BATTERY MODEL ESTIMATION BASED ON BATTERY TERMINAL VOLTAGE AND CURRENT TRANSIENT DUE TO LOAD POWERED FROM THE BATTERY

Abstract

A method of management of a battery that powers a component of a device may include monitoring a terminal voltage and a terminal current of the battery under a load that is drawing a current on the battery to provide power to a component of the device and modeling the battery as a battery model that approximates a relationship between the monitored terminal voltage and terminal current over at least one of: a certain frequency range; a certain duration, a certain amplitude range, an applied load, a set of conditions of the battery, and a set of conditions of the load. The relationship between the terminal voltage and the terminal current may have a frequency-dependent characteristic including at least two time constants. The two time constants may represent a time-varying relationship between an input and output of the battery model.

Description

RELATED APPLICATION

The present disclosure claims priority to U.S. Provisional Patent Application Ser. No. 63/107,727, filed Oct. 30, 2020, and to U.S. Provisional Patent Application Ser. No. 63/109,573, filed Nov. 4, 2020, both of which are incorporated by reference herein in their entireties.

FIELD OF DISCLOSURE

The present disclosure relates in general to circuits for electronic devices, including without limitation personal portable devices such as wireless telephones and media players, and more specifically, to battery model estimation for a battery which may be used to power components of an electronic device.

BACKGROUND

Portable electronic devices, including wireless telephones, such as mobile/cellular telephones, tablets, cordless telephones, mp3 players, and other consumer devices, are in widespread use. Such a portable electronic device may include a battery (e.g., a lithium-ion battery) for powering components of the portable electronic device.

In operation, the terminal voltage of a battery may droop under a load current due to internal output impedance of the battery. Such output impedance may be modeled in a number of suitable manners, including with an equivalent circuit model of a series of parallel-coupled resistors and capacitors. Knowledge of the detailed impedance of a battery may be useful for fuel-gauging algorithms (e.g., for determining a battery open-circuit voltage and state of charge, predicting power limits, and/or deriving safety limits or safe operation limits of the battery (e.g., a maximum voltage and maximum current of the battery terminal)).

Using traditional approaches, an impedance model is typically estimated using a low-level sinusoidal test current or short-term impulse. However, such approaches may impose loading conditions on the battery which are different from real-use-case conditions of a mobile device, which may lead to an inadequate model.

In addition, using traditional approaches, estimation of a battery impedance model having a multi-order complex impedance is computationally expensive.

SUMMARY

In accordance with the teachings of the present disclosure, one or more disadvantages and problems associated with existing approaches to modeling battery impedance may be reduced or eliminated.

In accordance with embodiments of the present disclosure, a method of management of a battery that powers a component of a device may include monitoring a terminal voltage and a terminal current of the battery under a load that is drawing a current on the battery to provide power to a component of the device and modeling the battery as a battery model that approximates a relationship between the monitored terminal voltage and terminal current over at least one of: a certain frequency range, a certain duration, a certain amplitude range, an applied load, a set of conditions of the battery, and a set of conditions of the load. The relationship between the terminal voltage and the terminal current may have a frequency-dependent characteristic including at least two time constants. The two time constants may represent a time-varying relationship between an input and output of the battery model.

In accordance with these and other embodiments of the present disclosure, a system for management of a battery that powers a component of a device may include battery monitoring circuitry configured to monitor a terminal voltage and a terminal current of the battery under a load that is drawing a current on the battery to provide power to a component of the device and a battery model estimator configured to model the battery as a battery model that approximates a relationship between the monitored terminal voltage and terminal current over at least one of: a certain frequency range, a certain duration, a certain amplitude range, an applied load, a set of conditions of the battery, and a set of conditions of the load. The relationship between the terminal voltage and the terminal current may have a frequency-dependent characteristic including at least two time constants. The two time constants may represent a time-varying relationship between an input and output of the battery model.

In accordance with these and other embodiments of the present disclosure, a method for estimating parameters of a battery impedance model that models an output impedance of a battery may include dividing the battery impedance model into a plurality of separate impedance stages, wherein each separate impedance stage approximates the battery impedance model within a particular frequency range, the respective impedance in each separate impedance stage comprises a primary impedance model with a primary set of defining impedance parameters and a secondary impedance model with a secondary set of impedance parameters, and the battery impedance model is defined by a series connection of the respective primary impedance models of the plurality of impedance stages. The method may also include monitoring operation of the battery to determine the primary set of defining impedance parameters and the secondary set of impedance parameters.

In accordance with these and other embodiments of the present disclosure, a system for estimating parameters of a battery impedance model that models an output impedance of a battery may include one or more inputs configured to receive information regarding operation of the battery and battery monitoring circuitry. The battery monitoring circuitry may be configured to divide the battery impedance model into a plurality of separate impedance stages, wherein each separate impedance stage approximates the battery impedance model within a particular frequency range, the respective impedance in each separate impedance stage comprises a primary impedance model with a primary set of defining impedance parameters and a secondary impedance model with a secondary set of impedance parameters, and the battery impedance model is defined by a series connection of the respective primary impedance models of the plurality of impedance stages, The battery monitoring circuitry may also be configured to monitor operation of the battery to determine the primary set of defining impedance parameters and the secondary set of impedance parameters.

Technical advantages of the present disclosure may be readily apparent to one skilled in the art from the figures, description and claims included herein. The objects and advantages of the embodiments will be realized and achieved at least by the elements, features, and combinations particularly pointed out in the claims.

It is to be understood that both the foregoing general description and the following detailed description are examples and explanatory and are not restrictive of the claims set forth in this disclosure.

BRIEF DESCRIPTION OF THE DRAWINGS

A more complete understanding of the present embodiments and advantages thereof may be acquired by referring to the following description taken in conjunction with the accompanying drawings, in which like reference numbers indicate like features, and wherein:

FIG. 1 illustrates a block diagram of selected components of an example power delivery network, in accordance with embodiments of the present disclosure;

FIG. 2 illustrates an example graph of an open circuit voltage of a battery versus the battery's state of charge, in accordance with embodiments of the present disclosure;

FIG. 3 illustrates a circuit diagram of selected components of an equivalent circuit model for a battery, in accordance with embodiments of the present disclosure;

FIG. 4 illustrates a circuit diagram of selected components of an equivalent circuit model for a battery depicting estimation of a battery impedance model in stages, in accordance with embodiments of the present disclosure;

FIG. 5 illustrates a circuit diagram of selected components of an equivalent impedance model for a first estimation stage, in accordance with embodiments of the present disclosure;

FIG. 6 illustrates a circuit diagram of selected components of an equivalent impedance model for a second estimation stage, in accordance with embodiments of the present disclosure;

FIG. 7 illustrates a circuit diagram of selected components of an equivalent impedance model for a third estimation stage, in accordance with embodiments of the present disclosure;

FIG. 8 illustrates a circuit diagram of selected components of an equivalent impedance model for a fourth estimation stage, in accordance with embodiments of the present disclosure;

FIG. 9 illustrates a block diagram of selected components of a battery model estimator, in accordance with embodiments of the present disclosure;

FIG. 10A illustrates a block diagram of a least-squares fit method for estimating primary and second battery model parameters, in accordance with embodiments of the present disclosure; and

FIG. 10B illustrates a block diagram of a least-squares fit method for estimating primary battery model parameters, in accordance with embodiments of the present disclosure.

DETAILED DESCRIPTION

FIG. 1 illustrates a block diagram of selected components of an example power delivery network 10, in accordance with embodiments of the present disclosure. In some embodiments, power delivery network 10 may be implemented within a portable electronic device, such as a smart phone, tablet, game controller, and/or other suitable device.

As shown in FIG. 1, power delivery network 10 may include a battery 12 and a load 18. As shown in FIG. 1, when loaded by load 18, battery 12 may generate a battery voltage V_CELL across its terminals and deliver a battery current I_CELL to load 18. In some embodiments, battery 12 may comprise a lithium-ion battery. Load 18 may represent any electric component, electronic component, and/or combination thereof. For example, load 18 may include any suitable functional circuits or devices of power delivery network 10, including without limitation power converters, processors, audio coder/decoders, amplifiers, display devices, etc. Further, although not explicitly shown in FIG. 1, power delivery network 10 may also include control circuitry for controlling operation of battery 12 and/or load 18.

As also shown in FIG. 1, power delivery network 10 may include battery monitoring circuitry 20. Battery monitoring circuitry 20 may include any suitable system, device, or apparatus configured to monitor battery voltage V_CELL and battery current I_CELL. Further, battery monitoring circuitry 20 may include a battery model estimator 24 configured to receive monitor battery voltage V_CELL and a sense voltage V_SNS across a sense resistor 22 indicative of battery current I_CELL, and based thereon, estimate a battery impedance model for battery 12, as described in greater detail below. Battery model estimator 24 may be implemented with a processing device, including without limitation a microprocessor, digital signal processor, application-specific integrated circuit, field-programmable gate array, electrically-erasable programmable read only memory, complex programmable logic device, and/or other suitable processing device. In some embodiments, battery monitoring circuitry 20 may monitor a temperature associated with battery 12, and battery model estimator 24 may estimate the impedance model based on battery voltage V_CELL, a sense voltage V_SNS, and the sensed temperature.

Lithium-ion batteries are typically known to operate from 4.5 V down to 3.0 V, known as an open circuit voltage V_OC of the battery (e.g., battery 12). As a battery discharges due to a current drawn from the battery, the state of charge of the battery may also decrease, and open circuit voltage V_OC (which may be a function of state of charge) may also decrease as a result of electrochemical reactions taking place within the battery, as shown in FIG. 2. Outside the range of 3.0 V and 4.5 V for open circuit voltage V_OC, the capacity, life, and safety of a lithium-ion battery may degrade. For example, at approximately 3.0 V, approximately 95% of the energy in a lithium-ion cell may be spent (i.e., state of charge is 5%), and open circuit voltage V_OC would be liable to drop rapidly if further discharge were to continue. Below approximately 2.4 V, metal plates of a lithium-ion battery may erode, which may cause higher internal impedance for the battery, lower capacity, and potential short circuit. Thus, to protect a battery (e.g., battery 12) from over-discharging, many portable electronic devices may prevent operation below a predetermined end-of-discharge voltage. Knowledge of the output impedance may be useful in determining open circuit voltage V_OC and other parameters of battery 12.

FIG. 3 illustrates a block diagram of selected components of an equivalent circuit model for battery 12, in accordance with embodiments of the present disclosure. As shown in FIG. 3, battery 12 may be modeled as having a battery cell 32 having an open circuit voltage V_OC in series with a plurality of parallel resistive-capacitive sections 34 (e.g., parallel resistive-capacitive sections 34-1, 34-2, . . . , 34-N) and further in series with an equivalent series resistance 36 of battery 12, such equivalent series resistance 36 having a resistance of R₀. Resistances R₁, R₂, . . . R_N, and respective capacitances C₁, C₂, . . . , C_N may model battery chemistry-dependent time constants τ₁, τ₂, . . . , τ_N, that may be lumped with open circuit voltage V_OC and equivalent series resistance 36. The series of impedance sections represented by resistive-capacitive sections 34 and equivalent series resistance 36 may represent diffusion processes that occur at different rates inside battery 12. Cutoff frequencies of parallel resistive-capacitive sections 34 may respectively be given by:

$f_{c1}=\frac{1}{{\tau }_1}=\frac{1}{2\pi R_1{C}_1};$ $f_{c2}=\frac{1}{{\tau }_2}=\frac{1}{2\pi R_2{C}_2};$ $\dots$ $f_{cN}=\frac{1}{{\tau }_3}=\frac{1}{2\pi R_N{C}_N};$

wherein π represents the well-known mathematical constant defined as the ratio of a circle's circumference to its diameter, and wherein parallel resistive-capacitive sections 34 are ordered such that f_cN< . . . <f_c2<f_c1.

Notably, an electrical node depicted with voltage V_CELL-EFF in FIG. 3 may capture the time varying discharge behavior of battery 12, and battery voltage V_CELL may be an actual voltage seen across the output terminals of battery 12. Voltage V_CELL-EFF may not be directly measurable, and thus battery voltage V_CELL may be the only voltage associated with battery 12 that may be measured to evaluate battery state of health. Also of note, at a current draw of zero (e.g., I_CELL=0), battery voltage V_CELL may be equal to voltage V_CELL-EFF which may in turn be equal to an open circuit voltage V_OC at a given state of charge.

To estimate an impedance model of battery 12, battery model estimator 24 may divide the output impedance of battery 12 into a number of stages, thus breaking down the estimation problem into several identification problems of lower order. Such estimation in stages may be possible due to the fact that cutoff frequencies (or time constants) of the diffusion processes represented by the various parallel resistive-capacitive sections 34 may be separated by an order of magnitude or more.

Although an impedance model of battery 12 may include any number N of stages as shown in FIG. 3, battery model estimator 24 may estimate an impedance model using any suitable number of stages and performing an estimation for each individual stage. For example, in some embodiments, battery model estimator 24 may estimate the impedance model of battery 12 in four estimation stages 40-1, 40-2, 40-3, and 40-4, as shown in FIG. 4. Estimation stage 40-1 may estimate resistance R₀, resistance R₁, and capacitance C₁, estimation stage 40-2 may estimate resistance R₂ and capacitance C₂, estimation stage 40-3 may estimate resistance R₃ and capacitance C₃, and estimation stage 40-4 may estimate resistance R₄ and capacitance C₄.

In operation, battery model estimator 24 may estimate the full battery impedance model as the sum of respective primary impedance models 42 (e.g., primary impedance models 42-1, 42-2, 42-3, and 42-4) for each of estimation stages 40. A primary impedance model 42 for a particular estimation stage 40 may define a main feature of the full battery impedance model within a frequency band FB_M centered around the cutoff frequency (e.g., f_c1, f_c3, f_c3, f_c4) associated with the primary impedance of such estimation stage 40. For instance, if a primary impedance model 42-M of an estimation stage 40-M has a single cutoff frequency defined by resistance R_M in parallel with capacitance C_M, then a frequency f_cM may be the cutoff frequency of such primary impedance model 42-M.

However, a primary impedance of an estimation stage 40-M may not be directly identified from characteristics of battery voltage V_CELL and battery current I_CELL in frequency band FB_M, because the primary impedances of other estimation stages 40 may contribute to the frequency response of the full battery impedance model within frequency band FB_M. Accordingly, battery model estimator 24 may estimate the impedance model for each estimation stage 40-M using a secondary impedance model for such estimation stage 40-M, such secondary impedance model modelling the residual impedance of the neighboring preceding estimation stages (e.g., estimation stages 40-1 to 40[M−1]) and subsequent estimation stages (e.g., estimation stages 40[M+1] to 40-N) over the frequency band FB_M for such estimation stage 40-M. For example, for estimation stage 40-1, the subsequent estimation stages 40-2, 40-3, and 40-4 with primary impedance models R₂∥C₂, R₃∥C₃, R₄∥C₄, have a residual impedance within frequency band FB₁ associated with estimation stage 40-1, which may be modeled with a lumped secondary impedance model 44-1 R₂∥C₂, as shown in FIG. 5.

As depicted in FIG. 5, an impedance model for estimate stage 40-1 may include a primary impedance model 42-1 with impedance R₀+R₁∥C₁ having primary parameters R₀, R₁, and C₁ modeling the effects of impedance near and above cutoff frequency f_c1. Primary impedance model 42-1 may be in series with secondary impedance model 44-1 with impedance R₂∥C₂ and having secondary parameters R₂ and C₂ modeling the effects of residual impedance of neighboring estimation stages 40 near cutoff frequency f_c1.

In operation, battery model estimator 24 may filter measurements of battery voltage V_CELL and battery current I_CELL, with a bandpass filter centered around cutoff frequency f_c1, or alternatively highpass filter such measurements with a high-pass filter having a cutoff frequency above cutoff frequency f_c2 but below cutoff frequency f_c1. Further, battery model estimator 24 may use a least-squares method or other appropriate fit approach in order to fit primary parameters R₀, R₁, and C₁ and secondary parameters R₂ and C₂ to the filtered battery voltage V_CELL and battery current I_CELL sequence.

Battery model estimator 24 may perform two estimation steps to estimate such parameters. In a first estimation step (e.g., as described with reference to FIG. 10A, below), a least-squares method may estimate both primary parameters R₀, R₁, and C₁ and secondary parameters R₂ and C₂ simultaneously. In a second estimation step (e.g., as described with reference to FIG. 10B, below), battery model estimator 24 may set secondary parameters R₂ and C₂ of the residual impedance by subsequent estimates of the primary impedance in the neighboring estimation stages 40, and may only least-squares fit primary impedance parameters R₀, R₁, and C₁ of estimation stage 40-1 during the second estimation step. Such second estimation step may iteratively improve accuracy of the fitted parameters. Thus, in this example, battery model estimator 24 may set secondary parameters R₂ and C₂ of estimation stage 40-1 from an estimation of the primary parameters R₂ and C₂ of estimation stage 40-2.

As depicted in FIG. 6, an impedance model for estimate stage 40-2 may include a primary impedance model 42-2 with impedance R₂∥C₂ having primary parameters R₂ and C₂ modeling the effects of impedance near cutoff frequency f_c2. Primary impedance model 42-2 may be in series with secondary impedance model 44-2a with impedance R₃∥C₃ having secondary parameters R₃ and C₃ and with another secondary impedance model 44-2b with impedance R_L1=R₀+R₁, wherein secondary parameters R_L1, R₃, and C₃ model the effects of residual impedance of neighboring estimation stages 40 near cutoff frequency f_c2.

In operation, battery model estimator 24 may filter measurements of battery voltage V_CELL and battery current I_CELL with a bandpass filter centered around cutoff frequency f_c2, or alternatively highpass filter such measurements with a high-pass filter having a cutoff frequency above cutoff frequency f_c3 but below cutoff frequency f_c2. Further, battery model estimator 24 may use a least-squares method or other appropriate fit approach in order to fit primary parameters R₂ and C₂ and secondary parameters R_L1, R₃, and C₃ to the filtered battery voltage V_CELL and battery current I_CELL sequence.

Battery model estimator 24 may perform two estimation steps to estimate such parameters. In a first estimation step (e.g., as described with reference to FIG. 10A, below), a least-squares method may estimate both primary parameters R₂ and C₂ and secondary parameters R_L1, R₃, and C₃ simultaneously. In the second estimation step (e.g., as described with reference to FIG. 10B, below), battery model estimator 24 may set secondary parameters R_L1, R₃, and C₃ of the residual impedance by subsequent estimates of the primary impedance in the neighboring estimation stages 40, and may only least-squares fit primary impedance parameters R₂ and C₂ of estimation stage 40-2 during the second estimation step. Such second estimation step may iteratively improve accuracy of the fitted parameters. Thus, in this example, battery model estimator 24 may set secondary parameters R₂ and C₂ of estimation stage 40-2 from an estimation of the primary parameters R₃ and C₃ of estimation stage 40-3 and from an estimation of the primary parameters R₀ and R₁ of estimation stage 40-1.

As depicted in FIG. 7, an impedance model for estimation stage 40-3 may include a primary impedance model 42-3 with impedance R₃∥C₃ having primary parameters R₃ and C₃ modeling the effects of impedance near cutoff frequency f_c3. Primary impedance model 42-3 may be in series with secondary impedance model 44-3a with impedance R₄∥C₄ having secondary parameters R₄ and C₄ and with another secondary impedance model 44-3b with impedance R_L2=R₀+R₁+R₂, wherein secondary parameters R_L2, R₄, and C₄ model the effects of residual impedance of neighboring estimation stages 40 near cutoff frequency f_c3.

In operation, battery model estimator 24 may filter measurements of battery voltage V_CELL and battery current I_CELL with a bandpass filter centered around cutoff frequency f_c3, or alternatively highpass filter such measurements with a high-pass filter having a cutoff frequency above cutoff frequency f_c4 but below cutoff frequency f_c3. Further, battery model estimator 24 may use a least-squares method or other appropriate fit approach in order to fit primary parameters R₃ and C₃ and secondary parameters R_L2, R₄, and C₄ to the filtered battery voltage V_CELL and battery current I_CELL sequence.

Battery model estimator 24 may perform two estimation steps to estimate such parameters. In a first estimation step (e.g., as described with reference to FIG. 10A, below), a least-squares method may estimate both primary parameters R₃ and C₃ and secondary parameters R_L2, R₄, and C₄ simultaneously. In the second estimation step (e.g., as described with reference to FIG. 10B, below), battery model estimator 24 may set secondary parameters R_L2, R₄, and C₄ of the residual impedance by subsequent estimates of the primary impedance in the neighboring estimation stages 40, and may only least-squares fit primary impedance parameters R₃ and C₃ of estimation stage 40-3 during the second estimation step. Such second estimation step may iteratively improve accuracy of the fitted parameters. Thus, in this example, battery model estimator 24 may set secondary parameters R₃ and C₃ of estimation stage 40-3 from an estimation of the primary parameters R₄ and C₄ of estimation stage 40-4 and from an estimation of the primary parameters R₀, R₁, and R₂ of estimation stages 40-1 and 40-2.

As depicted in FIG. 8, an impedance model for estimate stage 40-4 may include a primary impedance model 42-4 with impedance R₄∥C₄ having primary parameters R₄ and C₄ modeling the effects of impedance near cutoff frequency f_c4. Primary impedance model 42-4 may be in series with secondary impedance model 44-4 with impedance R_L3=R₀+R₁+R₂+R₃, wherein secondary parameter R_L3 models the effects of residual impedance of neighboring estimation stages 40 near cutoff frequency f_c4.

In operation, battery model estimator 24 may filter measurements of battery voltage V_CELL and battery current I_CELL with a bandpass filter centered around cutoff frequency f_c4. Further, battery model estimator 24 may use a least-squares method or other appropriate fit approach in order to fit primary parameters R₄ and C₄ and secondary parameter R_L3 to the filtered battery voltage V_CELL and battery current I_CELL sequence.

Battery model estimator 24 may perform two estimation steps to estimate such parameters. In a first estimation step (e.g., as described with reference to FIG. 10A, below), a least-squares method may estimate both primary parameters R₄ and C₄ and secondary parameter R_L3 simultaneously. In the second estimation step (e.g., as described with reference to FIG. 10B, below), battery model estimator 24 may set secondary parameter R_L3 of the residual impedance by subsequent estimates of the primary impedance in the neighboring estimation stages 40, and may only least-squares fit primary impedance parameters R₄ and C₄ of estimation stage 40-3 during the second estimation step. Such second estimation step may iteratively improve accuracy of the fitted parameters. Thus, in this example, battery model estimator 24 may set secondary parameters R₄ and C₄ of estimation stage 40-4 from an estimation of the primary parameters R₀, R₁, R₂, and R₃ of estimation stages 40-1, 40-2, and 40-3.

The first estimation step for each estimation stage 40 may only be needed once when no a priori estimate of secondary parameters is defined. Once the primary parameters of every estimation stage 40 have been estimated via this first estimation step, battery model estimator 24 may pass on the primary parameters as secondary parameters of neighboring estimation stages 40 to simplify and constrain the estimation process, as well as potentially leading to iteratively more accurate estimates of the full battery impedance model.

Once the full model parameters are estimated, battery model estimator 24 may estimate open circuit voltage V_OC, voltage V_CELL-EFF, internal overpotential states of battery 12, a lithium-ion anode potential, and/or other state representing a condition of the battery that may lead to degradation of its chemistry, by monitoring current I_CELL (e.g., which may be indicated by sense voltage V_SNS). Battery model estimator 24 may estimate such overpotential states with a simple filter model or with a Kalman filter, for example.

FIG. 9 illustrates a block diagram of selected components of battery model estimator 24, in accordance with embodiments of the present disclosure. As shown in FIG. 9, battery model estimator 24 may sample battery voltage V_CELL and battery current I_CELL with analog-to-digital converters (ADCs), and battery model estimator 24 may decimate such sampled battery voltage V_CELL and battery current I_CELL for each estimation stage 40 to a lower rate sufficient to obtain a suitable estimation. Such signal decimation may reduce an amount of processing required in estimating the battery impedance model while still maintaining desirable numerical precision. Such decimated signals are bandpass filtered near the respective cutoff frequencies f_c1, f_c2, f_c3, and f_c4 for estimation stages 40, and once filtered, battery model estimator 24 may perform least squares fits to determine the various primary and secondary parameters, as described above. Further, once the battery impedance model has been established (and which may be dynamically updated over time), battery model estimator 24 may apply such battery impedance model to the monitored battery current I_CELL to estimate open circuit voltage V_OC, voltage V_CELL-EFF, internal overpotential states of battery 12, a lithium-ion anode potential, and/or other state representing a condition of the battery that may lead to degradation of its chemistry.

FIG. 10A illustrates a block diagram of a least-squares fit method for estimating primary and second battery model parameters, in accordance with embodiments of the present disclosure. The least-squares fit method depicted in FIG. 10A may be used to perform the “first estimation step” described for each of estimation stages 40. As shown in FIG. 10A, linear combinations of delayed samples of the decimated, band-pass-filtered measurements of battery voltage V_CELL and battery current I_CELL may be used to form a set of regressors. A cross-correlation matrix of regressors may be low-pass filtered at the respective cutoff frequency (e.g., f_c1, f_c2, f_c3, f_c4) for the particular estimation stage 40. Such low-pass filter may comprise a simple accumulate-and-dump filter or a first-order low-pass filter with a bandwidth lower than the lowest cutoff frequency present in the impedance model for such estimation stage 40. Battery model estimator 24 may use the filtered correlations to determine a least-squares fit solution.

FIG. 10B illustrates a block diagram of a least-squares fit method for estimating primary battery model parameters, in accordance with embodiments of the present disclosure. The least-squares fit method depicted in FIG. 10B may be used to perform the “second estimation step” described for each of estimation stages 40. As shown in FIG. 10B, in the second estimation step, the least-squares solution may be further restrained by presetting the secondary parameters to lead to a more accurate estimate of the primary parameters.

In accordance with the foregoing, methods and systems may be provided to estimate parameters of a full battery impedance model that models an output impedance of a battery that powers a load and that is valid over a full range of frequencies of interest. The full battery model may be separated into a number N of separate impedance estimation stages, wherein N is an integer of 2 or greater. Each impedance estimation stage may approximate a full battery impedance model for a limited frequency range. For each of the separate impedance estimation stages, the respective impedance in each stage may include a primary impedance model with a primary set of defining parameters and a secondary impedance model with a secondary set of parameters. The full battery impedance model may be defined by a series connection of all the primary impedance models. Thus, the primary parameters may define the parameters of the full battery model.

A full range of frequencies of interest for the battery model may be from 1 mHz to 10 KHz. Methods and systems disclosed herein may further use measured voltage and current associated with a transient response of the battery under a switching load such as naturally exists in a mobile device. In some embodiments, a broadband test-excitation of the battery may be used to generate such voltage and current.

The primary impedance model of an impedance estimation stage M (wherein M is between 1 and N) may define a main feature (such as an electrical component or the combination of electrical components that define the dominant impedance over that frequency range) of the impedance of the full battery model over a frequency range for the impedance estimation stage M. The secondary impedance model of an impedance estimation stage M (M is between 1 and N) may define a residual feature of the primary impedance model of each impedance estimation stage 1 to M−1 and M+1 to N over the frequency range of the impedance estimation stage M. The secondary impedance model may be a lumped model of impedance of the primary impedance estimation stages 1 to M−1 and M+1 to N.

In accordance with the systems and methods disclosed herein, battery monitoring circuitry may monitor a terminal voltage and terminal current of a battery. For each impedance estimation stage M, the battery monitoring circuitry may band-pass filter the terminal voltage and terminal current in a frequency range associated with such impedance estimation stage M. The battery monitoring circuitry may apply a least-squares fit method to the bandpass-filtered terminal voltage and terminal current measurements to estimate parameters of the full battery impedance model over the impedance estimation stage M.

The least-squares fit method may be performed in two steps. The first estimation step may include estimating the primary and secondary parameters of the impedance estimation stage M; as in the first estimation step, sufficient a priori knowledge of the secondary parameters may not be available in order to perform estimation. The second estimation step may involve presetting the secondary parameters of the impedance estimation stage M and only estimating the primary parameters of the impedance estimation stage M using a least-squares fit method. In some instances, the secondary parameters may become stale and the battery monitoring circuitry may execute the first estimation step and the second estimation step again. In some embodiments, the first parameters step may be performed only once, and the second parameters step may be repeatedly performed (e.g., for tracking parameters).

In some embodiments, the battery monitoring circuitry may constrain some of the estimated parameters based on a priori knowledge of a valid range for the respective parameters. The battery monitoring step may constrain one or more estimated parameters by characterizing a population of samples of the battery.

The battery monitoring circuitry may use a multi-rate implementation wherein, for each impedance estimation stage M, the battery monitoring circuitry uses a sample rate that is sufficient to accurately fit the full battery model over the defined frequency range of such impedance estimation stage M but is low enough to reduce processing requirements of the least-squares fit. In some embodiments, such sample rate may be a sample rate that is at least two times a high cutoff frequency of the bandpass filter of the impedance estimation stage M.

The battery model estimation described herein may be disabled, or its adaptation rate reduced, when a signal-to-noise ratio (SNR) over a frequency band of interest is below a predetermined threshold. Such disabling of battery model estimation or reduction of the adaption rate for estimating parameters may occur if a load powered from a battery is idle or has strong spectral content out of the frequency band of interest. A noise figure required to estimate the SNR over the frequency band may be estimated a priori when the load is absent or a minimal load is detected in the frequency band (that may occur, for instance, when a mobile device is idle). A noise figure may be equally estimated from a voltage and current sensor noise specification and bandpass filter parameters.

Estimation of parameters of an impedance estimation stage M is disabled if content over a frequency band for the impedance estimation stage M is not sufficiently spectrally diverse to estimate the impedance estimation stage M parameters unambiguously (overfit). The spectral condition may be determined by monitoring some terms of a cross-correlation matrix of a least-squares fit.

The full battery impedance model may be implemented with a discrete time infinite impulse response filter, such as a Kalman filter, for example. The full battery impedance model and inferred states may be used for fuel-gauging algorithms or power limits estimation of the battery.

Battery monitoring circuitry may apply the full battery model to a monitored battery voltage and/or battery current to estimate internal states of the full battery model such as overpotentials and an open circuit voltage. In some embodiments, a terminal voltage and terminal current of the battery under a load that is drawing a current on the battery to provide power to the load of the device may be monitored. The battery may be modeled as a battery model that approximates the relationship between the monitored terminal voltage and terminal current over one of the following: (1) a certain frequency range, (2) a certain duration or amplitude range or an applied load, or (3) a set of conditions of the battery or the load, wherein battery monitoring circuitry may be able to estimate a voltage associated with the battery (e.g., open circuit voltage) from the terminal current.

The relationship between the monitored terminal voltage and terminal current may have a frequency dependent characteristic including at least two time constants. The at least two time constants represent a time-varying relationship between an input and an output of a linear or nonlinear model of the battery. As described above, the battery model may have model parameters. The model parameters may be determined through an optimization function. The model parameters and the battery model may be used to predict battery characteristics.

Although the foregoing contemplates battery monitoring circuitry configured to estimate a battery model, in some embodiments, a battery model may be determined in another manner. For example, a battery model may comprise one or more of the following: (1) a linear or non-linear model of the battery, (2) a parameterized equivalent circuit model that models impedance of the battery, (3) a physics-based model, (4) a combination of an equivalent circuit model and a physics-based model, or (5) a Kalman filter or extended Kalman filter. In the case of a parameterized equivalent circuit model, the parameters for modeling an impedance of the battery may include resistive, capacitive and/or inductive circuit elements that may be in parallel or in series and which element impedances may be time varying and/or may have nonlinear characteristics.

A filtered monitored terminal voltage and current may be used to derive an estimated signal-to-noise ratio (SNR) metric or power spectral density (PSD) metric. Parameters of the battery model relevant to the one or more frequency bands may be adapted if sufficient SNR is detected and if the PSD is sufficient to prevent overfitting the battery model. Adaptation of the parameters of the battery model relevant to the one or more frequency bands may be disabled during an idle period or under a constant direct current (DC) load. The adaptation rate of a parameter may be a function of the estimated SNR and of the expected parameter rate of change. A noise figure required for deriving the SNR may be determined when no load or very light load is present such as when the device powered by the battery enters into a sleep mode.

The optimization function may be a least square fit or a frequency or time weighted variant of a least square fit. The battery model may be used in conjunction with a fuel-gauging algorithm. The battery model may be used to estimate a limit of the battery and limit the power drawn from the battery. The battery model may be used to protect the battery from conditions that degrade the battery lifespan. The battery model may be used to design optimized charging or loading conditions of the battery.

A component drawing power from the battery may be a component on a mobile device, such as a central processing unit (CPU), a graphics processing unit (GPU), a power management unit (PMU), an amplifier, an audio, haptic or other actuator, a flash light-emitting-diode (LED), LED screen, a micro-electro-mechanical systems (MEMS) sensor or other sensor, and/or a passive component.

As used herein, when two or more elements are referred to as “coupled” to one another, such term indicates that such two or more elements are in electronic communication or mechanical communication, as applicable, whether connected indirectly or directly, with or without intervening elements.

This disclosure encompasses all changes, substitutions, variations, alterations, and modifications to the example embodiments herein that a person having ordinary skill in the art would comprehend. Similarly, where appropriate, the appended claims encompass all changes, substitutions, variations, alterations, and modifications to the example embodiments herein that a person having ordinary skill in the art would comprehend. Moreover, reference in the appended claims to an apparatus or system or a component of an apparatus or system being adapted to, arranged to, capable of, configured to, enabled to, operable to, or operative to perform a particular function encompasses that apparatus, system, or component, whether or not it or that particular function is activated, turned on, or unlocked, as long as that apparatus, system, or component is so adapted, arranged, capable, configured, enabled, operable, or operative. Accordingly, modifications, additions, or omissions may be made to the systems, apparatuses, and methods described herein without departing from the scope of the disclosure. For example, the components of the systems and apparatuses may be integrated or separated. Moreover, the operations of the systems and apparatuses disclosed herein may be performed by more, fewer, or other components and the methods described may include more, fewer, or other steps. Additionally, steps may be performed in any suitable order. As used in this document, “each” refers to each member of a set or each member of a subset of a set.

Although exemplary embodiments are illustrated in the figures and described below, the principles of the present disclosure may be implemented using any number of techniques, whether currently known or not. The present disclosure should in no way be limited to the exemplary implementations and techniques illustrated in the drawings and described above.

Unless otherwise specifically noted, articles depicted in the drawings are not necessarily drawn to scale.

All examples and conditional language recited herein are intended for pedagogical objects to aid the reader in understanding the disclosure and the concepts contributed by the inventor to furthering the art, and are construed as being without limitation to such specifically recited examples and conditions. Although embodiments of the present disclosure have been described in detail, it should be understood that various changes, substitutions, and alterations could be made hereto without departing from the spirit and scope of the disclosure.

Although specific advantages have been enumerated above, various embodiments may include some, none, or all of the enumerated advantages. Additionally, other technical advantages may become readily apparent to one of ordinary skill in the art after review of the foregoing figures and description.

To aid the Patent Office and any readers of any patent issued on this application in interpreting the claims appended hereto, applicants wish to note that they do not intend any of the appended claims or claim elements to invoke 35 U.S.C. § 112(f) unless the words “means for” or “step for” are explicitly used in the particular claim.

Claims

1. A method of management of a battery that powers a component of a device, comprising:

monitoring a terminal voltage and a terminal current of the battery under a load that is drawing a current on the battery to provide power to a component of the device;

modeling the battery as a battery model that approximates a relationship between the monitored terminal voltage and terminal current over at least one of: a certain frequency range, a certain duration, a certain amplitude range, an applied load, a set of conditions of the battery, and a set of conditions of the load;

wherein: the relationship between the terminal voltage and the terminal current has a frequency-dependent characteristic including at least two time constants; and the two time constants represent a time-varying relationship between an input and output of the battery model.

2. The method of claim 1, wherein the battery model has parameters and the method further comprises determining the model parameters through an optimization function.

3. The method of claim 2, wherein the optimization function is a least squares fit.

4. The method of claim 2, wherein the optimization function is a frequency- or time-weighted variant of a least squares fit.

5. The method of claim 1, wherein the battery model includes at least one of: a linear model of the battery, a non-linear model of the battery, a parameterized equivalent circuit model that models impedance of the battery, a physics-based model, a combination of an equivalent circuit model and a physics-based model, a Kalman filter, and an extended Kalman filter.

6. The method of claim 1, further comprising isolating and filtering the terminal voltage and the terminal current over one or more frequency bands in order to model the battery.

7. The method of claim 1, further comprising using the battery model to predict battery characteristics.

8. The method of claim 7, wherein the battery characteristics include at least one of: a maximum available power of the battery, a state of charge of the battery, a state of health of the battery, and an internal state of the battery.

9. The method of claim 8, wherein the internal state may include at least one of an open-circuit voltage of the battery, an internal overpotential state of the battery, a lithium-ion anode potential of the battery, and some other state representing a condition of the battery that may lead to degradation of its chemistry.

10. The method of claim 1, wherein:

the battery model includes a parameterized equivalent circuit model that models impedance of the battery; and

the battery model includes parameters for modeling an impedance of the battery including resistive, capacitive, and/or inductive circuit elements in parallel or in series.

11. The method of claim 10, wherein impedances of the circuit elements are time varying.

12. The method of claim 10, wherein impedances of the circuit elements have nonlinear characteristics.

13. A system for management of a battery that powers a component of a device, the system comprising:

battery monitoring circuitry configured to monitor a terminal voltage and a terminal current of the battery under a load that is drawing a current on the battery to provide power to a component of the device; and

a battery model estimator configured to model the battery as a battery model that approximates a relationship between the monitored terminal voltage and terminal current over at least one of: a certain frequency range, a certain duration, a certain amplitude range, an applied load, a set of conditions of the battery, and a set of conditions of the load;

wherein: the relationship between the terminal voltage and the terminal current has a frequency-dependent characteristic including at least two time constants; and the two time constants represent a time-varying relationship between an input and output of the battery model.

14. The system of claim 13, wherein the battery model has parameters and the battery model estimator is further configured to determine the model parameters through an optimization function.

15. The system of claim 14, wherein the optimization function is a least squares fit.

16. The system of claim 14, wherein the optimization function is a frequency- or time-weighted variant of a least squares fit.

17. The system of claim 13, wherein the battery model includes at least one of: a linear model of the battery, a non-linear model of the battery, a parameterized equivalent circuit model that models impedance of the battery, a physics-based model, a combination of an equivalent circuit model and a physics-based model, a Kalman filter, and an extended Kalman filter.

18. The system of claim 13, wherein the battery model estimator is further configured to isolate and filter the terminal voltage and the terminal current over one or more frequency bands in order to model the battery.

19. The system of claim 13, wherein the battery model estimator is further configured to predict battery characteristics using the battery model.

20. The system of claim 19, wherein the battery characteristics include at least one of: a maximum available power of the battery, a state of charge of the battery, a state of health of the battery, and an internal state of the battery.

21. The system of claim 20, wherein the internal state may include at least one of an open-circuit voltage of the battery, an internal overpotential state of the battery, a lithium-ion anode potential of the battery, and some other state representing a condition of the battery that may lead to degradation of its chemistry.

22. The system of claim 13, wherein:

the battery model includes a parameterized equivalent circuit model that models impedance of the battery; and

the battery model includes parameters for modeling an impedance of the battery including resistive, capacitive, and/or inductive circuit elements in parallel or in series.

23. The system of claim 22, wherein impedances of the circuit elements are time varying.

24. The system of claim 22, wherein impedances of the circuit elements have nonlinear characteristics.