 18 and / or the motor 22 are collectively referred to as a regenerative braking system in this article.
[0036] To facilitate the capture and supply of electrical energy, the energy storage component 14 may be electrically coupled to the vehicle's electrical system via a bus 26. For example, the bus 26 may enable the energy storage component 14 to receive electrical energy generated by the alternator 18 and / or the motor 22. Additionally, the bus 26 may enable the energy storage component 14 to output electrical energy to the ignition system 16 and / or the vehicle console 20. Thus, when a 12 volt battery system 12 is used, the bus 26 may carry electrical power typically between 8-18 volts.
[0037] In addition, as shown, the energy storage component 14 may include a plurality of battery modules. For example, in the illustrated embodiment, the energy storage component 14 includes a lead-acid (e.g., first) battery module 28 and a lithium-ion (e.g., second) battery module 30 according to the present embodiment, wherein each battery module 28, 30 includes one or more battery cells. In other embodiments, the energy storage component 14 may include any number of battery modules. In addition, although the first battery module 28 and the second battery module 30 are shown as being adjacent to each other, they may be positioned in different areas around the vehicle. For example, the second battery module 30 may be positioned inside or around the vehicle 10, while the first battery module 28 may be positioned under the hood of the vehicle 10.
[0038] In some embodiments, the energy storage component 14 may include multiple battery modules to utilize a variety of different battery chemistries. For example, the first battery module 28 may use a lead-acid battery chemistry and the second battery module 30 may use a lithium-ion battery chemistry. In such embodiments, the performance of the battery system 12 may be improved because lithium-ion battery chemistry generally has a higher coulombic efficiency and / or a higher power charging acceptance rate (e.g., a higher maximum charging current or charging voltage) than lead-acid battery chemistry. In this way, the capture, storage, and / or distribution efficiency of the battery system 12 may be improved.
[0039] To facilitate control of the capture and storage of electrical energy, the battery system 12 may additionally include a control module 32. More specifically, the control module 32 may control the operation of components in the battery system 12, such as the operation of relays (e.g., switches) within the energy storage component 14, the alternator 18, and / or the motor 22. For example, the control module 32 may adjust the amount of electrical energy captured / provided by each battery module 28 or 30 (e.g., de-rating and re-rating the battery system 12), perform load balancing between the battery modules 28 and 30, determine the state of charge of each battery module 28 or 30, determine the temperature of each battery module 28 or 30, determine a predicted temperature trajectory for either battery module 28 or 30, determine a predicted lifespan of the battery module 28 or 30, determine the fuel economy contribution of either battery module 28 or 30, control the magnitude of the voltage or current output by the alternator 18 and / or the motor 22, and the like.
[0040] Thus, the control module (e.g., unit) 32 may include one or more processors 34 and one or more memories 36. More specifically, the one or more processors 34 may include one or more application specific integrated circuits (ASICs), one or more field programmable gate arrays (FPGAs), one or more general purpose processors, or any combination thereof. In general, the processor 34 may execute computer readable instructions related to the processes described herein. Additionally, the processor 34 may be a fixed point processor or a floating point processor.
[0041] In addition, one or more memories 36 may include volatile memory, such as random access memory (RAM), and / or non-volatile memory, such as read-only memory (ROM), an optical drive, a hard disk drive, or a solid-state drive. In some embodiments, the control module 32 may include some parts of a vehicle control unit (VCU) and / or a separate battery control module. In addition, as shown, the control module 32 may be separated from the energy storage component 14, such as as a stand-alone module. In other embodiments, the battery management system 36 may be included in the energy storage component 14.
[0042] In certain embodiments, the control module 32 or the processor 34 may receive data from various sensors 38 disposed within and / or around the energy storage component 14. The sensors 38 may include various sensors for measuring current, voltage, temperature, etc., of the battery module 28 or 30. After receiving the data from the sensors 38, the processor 34 may convert the raw data into parameter estimates of the battery modules 28 and 30. In this way, the processor 34 may present the raw data as data that may provide valuable information about the operation of the battery system 12 to the operator of the vehicle 10, and may display information about the operation of the battery system 12 on the display 21. The display 21 may display various images generated by the device 10, such as a GUI or image data (including still images and video data) for operating the system. The display 21 may be any suitable type of display, such as a liquid crystal display (LCD), a plasma display, or an organic light emitting diode (OLED) display. In addition, the display 21 may include a touch-sensitive element that may provide input for adjusting parameters of the control module 32 or data processed by the processor 34.
[0043] The energy storage component 14 can have dimensions comparable to those of a typical lead-acid battery to avoid having to make changes to the vehicle 10 design in order to accommodate the battery system 12. For example, the energy storage component 14 can have dimensions similar to an H6 battery, which can be approximately 13.9 inches x 6.8 inches x 7.5 inches. As shown, the energy storage component 14 can be included in a single continuous housing. In other embodiments, the energy storage component 14 may include a plurality of housings coupled together (e.g., a first housing including a first battery 28 and a second housing including a second battery 30). In yet other embodiments, as described above, the energy storage component 14 may include a first battery module 28 located under the hood of the vehicle 10, and the second battery module 30 may be located inside the vehicle 10.
[0044] More specifically, Figure 3 A schematic diagram of the components of the battery system 12 is shown. Figure 2As mentioned in the discussion of , the control module 32 can adjust the amount of electrical energy captured / provided by each battery module 28 or 30 (e.g., de-rating and re-rating the battery system 12), perform load balancing between the battery modules 28 and 30, determine the state of charge of each battery module 28 or 30, determine the temperature of each battery module 28 or 30, determine the predicted temperature trajectory of the battery module 28 or 30, determine the predicted life of the battery module 28 or 30, determine the fuel economy contribution of the battery module 28 or 30, control the magnitude of the voltage or current output by the AC generator 18 and / or the motor 22, etc. In particular, the control module 32 can measure the state of charge (SOC) and / or the state of health (SOH) based on the battery parameters measured by the sensor 38.
[0045] In some embodiments, the energy storage component 14 may include a single lithium-ion battery or multiple lithium-ion batteries coupled in series. In addition, other rechargeable battery chemistries are contemplated. The energy storage component 14 may release the stored energy to a load 40, which may include an ignition system 16, a vehicle console 20, a display 21, a motor 22, and any other electrical components of the vehicle 10. When the energy storage component 14 releases the stored energy to the load 40, the alternator 18 and / or the motor 22 may provide energy to the energy storage component 14 to replenish the stored energy previously released to the load 40. The sensor 38 may measure battery parameters of the energy storage component 14, and the sensor 38 may transmit the measurements to the control module 32. The battery parameters of the energy storage component 14 may include terminal voltage measurements, terminal current measurements, and battery temperature measurements. The control module 32 processes the measured battery parameters, as described in detail below, to estimate the SOC of the energy storage component 14, two resistances associated with an equivalent circuit model of the energy storage component 14, and a capacitance associated with the equivalent circuit model. As described below with respect to Figure 4 As discussed, the equivalent circuit model may be a 1-RC equivalent circuit model.
[0046] Furthermore, it will be appreciated that the systems and methods described herein may be used to vary the chemistry of the energy storage component 14. For example, the SOC, resistance, and capacitance of the energy storage component 14 may represent a single-cell lithium-ion battery or a multi-cell lithium-ion battery, a single-cell lead-acid battery or a multi-cell lead-acid battery, some combination thereof (e.g., a lithium-ion battery electrically coupled in parallel with a lead-acid battery), or any other single-cell battery chemistry or multi-cell battery chemistry. Furthermore, in an energy storage component 14 having multiple battery chemistries electrically coupled in parallel, the SOC, resistance, and capacitance may represent the entire energy storage component 14, or the SOC, resistance, and capacitance may be calculated for each of the multiple battery chemistries.
[0047] Figure 4 A 1-RC equivalent circuit model 42 of the energy storage component 14 is shown. The 1-RC equivalent circuit model 42 relates battery parameters (e.g., open circuit voltage (OCV) 44, resistances 46 and 48, and capacitance 50) to measured parameters measured by the sensor 38 (e.g., terminal voltage 52, terminal current, and battery temperature). In addition, the 1-RC equivalent circuit model 42 provides a mechanism for estimating OCV in real time during operation of the vehicle 10. Using other methods to measure OCV, such as a coulomb counting method, the OCV may be measurable when the battery system 12 is in a dormant state for a long period of time. That is, the OCV of the battery system 12 may be measurable after the battery system 12 has been dormant for one or more hours. Therefore, by using the 1-RC equivalent circuit model 42, the dormant period is no longer used, and the OCV may be estimated while the battery system 12 is operating under the load 40.
[0048] In the 1-RC equivalent circuit model 42, the resistor 46 (i.e., R0) represents the ohmic resistance of the current path of the energy storage component 14, the resistor 48 (e.g., R1) represents the charge transfer resistance of the energy storage component 14, and the capacitor 50 (e.g., C1) represents the double layer capacitance of the energy storage component 14. The 1-RC equivalent circuit model 42 is considered to be a 1-RC equivalent circuit model attributed to a single resistor-capacitor pair (e.g., resistor 48 and capacitor 50). Using the 1-RC equivalent circuit model 42 enables the OCV 44 to be determined during the real-time driving state of the vehicle 10. In the 1-RC equivalent circuit model 42, the resistors 46 and 48 and the capacitor 50 can generally be time-invariant parameters of the energy storage component 14. Alternatively, the OCV 44 that can be used to determine the state of charge of the energy storage component 14 can generally be a time-varying parameter of the energy storage component 14. That is, when the energy storage component 14 is charged and discharged over a period of time, the OCV 44 will increase and decrease over time.
[0049] Accurate estimation of OCV 44, resistors 46 and 48, and capacitor 50 may be beneficial in controlling energy storage component 14 to achieve longer battery life and improve fuel efficiency of a hybrid electric vehicle. Figure 5A graph 54 is shown that provides a relationship between the OCV 44 and the state of charge (SOC) of the energy storage component 14. The SOC is displayed as a percentage along the abscissa 56 of the graph 54. In addition, the OCV 44 is displayed as a voltage along the ordinate 58 of the graph 54. Curve 60 represents the relationship between the OCV 44 and the SOC of the energy storage component 14. For example, the curve 60 can be used as a lookup table to provide an accurate SOC representation of the energy storage component 14. When the OCV 44 of the 1-RC equivalent circuit model 42 is determined based on the measured battery parameters, the value of the OCV 44 can be matched with the corresponding SOC percentage. The SOC percentage can provide an accurate indication of the remaining battery life of the energy storage component 14 to the operator of the vehicle 10 in real time during the operation of the vehicle 10. In addition, it can be understood that the OCV 44 changes with the change of the SOC. For example, when the SOC of the energy storage component 14 increases or decreases, the OCV 44 will not stabilize at a certain voltage.
[0050] return Figure 4 1-RC equivalent circuit model 42 can be initially derived in discrete time to relate estimates of OCV 44, resistances 46 and 48, and capacitance 50 (i.e., battery parameter estimates) to measured terminal voltage 52 and measured current of energy storage component 14. A Kalman filter is used for this association to determine the battery parameter estimates based on the measured terminal voltage 52 and measured current. Using the Kalman filter, control module 32 can update the battery parameter estimates in real time with limited reliance on predefined battery parameters.
[0051] For any arbitrary current source I, the voltage 52 (eg, V) of the energy storage component 14 can be calculated by the Duhamel superposition theorem:
[0052]
[0053] where ξ is the imaginary variable of integration. The first two terms on the right side of Equation 1 (i.e., V oc and IR0) give an ohmic description of the energy storage component 14, since the voltage 52 is related to the OCV 44 minus the ohmic drop IR0. Furthermore, the third term on the right side of Equation 1 corresponds to a superposition integral by which the impact of past current on the OCV 44 exceeds the first order effect of changes in the average SOC characterized by the energy storage component 14. Due to the exponential weighting function, the impact of older current-potential data points is exponentially smaller than the most recent data points.
[0054] At two arbitrary time steps t k-1 and t k Evaluating Equation 1 yields:
[0055]
[0056] and
[0057]
[0058] It can be assumed that the battery current and voltage are measured at regular time intervals, for example:
[0059]
[0060] Therefore, equations 2 and 3 can be combined to obtain:
[0061]
[0062] If I(ξ) is approximately l k-1 If the step current is , then equation 5 can be subtracted to obtain:
[0063] V k =aV k-1 +(1-a)V OC -I k R0-I k-1 [(1-a)R1-aR0] (6)
[0064] in In addition, if (I k-1 +I k ) / 2 step current pair I(ξ), then equation 5 can be subtracted to obtain:
[0065]
[0066] In addition, if the piecewise linear equation (I k-1 +(I k -I k-1 )) / ((t k -t k-1 )x(ξ-t k-t )) To approximate I(ξ), we can subtract Equation 5 to obtain:
[0067]
[0068] Due to the continuous nature of I(ξ) for all time intervals, Equation 8 can calculate the battery voltage more accurately than Equations 6 and 7, which are derived based on discontinuous step currents. For convenience, Equation 8 can be rewritten as:
[0069] V k =θ1-I k θ3-I k-1 θ2+θ4V k-1 (9)
[0070] in
[0071] θ1=(1-θ4)V OC (10);
[0072]
[0073] and
[0074]
[0075] In this embodiment, Equation 9 is used as a battery model to simultaneously recursively estimate four parameters, θ1-θ4, from the measured voltage and current data using a Kalman filter approach. Physical parameters, such as V oc (e.g., OCV 44), R0 (e.g., resistor 46), R1 (e.g., resistor 48), C1 (e.g., capacitor 50), and the time constant τ=R1C1 are extracted from θ1-θ4 according to the following equation: (16)
[0077]
[0078]
[0079] and
[0080]
[0081] The standard Kalman filter method can be implemented using a dual model process. In order to estimate these four parameters θ1-θ4, using the standard Kalman filter method, the state transition model can be described as:
[0082] Θ(k+1)=Θ(k)+R(k) (18)
[0083] Where R(k) is the process noise vector, Θ(k) is the state vector, and θ1-θ4 are the four parameters of the state vector. In contrast to Equation 18 and other Kalman filter methods, the alternative Kalman filter method described below does not explicitly calculate the state transitions. Instead, the state measurement pattern can be described using the following equation:
[0084] V(k)=Φ(k)′Θ(k)+W(k) (19)
[0085] Where W(k) is the measurement noise and Φ(k) is the regression vector described by the following equation:
[0086] Φ(k)=[1-I(k-1)-I(k)V(k-1)]′ (20)
[0087] Furthermore, by using an alternative Kalman filter approach, the Kalman gain K(k) for SOC estimation is calculated using the following equation:
[0088]
[0089] Therefore, the state transition model can be updated as follows:
[0090] Θ(k)=Θ(k-1)+K(k)[y-Φ′Θ(k-1)] (22)
[0091] Where y is the battery voltage 52 measured at time step k. In addition, the covariance matrix P(k) is updated by the following equation:
[0092] P(k)=[IK(k)Φ(k)′]P(k-1)+R (23)
[0093] In the alternative Kalman filter approach, θ1 is identified as a fast time-varying parameter and θ2-θ4 are identified as slow time-varying parameters. To implement this feature using the alternative Kalman filter approach, the process noise matrix R is defined using the following equation:
[0094]
[0095] Where r is a small value (e.g., r can be set to about 0.001) for controlling the variability of θ1 over time using the random walk concept. This enables the Kalman filter to estimate the time-varying parameters of the rechargeable battery without using a state transition model, which is typically associated with calculating SOC changes by current integration or coulomb counting. In addition, by establishing the three other values on the diagonal of the process noise matrix R(k) to zero, the invariant parameters or slowly varying parameters θ2-θ4 are generally kept constant.
[0096] The calculation of the covariance matrix P(k) plays a role in improving the numerical efficiency and accuracy of estimating battery parameters. Due to truncation and rounding errors in the computer, the alternative Kalman filter method may cause the loss of symmetry and positive definiteness of P(k), or may cause divergence. Therefore, in the alternative Kalman filter method, a UD factorization method for calculating P(k) is implemented. The UD factorization method improves the positive definiteness and symmetry of P(k), which can lead to high estimation accuracy and robustness.
[0097] Implementing the Kalman filter function using the UD factorization method may include calculating the following equation:
[0098] F=U′Φ, G=DF, α0=1 (25)
[0099] When calculating Equation 25, for a range of values from j=1 to j=M, the following equation is calculated:
[0100]
[0101] Additionally, for the range of values from j=1 to j=i-1, the following equation is calculated:
[0102]
[0103] In addition, the value of r is added to D(1,1) by the following equation:
[0104] D(1,1)=D(1,1)+r (28)
[0105] The gain is also calculated using the following equation:
[0106] K=b / α (29)
[0107] After completing the calculations in Equations 25-29, calculate the new estimate using the following equation:
[0108] Θ=Θ+K(y-Φ′Θ) (30)
[0109] Now turn to Fig. 6A and 6B , a flow chart of method 70 shows a method for determining battery parameters of the energy storage component 14 using the above-mentioned alternative Kalman filter method. Initially, at box 72, initialization is performed by the control unit 32. During initialization, arbitrary initial values of the parameters θ1-θ4 can be assigned to the state vector θ. The initial values of the parameters can be arbitrary because the alternative Kalman filter method uses a starting point to determine the final value of the parameter, and the alternative Kalman filter method does not require a starting point close to the actual value of the parameter θ1-θ4. Therefore, the state vector can be initialized to [0 0 0 0]' during initialization. In addition, during initialization, U, can be initialized to a unit diagonal matrix (i.e., a matrix with 1 values assigned to the diagonal elements), and D can be initialized to a diagonal matrix with a large value (e.g., 1000) assigned to each diagonal element. When new voltage and current measurements of the energy storage component 14 from the sensor 38 become available to the control module 32, the initialization can enable the method 70 to start training the measurement model, which is represented by equation 9 above. In addition, the initialization can include setting the values of the relative error tolerance (RTOL) and the scale factor. RTOL may represent a smoothness and flatness criterion, and the scaling factor may be a factor of the battery current to avoid overflow math errors that may be encountered on fixed-point microprocessors.
[0110] Subsequently, at box 74, the data counter may be updated as new data from the sensor 38 becomes available at the control module 32. For example, the following equation may represent the data counter:
[0111] k=k+1 (31)
[0112] Where k is the current data count. In addition, at block 76, the data values measured at the current time step and the previous time step are assigned to the regression vector Φ(k). The measured data values may include measured battery voltage and current values.
[0113] At box 78, a Kalman filter function is executed. The Kalman filter function may include equations 25-30 described above. At box 80, by executing the Kalman filter function, values of battery parameters (e.g., OCV 44, resistors 46 and 48, capacitor 50, and time constant τ) may be extracted from the values of θ1-θ4. Additionally, at box 82, using the value of OCV 44 extracted from the values of θ1-θ4, the SOC of the energy storage component 14 may be determined based on an OCV to SOC lookup table stored in the memory 36 of the control module 32. The OCV to SOC lookup table may generally be based on a formula similar to Figure 5 The curve 60 shown in FIG.
[0114] Boxes 84-96 involve monitoring the convergence of the estimated values of resistors 46 and 48. By monitoring the convergence of resistors 46 and 48, control module 32 can confirm that the estimated values of resistors 46 and 48 are time-invariant. Therefore, at box 84, the mean and variance of a moving sample window of sample number L of resistors 46 and 48 are recursively determined. Specifically, the following equations can be used to determine the mean and variance of resistors 46 and 48:
[0115] If k <= L, then
[0116]
[0117] otherwise
[0118]
[0119] where μ k and σ k 2 is the resistance R evaluated at time step k j As shown in equations 32 and 33, the recursive formula developed for calculating the sample mean and variance does not involve each data copy, and only the new data and the oldest data are involved in the mean and variance calculation. Therefore, the control module 32 effectively calculates equations 32 and 33.
[0120] Subsequently, in box 86, the control module 32 calculates the ratio of the square of the variance and the mean of each resistor 46 and 48, and the control module 32 determines which of the two resistors 46 and 48 has a larger ratio. In addition, once it is determined which ratio is larger, the control module 32 compares the larger ratio with the square of the relative error tolerance (RTOL). This comparison is used as a smoothness and flatness test. The smoothness and flatness test can be used to determine whether the estimated parameters no longer change over time after the learning period of the system. Therefore, in order to determine the convergence of the resistors 46 and 48, the smoothness and flatness test needs to be qualified N times in a row. The value of N can be 5 cycles, 10 cycles, 15 cycles, or any other number of cycles that can reliably indicate the convergence of the resistors 46 and 48. In addition, if a sampling frequency of 1 second is used for the method 70, each cycle of the smoothness and flatness test of box 86 can run for a total time period of 0.1 nanoseconds.
[0121] Then, at block 88, if the larger ratio is less than RTOL squared, the loop counter (e.g., iCheck) is updated by adding 1 to the previous loop counter value. Conversely, if the larger ratio is greater than RTOL squared, then at block 90, the loop counter is reset to zero and the method 70 returns to initialization at block 72. If the loop counter is updated at block 88 (i.e., if the larger ratio is less than RTOL squared), then at block 92, it is determined whether the loop counter exceeds the value of N. If the loop counter does not exceed the value of N, then the method returns to initialization at block 72.
[0122] However, if the cycle counter has exceeded the value of N, then at box 94, the average value of the convergence resistances 46 and 48 and the temperature value are stored in the memory 36 of the control module 32. By measuring the value of the resistances 46 and 48 as a function of temperature, the control module 32 can observe how the resistances 46 and 48 decrease over time. Measuring the decrease in resistances 46 and 48 can enable the state of health (SOH) of the energy storage component 14 to be characterized from the perspective of the increase in the ohmic resistance of the energy storage component 14. Once the average value of the convergence resistances 46 and 48 and the temperature value are stored, the cycle counter is reset to zero at box 96 and the method 70 returns to initialization at box 72. The method 70 can be repeated until the control module 32 provides an indication to stop operation.
[0123] Now go to Fig. 7A and 7B , the recursive calculation results of the sample mean and variance of the resistors 46, 48 with the sample number L (for example, as Fig. 6A8) may be stored in a circular buffer 100 having L storage locations 102. For example, for a sample number L of 10, the circular buffer 100 may include 10 storage locations 102, as shown. When new data 104 including the sample mean and variance of resistors 46 and 48 is calculated, the new data 104 may be stored in one of the storage locations 102. Fig. 7A In the example, empty storage locations 102 are filled with new data 104 in the order in which each sample was recorded. Figure 7B As shown, new data 104 is stored in the storage location 102 of the oldest data value in the circular buffer 100 (e.g., R j (k-9)): The new data 104 is stored in the ring buffer 100 and the old data 106 is removed from the ring buffer 100 (eg, R j (k-9)). Therefore, only L storage locations 102 are available to store the sample means and variances of the sample number L, and the old data 106 is removed from the ring buffer 100.
[0124] There may also be benefits in estimating the capacity of the energy storage component 14 in real time. The capacity of the energy storage component 14 may be referred to as the state of health (SOH) of the energy storage component 14. The SOH of the energy storage component 14 may indicate a change in the rated capacity of the energy storage component. Discussed in detail below are two complementary methods for estimating the capacity of the energy storage component 14. In combination Fig. 8A and Figure 8B The first method discussed uses a Kalman filter and an equivalent circuit battery model to provide a linear regression of the real-time battery current and voltage. Fig. 9A and Fig. 9B The second method discussed involves monitoring two open circuit voltage relaxation events and calculating the SOH from the two open circuit voltage relaxation events. The control module 32 may perform each of the two methods to estimate the capacity of the energy storage component.
[0125] Now turn to Fig. 8A and 8B, which shows a method 120 for calculating the SOH of an energy storage component 14. The method 120 is based on two linear regression models that are run in series at each time step. Because the method 120 relies on a linear regression model, the method 120 can provide numerical stability for estimating the SOH because the parameters of the linear regression model will converge even when the initial values of the parameters are arbitrarily selected. The arbitrarily selected initial values of the parameters can provide a unique advantage over the extended Kalman filter (EKF) model, which generally relies on an accurate initial guess of the parameters of the rechargeable battery due to the nonlinearity of the EKF model. In addition, the method 120 can provide a higher degree of tolerance to measurement noise and data imperfections than the EKF model.
[0126] The partial discharge of the energy storage component 14 may be used to determine the SOH (ie, capacity) of the energy storage component 14 using the following equation:
[0127]
[0128] Where Q is the capacity of the energy storage component 14 in ampere hours (Ah), I is the current in amperes, SOC is the current SOC value, and SOC0 is the initial SOC value. As described above, when an estimate or known value of the open circuit voltage is available, it can be obtained from a method similar to Figure 5 The SOC is obtained by using the lookup table of curve 60. Equation 34 can also be rewritten as the following equation:
[0129] SOC=SOC0+w∫0 t Idt (35) where w is equal to 100 / Q and is the accumulated Ah throughput of the energy storage component 14. Equation 35 can be used as the main equation for battery capacity estimation in method 120. Both SOC0 and w are time-invariant parameters that can be estimated if the SOC value is known. As shown above in Fig. 6A and 6B As discussed in the discussion, the SOC value can be estimated using a Kalman filter technique. Therefore, in method 120, two linear regression models are used. That is, blocks 122 to 132 can determine the SOC value of the energy storage component 14, and blocks 134-156 can use the SOC and Ah throughput as inputs and the convergence determination to provide estimates of SOC0 and w. Therefore, the battery capacity can be easily calculated based on the estimated value of w.
[0130] At box 122, the method 120 is initialized. During initialization, arbitrary initial values can be assigned to θ1-θ4 for the state vector θ. The initial values of the parameters can be arbitrary because the alternative Kalman filter method uses a starting point to determine the final value of the parameter, but the alternative Kalman filter method does not rely on the starting point being close to the actual value of the parameter θ1-θ4. Therefore, the state vector θ can be initialized to [0 0 0 0]' during the initialization event. In addition, during initialization, U, can be initialized to a unit diagonal matrix (i.e., a matrix with 1 values assigned to the diagonal elements), and D can be initialized to a diagonal matrix with a large value (e.g., 1000) assigned to each diagonal element. When new voltage and current measurements of the energy storage component 14 from the sensor 38 become available to the control module 32, the initialization can enable the method 70 to begin training the measurement model, which is represented by equation 9 above. In addition, the initialization can include setting the values of the relative error tolerance (RTOL) and the scaling factor. RTOL may represent a smoothness and flatness criterion and the scaling factor may be a factor of the battery current to avoid overflow math errors that may be encountered on fixed point microprocessors. The initialization at block 122 may be performed for both the linear regression model for SOC estimation and the linear regression model for SOC0 and w.
[0131] At box 124, when new data from the sensor 38 becomes available at the control module 32, the data counter may be updated. For example, equation 31 may be used as a representation of the current data count. In addition, when data is available, the amp-hour throughput may also be updated using the following equation:
[0132] q=q+Δq k (36)
[0133] Where q is the ampere-hour throughput. In addition, at block 126, the data values measured at the current time step and the previous time step are assigned to the regression vector Φ(k). The measured data values may include measured battery voltage and current values.
[0134] At box 128, a Kalman filter function is executed. The Kalman filter function may include equations 25-30 described above. By executing the Kalman filter function, the value of OCV 44 may be extracted from the values of θ1-θ4 at box 130 in addition to the resistors 46 and 48, the capacitor 50, and the time constant τ. Additionally, at box 132, using the value of OCV 44 extracted from the values of θ1-θ4, the SOC of the energy storage component 14 may be determined based on an OCV to SOC lookup table stored in the memory 36 of the control module 32. The OCV to SOC lookup table may generally be based on a formula similar to Figure 5 The curve 60 shown in FIG.
[0135] Then, at box 134, the control module 32 determines whether the estimated SOC can be used as an input to proceed with subsequent capacity estimation of the energy storage component 14. For example, the control module 32 can make a determination based on whether the battery terminal temperature requirement is met (e.g., the temperature of the terminals of the energy storage component is less than about 25 degrees Celsius) and whether the minimum SOC learning time period has been exceeded (e.g., the time interval between the current measurement and the initial condition is greater than 100 seconds). If any of these conditions are not met, then at box 136, the current and voltage measurements are reset to the previous values, and the method 120 restarts at the data counter update of box 124.
[0136] If the SOC is available for capacity estimation, then at box 138, the minimum and maximum SOC values are tracked for use in calculating the maximum SOC swing. In addition, the ampere-hour throughput is assigned to the second regression vector at box 140. The second regression vector is represented by ΦI and represents the following equation:
[0137] Φ1(k)=[1q / scale 1]′ (37)
[0138] where q is the ampere-hour throughput and scale l is a scaling factor for the estimated capacity to avoid overflow math errors when calculating on a fixed-point microprocessor.
[0139] Once the ampere-hour throughput is assigned to the second regression vector, a Kalman filter function is executed at block 142. Executing the Kalman filter function may update the covariance matrices U1 and D1 and the parameter vector θ1 for capacity estimation, which is initialized to [0 0 0 0]' together with the parameter vector θ for SOC estimation. In addition, the Kalman gain K(k) for capacity estimation is represented by the following equation:
[0140]
[0141] And the covariance matrix P1(k) is updated by the following equation:
[0142] P1(k)=[I-K1(k)Φ1(k)′]P1(k-1) (39).
[0143] When compared to Equation 23 for SOC estimation, Equation 39 does not include the process noise matrix R since both SOC0 and w are time-invariant parameters.
[0144] Then, at block 144, an estimated capacity value Q may be extracted from the parameter vector θ1. For example, the estimated capacity value Q may be represented by the following equation:
[0145] Q=-100 / θ12 (40)
[0146] where θ12 is the second parameter from the parameter vector θ1. After extracting the estimated capacity value Q, blocks 146-156 may be used to monitor the convergence of the capacity estimate.
[0147] Therefore, at box 146, the mean and variance of the moving sample window with the number of samples L of the estimated capacity value Q are recursively determined and stored in a circular buffer similar to the circular buffer 100 described above. After determining the mean and variance, at box 148, the ratio of the variance to the mean square value squared is compared with the square of RTOL. This comparison is used as a smoothness and flatness test. The smoothness and flatness test can be used to determine whether the estimated parameter (e.g., the estimated capacity value Q) no longer changes over time after the learning period of the system. Therefore, in order to determine the convergence of the estimated capacity value Q, the smoothness and flatness test is qualified N times in a row. The value of N can be 5 cycles, 10 cycles, 15 cycles, or any other number of cycles that can reliably indicate the convergence of the estimated capacity value Q. In addition, if a sampling frequency of 1 second is used for method 120, each cycle of the smoothness and flatness test of box 148 can run for a total period of 0.1 nanoseconds.
[0148] Subsequently, at box 150, if the ratio is less than RTOL squared, the cycle counter (e.g., iCheck) is updated by adding 1 to the previous cycle counter value. Conversely, if the ratio is greater than RTOL squared, at box 136, the current and voltage measurements are reset to the previous values, and the method 120 restarts at the data counter update at box 124. If the cycle counter is updated at box 150 (i.e., the larger ratio is less than RTOL squared), at box 152, it is determined whether the cycle counter has exceeded the value of N and whether the SOC swing (maximum SOC minus minimum SOC) is greater than 20% of SOC0. The threshold for the SOC swing at box 152 can also be another percentage of SOC0, depending on the specific energy storage component 14. For example, the percentage can be 25% or up to 50%, or the percentage can be 10% or 15% of SOC0. If the cycle counter does not exceed the value of N and / or the SOC swing is not large enough, the method 120 returns to box 136 to reset the voltage and current values.
[0149] If the cycle counter has exceeded the value of N and the SOC swing is sufficient, then at box 154, the extracted estimated capacity value Q is saved to the memory 36. After the estimated capacity value Q is saved to the memory 36, the cycle counter is reset to zero and the method 120 returns to box 136. In addition, the method 120 may be repeated as described above until the control module 32 receives an indication to stop estimating the estimated capacity value Q of the energy storage component 14.
[0150] Fig. 9A and 9B A method 160 is shown for calculating the SOH of an energy storage component 14 by monitoring two valid open circuit voltage (OCV) relaxation events and integrating the current over time between the two valid OCV relaxation events. The first of the two valid OCV relaxation events may be measured immediately upon starting the vehicle 10 after a longer sleep period (e.g., after the vehicle has been turned off for more than 2 hours), and the second of the two valid OCV relaxation events may be measured when the energy storage component 14 has relaxed for a longer period of time (e.g., after the vehicle 10 has been turned off and a predetermined amount of time has passed). In addition, the time between the two valid OCV relaxation events may be limited to limit the accumulation of current offset errors and improve the accuracy of the SOH measurement.
[0151] At box 162, the method 160 is initialized. During initialization, the time counters t1 and t2 for the start ampere-hour throughput ∑q, and the binary OCV validity state iOCV are set. If the OCV of the energy storage component 14 is fully relaxed at the time the control module 32 wakes up (e.g., when the vehicle 10 is started), the value of iOCV can be set to 1 during initialization. Otherwise, the value of iOCV can be set to 0 during initialization. After parking the vehicle at room temperature for a long period of time (e.g., greater than one hour), the OCV can be fully relaxed, with a relaxation current threshold l less than 1. relax One way to confirm whether the OCV is fully relaxed is to monitor the rate of change of the OCV over time, as the OCV may change asymptotically over time. After a long period of time at room temperature, the OCV may change very slowly over time, and this OCV may be used for method 160 to represent the relaxed OCV with sufficient accuracy.
[0152] After initialization, at box 164, the counter is updated when new data from the sensor 38 becomes available at the control module 32. For example, equation 31 can be used as a representation of the current data count. Once the counter is updated, at box 166, data is read from the sensor 38 of the energy storage component 14. The data may include step changes in battery current, voltage, and amp-hour throughput. Therefore, with a step change in amp-hour throughput, at box 168, the accumulated amp-hour throughput is updated using the following equation:
[0153] ∑q=∑q+Δq k (41)
[0154] where ∑q represents the cumulative ampere-hour throughput, and Δq k Represents a step change in ampere-hour throughput.
[0155] At block 170, the control module 32 determines whether the battery current satisfies a predetermined relaxation current threshold l. Relax This can be done by measuring the absolute value of the battery current at the current time step and comparing this absolute value with I Relax This determination is made by comparing the absolute value of the battery current to the Relax , the battery is not sufficiently relaxed. At this point, the time step may be monitored and the iOCV value may remain at 0 or be set to 0 at block 172. Thus, the method 160 may begin again at block 164 where the counter is updated.
[0156] Or, if the absolute value of the battery current is less than I Relax , the iOCV validity state may be determined at block 174. That is, if the value of iOCV is 1, the energy storage component 14 has been in a relaxed state for a sufficient amount of time, and the OCV may be set to the current voltage reading V at block 178. k Alternatively, if iOCV is 0, then at block 176, a determination is made as to whether a sufficient amount of time has elapsed to allow the energy storage component 14 to relax sufficiently. For example, the amount of time from initialization to the current time step may be compared to the relaxation threshold time t Relax As mentioned above, under standard operating conditions, t Relax It can be set to 1 hour, 2 hours or longer. Relax Can be increased or decreased based on external conditions such as battery temperature. Generally, the higher the battery temperature, the lower the t Relax The time may be shorter. If the relaxation threshold time t is not met Relax , the method 160 returns to the counter update of box 164 to start the method 160 again.
[0157] Alternatively, if the relaxation threshold time t has been met Relax , the energy storage component 14 has been in a relaxed state for a sufficient amount of time, then the OCV may be set to the current voltage reading V at block 178. k . According to the current voltage reading V k The OCV value can be determined from a Figure 5 The OCV vs. SOC lookup table of curve 60 determines the current SOC.
[0158] Subsequently, at box 182, it is determined whether the counter value k is greater than the number of samples L. If the counter value k is greater than L, the value of the modulus variable M is updated by incrementing the value by 1 at box 184. The modulus variable M represents the cumulative number of times the OCV standard has been met. Alternatively, if the counter value k is not greater than the number of samples L, the value of L is set to the counter value k plus 1 at box 186. After updating the value of the modulus variable M and / or the number of samples L, a modulus operation is performed at box 188 to alternately record the current SOC value between two memory storage locations within the memory 36 of the control module 32. Therefore, the method 160 will alternately store the current SOC value as SOC1 at box 190 and store the current SOC value as SOC2 at box 192. In addition, the time and cumulative ampere-hour throughput may also be stored in the memory at boxes 190 and 192.
[0159] At block 194, it is determined whether there is sufficient SOC swing between the value stored for SOC1 and the value stored for SOC2. Sufficient SOC swing can be described as SOC MIN , and SOC MIN can represent a threshold percentage difference between two stored SOC values. MIN There may be a 5% swing, a 10% swing, a 15% swing, or another percentage swing that establishes a sufficient difference between the two SOC values for accurate capacity estimation of the energy storage component 14. If the swing is insufficient, the method 160 may return to the counter update at block 164.
[0160] If the SOC swing is sufficient, then at block 196, the control module 32 may determine whether the total current integration time is less than a maximum allowed time. For example, if the total current integration time is greater than the maximum allowed time, the current offset error may affect the capacity estimate in an undesirable manner. Therefore, it may be desirable to limit the total current integration time to less than about 50 hours. If the maximum allowed time has been exceeded, the method 160 may return to the counter update of block 164.
[0161] Alternatively, if the maximum allowed time has not been exceeded, an estimated battery capacity Q of the energy storage component 14 may be calculated at block 198. To calculate the estimated battery capacity Q, the following equation may be used:
[0162]
[0163] The values used for calculation are obtained from the values stored in blocks 190 and 192. Subsequently, at block 200, the estimated battery capacity Q is stored in the memory 36. In addition, when the estimated battery capacity Q is stored, the method 160 may return to the counter update of block 164, and the method 160 may operate recursively until an indication to stop the method 160 is provided to the control module 32.
[0164] The two capacity estimation methods 120 and 160 described above complement each other in practice. Method 120 may not use SOC swings between two OCV relaxation events, but may use real-time SOC swings when the vehicle 10 is running. Therefore, due to the design principles of early start-stop and hybrid electric vehicles, method 120 may be particularly suitable for early start-stop and hybrid electric vehicles to maximize the charging rate of harvested energy when there is an excess supply of kinetic energy and maximize the discharge rate when there is a peak power consumption demand to improve fuel economy. Therefore, the energy storage component 14 can establish a standby (sleep) mode at approximately 50% SOC. Therefore, the energy storage component 14 typically has a small interval between two OCV measurements. Alternatively, due to the simplicity, accuracy and low implementation cost of method 160, when the vehicle 10 undergoes several sleep periods during typical operation, method 160 may be applicable to enhance the robustness and accuracy of the capacity estimation value of the energy storage component 14.
[0165] It may be beneficial to occasionally verify the capacity estimate of the energy storage component 14. Fig.10 210 is a flowchart describing a validation process for the estimated battery capacity Q of the energy storage component 14. At block 212, the control module 32 determines whether the capacity estimate of the energy storage component 14 is valid. For example, if there have not been any valid capacity estimate results from the methods 120 and 160 for a long amount of time (e.g., more than two months since the last valid capacity estimate result), if the most recently calculated capacity estimate exceeds a change threshold (e.g., the most recently calculated capacity estimate is 5%, 10%, 15%, or more than the previously calculated capacity estimate), or if the accuracy of the valid capacity estimate is not within an acceptable range, the capacity estimate may be invalid. When any of these criteria are met, the control module 32 may initiate the validation process of the flowchart 210.
[0166] In determining whether the accuracy of the effective capacity estimate is not within an acceptable range, the error in the SOC / OCV measurement and the error in the capacity estimate calculated via the current integration that represents the current measurement accuracy can be used. For example, the Kalman filter method may include an error estimate or a maximum accuracy estimate (e.g., approximately 3%), and evaluation of the convergence criteria as described herein can refine such an error estimate. After refining the error estimate, an error increase related to time or energy throughput based on expected aging under observed conditions of the battery system 12 can be applied to the error estimate. Therefore, if the capacity estimate is determined within a certain accuracy, the error steadily increases until there is an opportunity to reach a new estimate with a lower error, which will reset the error estimate.
[0167] If it is determined that the capacity estimate is valid, the current capacity estimate may be updated at block 214. The capacity updated at block 214 may be used as the actual capacity of the energy storage component 14 during a current integration process using the current capacity estimate (e.g., using equation 35 above) at block 216. The result of the current integration process using the actual capacity of the energy storage component is a value of SOC(1) calculated over an integration period at block 218.
[0168] Alternatively, if it is determined that the capacity estimate is invalid (e.g., the average error is too high and resetting will not improve accuracy), a candidate capacity may be used in a current integration process in parallel with the current integration process of block 216 at block 220. The candidate capacity may be 5% less than the actual capacity used in the current integration process of block 216. The result of the current integration process using the candidate capacity is a value of SOC(2) over an integration time period calculated at block 222. Additionally, measured parameters 224 of the energy storage component 14 may be used when performing the current integration process at blocks 216 and 220. The measured parameters 224 may include system / sensor specifications, temperature, current, and voltage of the energy storage component 14.
[0169] In order to effectively use the verification process of flowchart 210, an accurate initial SOC at the beginning of the parallel current integration process can be used. Therefore, the measured parameter 224 can provide a value for the battery control module 32 to calculate an accurate initial SOC. For example, if the vehicle 10 exits a longer sleep period of the energy storage component 14 before the parallel current integration process at boxes 216 and 220, the open circuit voltage 44 of the energy storage component 14 can be measured from the measured parameter 224, as discussed in detail above. Using the open circuit voltage 44, an open circuit voltage-state of charge lookup table stored in the memory 36 can be consulted by the control module 32 to calculate an accurate initial SOC value. It can also be understood that a significant SOC increment (e.g., SOC swing) between the initial SOC at the beginning of the parallel current integration process and the final SOC at the time point when the changing capacity estimate is evaluated may be beneficial to accurately evaluate the effectiveness of the capacity estimate. This value can be calculated by ΔSOC δQ express.
[0170] Additionally, a Kalman filter may be used at box 226 at times when a change in the estimated capacity value is being evaluated for comparison with the SOC value generated by the parallel current integration process. For example, using the method 70 discussed above in connection with 6A and 6B, an estimated SOC value (SOC(3)) for the energy storage component 14 may be calculated by the control module 32 at box 228. Alternatively or additionally, if the energy storage component 14 is in a dormant state, the OCV 44 may be obtained from the energy storage component 14 at box 230. Consultation of the OCV to SOC lookup table by the control module 32 at box 232 results in the calculation of SOC(4) at box 234.
[0171] After calculating SOC(1)-SOC(3) and / or SOC(4), a directional comparison of the SOC values may be performed by the control module 32 at box 236. For example, the values of SOC(3) and SOC(4) at the time point at which the estimated capacity value is evaluated to have changed may be compared to the values of SOC(1) and SOC(2) at the same time point. This comparison may provide many details about the new estimated capacity value of the energy storage component 14. If the values of SOC(3) and / or SOC(4) fall within the range of the values of SOC(1) and SOC(2) or the values of SOC(3) and / or SOC(4) are in the same direction relative to SOC(1) as the value of SOC(2), the control module 32 may deem the estimated capacity value calculated from SOC(3) and / or SOC(4) to be valid. In this case, the control module 32 may calculate the capacity at this point in time and update the actual capacity value to the newly calculated estimated capacity at box 214.
[0172] Alternatively, if the values of SOC(3) and / or SOC(4) do not fall within the range of the values of SOC(1) and SOC(2), the control module 32 may deem the estimated capacity values from SOC(3) and / or SOC(4) invalid at block 236. In this case, the process of flowchart 210 may restart at a different initial SOC value, or SOC(3) and / or SOC(4) may be calculated at a later time to determine whether the capacity values based on SOC(3) and / or SOC(4) are valid.
[0173] Go to Fig.11 , the diagram 240 shows Fig.10 210. The vertical axis 242 represents the SOC of the energy storage component 14 in percentage. The horizontal axis 244 represents the time when the SOC is measured. Line 246 represents the current integral of the energy storage component 14 when the actual capacity of the energy storage component is used, and line 248 represents the parallel current integral of the energy storage component 14 when the candidate capacity of the energy storage component 14 is used, which is shown as 5% lower than the actual capacity of the energy storage component 14.
[0174] As shown, the parallel current integration process begins at time t1 with an accurate initial SOC value 250. The accurate initial SOC value 250 can be obtained through open circuit voltage (OCV) measurements and comparison of the OCV measurements with an OCV to SOC lookup table. In addition, the parallel current integration process can be performed until time t2. Time t2 can represent the ΔSOC of the parallel current integration process. Q The state of charge swing may be a time (e.g., a state of charge swing) sufficient to accurately verify the estimated capacity of the energy storage component 14. For example, the state of charge swing may be a swing of about 20% to accurately verify the estimated capacity. In other cases, the state of charge swing may be a swing of about 10%, 15%, or up to 25% or more to accurately verify the estimated capacity.
[0175] As the parallel integration process approaches time t2, several SOC measurements 252 and 254 of the energy storage component, calculated, for example, using a Kalman filter method, are plotted. The darker SOC measurements 252 represent SOC measurements that are invalid for capacity estimation updates. For example, the SOC swing (e.g., ΔSOC) of the parallel current integration lines 246 and 248 is not valid. δQ ) compared to the SOC change (eg, ΔSOC k ) is in the opposite direction. Therefore, if the value of the dark SOC measurement 252 is calculated by the control module 32, the control module 32 will wait for a subsequent valid SOC measurement to update the capacity estimate.
[0176] Alternatively, the light-colored SOC measurement 254 represents the SOC measurement that is valid for the capacity estimate update. For example, the SOC change (eg, ΔSOC k ) and the SOC swing (eg, ΔSOC δQ ) are in the same direction. Therefore, if the value of the light SOC measurement 254 is calculated by the control module 32, the control module 32 will update the capacity estimate when calculating the light SOC measurement 254.
[0177] One or more of the disclosed embodiments may provide one or more technical effects, alone or in combination, including determining a battery time-varying variable and a non-time-varying variable, determining a battery state of charge, determining a battery state of health, and verifying a battery state of health. The technical effects and technical problems in the specification are exemplary and not restrictive. It should be noted that the embodiments described in the specification may have other technical effects and may solve other technical problems.
[0178] Although only certain features and embodiments have been shown and described, many modifications and changes (e.g., changes in size, dimensions, structure, shape and proportion of various elements, values of parameters (e.g., temperature, pressure, etc.), mounting arrangements, use of materials, colors, orientations, etc.) may occur to those skilled in the art without materially departing from the novel teachings and advantages of the disclosed subject matter. The order or sequence of any process or method steps may be changed or reordered according to alternative embodiments. It should be understood, therefore, that the appended claims are intended to cover all such modifications and changes within the true spirit of the disclosure. In addition, in order to provide a concise description of the exemplary embodiments, not all features of an actual implementation may be described. It should be understood that in the development of any such actual implementation, as in any engineering or design project, many implementation specific decisions may be made. Such development efforts may be complex and time consuming, but are a routine task of design, production, and manufacture for an ordinary technician who has the benefit of this disclosure without undue experimentation.
Claims
1. A battery system, comprising: a battery configured to be coupled to an electrical system; and a battery control module configured to be electrically coupled to the battery, wherein the battery control module is configured to: performing a parallel current integration process using the actual capacity and the candidate capacity of the battery at an initial charging state; performing a directional comparison between an estimated state of charge of the battery and a result of the parallel current integration process; as well as determining validity of the estimated state of charge based at least in part on the directional comparison between the estimated state of charge of the battery and a result of the parallel current integration process; wherein the parallel current integration process includes current integration of the initial state of charge using the actual capacity of the battery and current integration of the initial state of charge using the candidate capacity of the battery, the candidate capacity being 5% smaller than the actual capacity; wherein the battery control module is configured to calculate the estimated state of charge of the battery using a Kalman filter method, an open circuit voltage-state of charge lookup table, or both; wherein the battery control module is configured to start the parallel current integration process at the initial charging state, and the battery control module is configured to perform the parallel current integration process for a period of time, the period of time being long enough to allow the charging state to swing by more than 20%; The battery control module is configured to update the actual capacity by calculating an estimated value of the effective capacity at a certain moment when calculating the estimated state of charge.
2. The battery system of claim 1 , wherein the battery control module is configured to perform the parallel current integration process in response to determining that a valid capacity estimate of the battery has not been received for more than two months.
3. The battery system of claim 1 , wherein the battery control module is configured to perform the parallel current integration process in response to an indication that the capacity has changed by more than 10%.
4. The battery system of claim 1 , wherein the initial state of charge is calculated by the battery control module using an open circuit voltage measurement and an open circuit voltage-state of charge lookup table.
5. The battery system of claim 1 , wherein the battery control module is configured to validate the estimated state of charge in response to receiving a value of the estimated state of charge, the received value of the estimated state of charge falling within a value of the parallel current integration process when the estimated state of charge is obtained by the battery control module.
6. The battery system of claim 1 , wherein the battery control module is configured to compare a subsequent estimated state of charge with a result of the parallel current integration process in response to the estimated state of charge being greater than a result of the parallel current integration process for the actual capacity and the result of the parallel current integration process for the candidate capacity being less than a result of the parallel current integration process for the actual capacity. 7 . The battery system of claim 1 , wherein the battery control module is configured to repeat validity determinations until a value of the estimated state of charge falls within a result of the parallel current integration process. 8 . The battery system of claim 1 , wherein the battery control module is configured to update the actual capacity when the battery control module verifies that the estimated state of charge is valid.
9. A method of verifying a capacity estimate of a rechargeable battery configured to be coupled to an electrical system, comprising: performing a parallel current integration process by a battery control module of the rechargeable battery using an actual capacity of the rechargeable battery and a candidate capacity of the rechargeable battery in an initial charging state; Calculating an estimated state of charge of the battery using a Kalman filter method, an open circuit voltage-state of charge lookup table, or both; directional comparison, by the battery control module, of the estimated state of charge of the rechargeable battery and a result of the parallel current integration process; determining, by the battery control module, validity of the estimated state of charge based at least in part on a directional comparison of the estimated state of charge and a result of the parallel current integration process; as well as When calculating the estimated state of charge, updating the actual capacity by calculating an estimated value of the effective capacity at a certain moment; wherein performing the parallel current integration process includes performing the current integration of the initial charging state using the actual capacity of the battery and performing the current integration of the initial charging state using the candidate capacity of the battery, the candidate capacity being 5% smaller than the actual capacity; as well as The parallel current integration process is started at the initial charging state and is performed for a period of time, which is long enough to allow the charging state to swing by more than 20%.
10. The method of claim 9, wherein the parallel current integration process is performed when the battery control module has not received a valid capacity estimate for more than two months and / or when the battery control module calculates that the actual capacity has changed by more than 10%.
11. The method of claim 9, wherein determining the validity of the estimated state of charge comprises determining whether a value of the estimated state of charge falls within a result of the parallel current integration process.
12. The method according to claim 9, comprising: When the value of the estimated state of charge falls within the result of the parallel current integration process, the capacity estimation value of the rechargeable battery is updated to a new capacity estimation value based on the estimated state of charge.
13. A battery system (12), comprising: - a battery configured to be coupled to an electrical system; and - a battery control module (32), the battery control module being configured to be electrically coupled to the battery, wherein the battery control module (32) is configured to: - performing a parallel current integration process, wherein, in order to perform the parallel current integration process, the battery control module (32) is configured as follows: - determining a first estimated state of charge of the battery using a first current integration process based on a first estimated capacity, an initial state of charge of the battery at the beginning of a current integration period, and a current indicated by a battery parameter measured during the current integration period by one or more sensors communicatively coupled to the battery control module (32); - determining a second estimated state of charge of the battery using a second current integration process in parallel with the first current integration process based on a second estimated capacity of the battery, the initial state of charge of the battery, and a current indicated by the battery parameter measured during the current integration period, wherein the second estimated capacity of the battery is 5% less than the first estimated capacity of the battery; - determining a third estimated state of charge of the battery using a Kalman filter method, an open circuit voltage-state of charge lookup table, or both; - performing a directional comparison between the third estimated state of charge of the battery and the first and second estimated states of charge of the battery to determine whether the value of the third estimated state of charge falls within a range of the values of the first and second estimated states of charge; - determining the validity of the third estimated state of charge of the battery based at least in part on the directional comparison between the third estimated state of charge of the battery and the first estimated state of charge and the second estimated state of charge of the battery, wherein the third estimated state of charge is determined to be valid if the value of the third estimated state of charge falls within the range of the value of the first estimated state of charge and the value of the second estimated state of charge; and - in case the third estimated state of charge is valid, calculating the actual capacity of the battery; wherein the battery control module (32) is configured to start the parallel current integration process at an initial charging state, and the battery control module (32) is configured to perform the parallel current integration process within the current integration period, the current integration period being a period of time long enough to allow the charging state to swing by more than 20%; wherein the battery control module is configured to update the first estimated capacity when the battery control module verifies that the third estimated charging state is valid; and The battery control module (32) is configured to update the first estimated capacity to the calculated actual capacity after calculating the third estimated state of charge.
14. A method of verifying a capacity estimate of a rechargeable battery configured to be coupled to an electrical system, comprising: - executing a parallel current integration process by a battery control module (32) of the rechargeable battery, wherein executing the parallel current integration process comprises: - performing, by the battery control module (32), a first current integration process to determine a first estimated state of charge of the battery based on a first estimated capacity, an initial state of charge of the battery at the start of a current integration period, and a current indicated by a battery parameter measured during the current integration period by one or more sensors communicatively coupled to the battery control module (32); - performing, by the battery control module (32), a second current integration process in parallel with the first current integration process to determine a second estimated state of charge of the battery based on a second estimated capacity of the battery, the initial state of charge of the battery, and the current indicated by the battery parameter measured during the current integration period, wherein the second estimated capacity of the battery is 5% less than the first estimated capacity of the battery; - calculating a third estimated state of charge of the battery using a Kalman filter method, an open circuit voltage-state of charge lookup table, or both; - directionally comparing, by the battery control module (32), the third estimated state of charge of the rechargeable battery with the first estimated state of charge and the second estimated state of charge of the battery to determine whether the value of the third estimated state of charge falls within the range of the value of the first estimated state of charge and the value of the second estimated state of charge; - determining, by the battery control module (32), the validity of the third estimated state of charge of the battery based at least in part on the directional comparison between the third estimated state of charge of the battery and the first estimated state of charge and the second estimated state of charge, wherein the third estimated state of charge is determined to be valid if the value of the third estimated state of charge falls within a range between the value of the first estimated state of charge and the value of the second estimated state of charge; - in case the third estimated state of charge is valid, calculating the actual capacity of the battery; and - after calculating the third estimated state of charge, updating the first estimated capacity to the calculated actual capacity, wherein the parallel current integration process is performed starting at the initial state of charge and during the current integration period, the current integration period being a period of time long enough to allow the state of charge to swing by more than 20%.
