A method of estimating a state of charge of a battery

Abstract

The application relates to a battery power state estimation method, which comprises the following steps: acquiring real-time operation data of voltage, current and temperature of a battery, and obtaining a reference SOC value and a dynamic SOC value based on the real-time operation data; establishing a hierarchical calibration arbitrator comprising multiple calibration modes; the hierarchical calibration arbitrator judges and selects a highest priority calibration mode whose trigger condition is met from high to low according to a preset priority based on the reference SOC value, the dynamic SOC value and the real-time operation data, so as to generate a calibration SOC value; and according to the source mode of the calibration SOC value and the corresponding confidence, a preset correction strategy is adopted to correct the reference SOC value by using the calibration SOC value, so as to obtain a final SOC estimation value. The application has the advantages of not depending on ideal conditions and being able to maintain long-term accurate estimation under complex dynamic working conditions.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power estimation, in particular to a battery state of charge estimation method. BACKGROUND

[0002] The battery management system is a core component in electrochemical energy storage units, one of its key functions is to accurately estimate the state of charge of the battery, usually referred to as SOC. Accurate SOC estimation is of great significance for optimizing energy use strategies, ensuring safe operation of the system, and providing reliable information to users about the remaining range.

[0003] In the prior art, the ampere-hour integration method is a widely used SOC estimation method. This method tracks the change in power by integrating the current flowing into and out of the battery in real time, and under ideal conditions it can better reflect the short-term dynamics of the power. However, the ampere-hour integration method is essentially an open-loop estimation method, which is sensitive to measurement errors of the current sensor and uncertainty of the initial SOC value of the battery. These small deviations will accumulate over time, causing the SOC estimation result to gradually deviate from its actual value, resulting in significant cumulative errors.

[0004] In order to correct the cumulative error of the ampere-hour integration method, a calibration method based on the open-circuit voltage of the battery is usually introduced. The basis of this method is that the open-circuit voltage of the battery has a relatively stable relationship with its SOC. However, to obtain an accurate open-circuit voltage, the battery needs to be left alone for a long period of time after the load is disconnected, so that the internal electrochemical reactions reach equilibrium. In many practical application scenarios, such as continuously driving electric vehicles or participating in grid frequency regulation energy storage systems, such long periods of rest are not common. Therefore, the calibration method that relies on rest conditions has limited applicability in these dynamic operating conditions, and it is difficult to effectively suppress the error divergence of the ampere-hour integration method, resulting in unreliable long-term estimation accuracy of the SOC. SUMMARY

[0005] In order to solve the problem that the existing battery SOC estimation method relies too much on ideal calibration conditions and the long-term estimation accuracy is difficult to guarantee in complex and dynamic operating conditions, the present application provides a battery state of charge estimation method.

[0006] The battery state of charge estimation method provided by the present application adopts the following technical solution:

[0007] A battery state of charge estimation method, comprising the following steps:

[0008] S1. Obtain real-time running data of voltage, current and temperature of the battery, and obtain a reference SOC value which is a long-term power accumulation reference by ampere-hour integration method based on the real-time running data, and obtain a dynamic SOC value which can reflect current voltage characteristics of the battery in real time by a filter based on a battery model;

[0009] S2. Establish a hierarchical calibration arbitrator including multiple calibration modes, and the hierarchical calibration arbitrator judges and selects a highest priority calibration mode whose trigger condition is met according to preset priorities from high to low based on the reference SOC value, the dynamic SOC value and the real-time running data, to generate a calibration SOC value;

[0010] S3. According to the source mode of the calibration SOC value and the corresponding confidence degree, a preset correction strategy is adopted to correct the reference SOC value by using the calibration SOC value, to obtain a final SOC estimation value.

[0011] Optionally, the S1 includes the following steps:

[0012] S11. Based on the reference SOC value of the last period, the updated reference SOC value is calculated by time integration of the current in the real-time running data of the current period, combined with the rated capacity and coulomb efficiency of the battery;

[0013] S12. An equivalent circuit model of the battery is established as a state space equation, and the current in the real-time running data is used as the input of the state space equation to obtain a predicted SOC value;

[0014] S13. The voltage in the real-time running data is used as the observation value of the state space equation, and the predicted SOC value is corrected by comparing the error between the observation value and the predicted output voltage of the equivalent circuit model, to determine the dynamic SOC value.

[0015] Optionally, the calibration modes of the hierarchical calibration arbitrator include: absolute zero point calibration mode, floating anchor point calibration mode, quasi-static voltage calibration mode, on-demand dynamic impedance calibration mode.

[0016] Optionally, the S2 includes the following steps:

[0017] S21. With the highest priority, it is judged whether the real-time running parameters of the battery reach a preset charge-discharge physical boundary condition, and if so, the absolute zero point calibration mode is selected;

[0018] S22. If the physical boundary condition is not met, then with the second highest priority, determine whether a preset electrochemical feature indicative of the internal state of the battery can be identified in the operational data of the battery, and if so, select the floating anchor point calibration mode;

[0019] S23. If the electrochemical feature is not identified, then with the third highest priority, determine whether the battery satisfies a preset quasi-static state defined by a combination of a duration threshold and a current threshold, and if so, select the quasi-static voltage calibration mode;

[0020] S24. In parallel with S21-S23, maintain a counter for tracking the cumulative operation since the last successful calibration;

[0021] S25. With the lowest priority, determine whether the value of the counter reaches a preset threshold, and if so, and none of the aforementioned physical boundary condition, electrochemical feature and quasi-static state is met in the current operation cycle, then select the on-demand dynamic impedance calibration mode, wherein the on-demand dynamic impedance calibration mode is characterized in that a probing signal is actively injected into the battery for calibration;

[0022] S26. After any calibration mode is selected and successfully executed, reset the counter.

[0023] Optionally, the step of selecting the absolute zero point calibration mode to generate the calibration SOC value comprises:

[0024] S211. Monitor whether the battery reaches a preset full charge state or a preset full discharge state;

[0025] S212. When the battery reaches the full charge state, generate the calibration SOC value as 100%;

[0026] S213. When the battery reaches the full discharge state, generate the calibration SOC value as a preset minimum safe SOC value.

[0027] Optionally, the step of selecting the floating anchor point calibration mode to generate the calibration SOC value comprises:

[0028] S221. Determine whether the battery is in a continuous and current-stable charging or discharging process;

[0029] S222. During the process, calculate an online voltage change rate characteristic curve based on the trajectory of the voltage change with respect to the cumulative charge;

[0030] S223. identifying one or more preset characteristic peaks from the characteristic curve, and determining a current SOC value as the calibration SOC value according to a pre-stored correspondence between the characteristic peaks and SOC.

[0031] Optionally, the step of selecting the quasi-static voltage calibration mode to generate the calibration SOC value comprises:

[0032] S231. judging whether the battery is in a quasi-static state in which the absolute value of its current is less than a preset threshold value;

[0033] S232. when the battery is in the quasi-static state, using the dynamic SOC value output by the battery model-based filter as the calibration SOC value.

[0034] Optionally, the step of selecting the on-demand dynamic impedance calibration mode to generate the calibration SOC value comprises:

[0035] S241. judging whether other calibration modes of higher priority have failed to be triggered within a preset time period or a preset cumulative charge period;

[0036] S242. if other calibration modes of higher priority have failed to be triggered, actively requesting a power conversion system connected to the battery to superimpose a detection current signal;

[0037] S243. calculating a dynamic impedance characteristic of the battery according to its voltage response to the detection current signal;

[0038] S244. determining a current SOC value as the calibration SOC value based on a pre-stored correspondence between the dynamic impedance characteristic and SOC.

[0039] Optionally, the S3 comprises the following steps:

[0040] S31. judging which specific calibration mode in the hierarchical calibration arbitrator generates the calibration SOC value;

[0041] S32. if the calibration SOC value is derived from the absolute zero-point calibration mode, instantaneously resetting the reference SOC value to the calibration SOC value as the final SOC estimation value;

[0042] S33. if the calibration SOC value is derived from a calibration mode other than the absolute zero-point calibration mode, gradually smoothing the reference SOC value to the calibration SOC value through a preset convergence algorithm within a preset period of time or a plurality of operation cycles, and outputting the reference SOC value during the dynamic correction process as the final SOC estimation value in real time.

[0043] In summary, the present application includes at least one of the following beneficial technical effects:

[0044] 1. By establishing a multi-mode, hierarchical calibration arbitration mechanism, the inherent cumulative error of the ampere-hour integral method can be continuously corrected. Compared with the existing technology, which will obviously decrease in accuracy after a long time of operation, the present method can ensure that the SOC estimation result always maintains high accuracy and long-term stability during the battery's life cycle and multiple charge-discharge cycles.

[0045] 2. By fusing multiple calibration modes such as absolute zero point, floating anchor point, quasi-static voltage, and on-demand dynamic impedance, the present method reduces the dependence on specific ideal conditions such as long-term static or complete charge-discharge cycles. Whether the battery is in a partial capacity interval for frequent use or in continuous dynamic operation, the method can find applicable calibration strategies, thereby effectively working in more extensive and complex actual application scenarios.

[0046] 3. The present application improves the intelligence level of the battery management system and the reliability of the output data. The hierarchical calibration arbitrator makes decisions based on clear technical criteria, always preferring calibration methods with better data quality. This not only provides the system with an accurate SOC value, but more importantly, improves the credibility of the value, thereby providing a more solid and reliable data foundation for the energy management, safety protection, and endurance prediction of the upper-layer application of the vehicle or energy storage system. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 A flow chart of the battery state of charge estimation method in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0048] The present application will be further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely intended to explain the present application and are not intended to limit the present application.

[0049] In the following description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the inventive concept. As part of the description, some of the diagrams in the present disclosure are represented in block diagram form to avoid obscuring the disclosed principles. Not all features of a practical implementation are necessary in order to describe the disclosed principles. Further, the language used in the present disclosure has been principally selected for readability and instructional purposes and can not have been selected to delineate or circumscribe the inventive subject matter, resort to the claims being necessary to determine such an inventive subject matter. Reference in the specification to "one implementation" or "an implementation" means that a particular feature, structure, or characteristic described is included in at least one implementation, and multiple references to "one implementation" or "an implementation" do not necessarily all refer to the same implementation.

[0050] The terms "a," "an," and "the" are not intended to refer to singular entities, but include the general class of which a specific example can be used for illustration. The use of the terms "a" or "an" can mean any number of, including "one," "one or more," "at least one," and "one or more than one." The term "or" means any one of the alternatives, as well as any combination of the alternatives, including all of the alternatives, unless the alternatives are expressly indicated to be mutually exclusive. The phrase "at least one of" followed by a list of items means any individual item in the list, as well as any combination of items in the list, unless expressly indicated to the contrary. The phrase "at least one of" followed by a list of items means any individual item in the list, as well as any combination of items in the list, unless expressly indicated to the contrary.

[0051] As a key component of electrochemical energy storage units, one of the core tasks of battery management systems is to accurately estimate the state of charge (SOC) of the battery. The accuracy of SOC estimation directly affects the energy management strategy, operational safety, and reliability of the range information of the system.

[0052] In the prior art, a common SOC estimation strategy is to combine the ampere-hour integration method with the open-circuit voltage calibration method. The ampere-hour integration method tracks the change in charge by integrating the current, but as an open-loop estimation method, it is susceptible to the influence of inaccurate initial values and sensor measurement deviations, resulting in cumulative errors over time.

[0053] To correct the cumulative error of the ampere-hour integration method, the prior art usually adopts an open-circuit voltage calibration method. This method relies on the characteristics that the terminal voltage of the battery can stabilize and reflect the SOC after a long period of rest. However, in many actual dynamic working conditions such as continuous operation of electric vehicles or uninterrupted participation of energy storage systems in grid services, the long rest condition is difficult to meet. Therefore, the lack of calibration opportunities makes it impossible to effectively suppress the error of the ampere-hour integration method, resulting in a decrease in the accuracy of SOC estimation in long-term use, which limits the reliability of this technical solution in complex application scenarios.

[0054] Therefore, the present application proposes a method for estimating the state of charge of a battery, with reference to Figure 1 , comprising the following steps S1-S3.

[0055] S1. Obtain real-time operating data of the voltage, current and temperature of the battery, and obtain a reference SOC value which is a long-term charge accumulation reference by the ampere-hour integration method in parallel based on the real-time operating data, and obtain a dynamic SOC value which can reflect the current voltage characteristics of the battery in real time by a filter based on the battery model.

[0056] The main control unit of the battery management system periodically collects sensor data through a hardware interface to form the real-time operating data. Specifically, the current data is measured by deploying a high-precision shunt or a Hall sensor in the main loop of the battery pack; the terminal voltage data of the battery pack is collected by a resistance voltage dividing network; and the temperature data is obtained by attaching a thermistor to the surface of the battery cell. After the collected analog signals are processed by an analog-to-digital converter, a digital information stream is formed for the algorithm to use.

[0057] The current data accurately represents the charge rate flowing into or out of the battery, which is the input quantity for the ampere-hour integration method to calculate the change in charge. The voltage data is the manifestation of the internal electrochemical state of the battery on the external terminal, which provides an observation value for the filter based on the battery model, used to correct and constrain the estimation results of the internal state. The temperature data is used to compensate the internal parameters of the battery model and the coulomb efficiency of the ampere-hour integration method in real time, by modifying the model parameters to adapt to the battery characteristics at different working temperatures.

[0058] The ampere-hour integration method is an estimation technique that accumulates the total amount of charge flowing into or out of the battery by continuously integrating the current over time, thereby calculating the amount of state-of-charge change. The reference SOC value is an estimation result obtained by the ampere-hour integration method, mainly reflecting the long-term cumulative change in the amount of charge, and is used as the main reference for SOC change in the present method. The filter based on the battery model is an algorithm that describes the electrochemical characteristics of the battery using mathematical equations, and estimates the internal state of the battery by fusing the current information as the model input and the voltage information as the model observation. The dynamic SOC value is the output result of the filter based on the battery model, and its value can quickly respond to real-time changes in the battery terminal voltage, thereby sensitively reflecting fluctuations in battery characteristics caused by load dynamics.

[0059] Specifically, in an embodiment, the S1 includes steps S11-S13.

[0060] S11. Based on the reference SOC value of the previous cycle, the current in the real-time operation data of the current cycle is time-integrated, and the rated capacity and coulomb efficiency of the battery are combined to calculate an updated reference SOC value;

[0061] S12. An equivalent circuit model of the battery is established as a state space equation, and the current in the real-time operation data is used as the input of the state space equation to obtain a predicted SOC value;

[0062] S13. The voltage in the real-time operation data is used as the observation value of the state space equation, and the predicted SOC value is corrected by comparing the error between the observation value and the predicted output voltage of the equivalent circuit model to determine the dynamic SOC value.

[0063] The coulomb efficiency refers to the ratio of the discharged charge to the charged charge in the charge and discharge cycle of the battery, and is used to quantify the loss of electric quantity in electrochemical reactions. The equivalent circuit model is a circuit network constructed using basic electronic components such as resistors and capacitors, which is used to simulate the voltage response characteristics of the battery under different working conditions in a macroscopic manner. The state space equation is a mathematical form describing the dynamic behavior of a physical system, which expresses the internal relationship of the system through a set of first-order differential equations containing state variables, inputs, and outputs. The predicted SOC value is a priori estimation made by the filter based on the state of the system at the previous time and the current input, while the observation value is a physical quantity related to the internal state that can be directly measured in the system.

[0064] In S11, the master unit obtains the reference SOC value of the last calculation period, and integrates the current data collected in the current period to calculate the charge variation. By combining the pre-calibrated battery capacity and the coulomb efficiency related to the current temperature and current direction, the charge variation is normalized and compensated to obtain the updated reference SOC value. In the example above, in one operating period, if the reference SOC value of the last period is 50.5%, the rated capacity is 100 ampere-hours, the measured charging current in the current period is 100 amperes, the period length is 0.1 seconds, and the coulomb efficiency under the current working condition is 0.99, the system updates the reference SOC value by calculating the effective charge amount charged in the period.

[0065] In S12, the algorithm converts the equivalent circuit model into a state space equation, uses real-time current as the system input, and makes prior prediction of the state variables including SOC. Continuing the example above, based on the initial state of 50.5% and the charging current of 100 amperes, the system predicts that the SOC value of the next period is 50.527%, and the predicted terminal voltage, i.e. the predicted output voltage, is 350.2 volts. In step S13, the algorithm takes the actually measured battery terminal voltage, for example 350.5 volts, as the observation value. By calculating the 0.3 volt error between the observation value and the model predicted output voltage, and combining the gain matrix inside the filtering algorithm, the prior predicted SOC value is post-processed to finally determine a dynamic SOC value that matches the actual voltage response better, for example 50.531%.

[0066] S2. Establish a hierarchical calibration arbitrator containing multiple calibration modes, which judges and selects the highest priority calibration mode that meets the trigger condition from high to low according to the preset priority based on the reference SOC value, the dynamic SOC value, and the real-time operating data, to generate a calibrated SOC value. The calibration modes of the hierarchical calibration arbitrator include: absolute zero-point calibration mode, floating anchor point calibration mode, quasi-static voltage calibration mode, and on-demand dynamic impedance calibration mode.

[0067] The hierarchical calibration arbitrator is a set of decision logic algorithms implemented in the master unit of the battery management system, aiming to solve the technical limitations of single calibration method that cannot adapt to working conditions. By establishing a priority system based on data reliability and intervention initiative, the arbitrator can dynamically select the optimal one from multiple calibration modes according to real-time working conditions, so as to find effective calibration opportunities in various complex application scenarios. The hardware principle is rooted in the master unit of the battery management system, such as a microcontroller MCU, which executes the arbitrator algorithm instructions contained in the firmware, makes logical judgments on the input SOC estimation value and real-time data, and outputs the decision result.

[0068] The four calibration modes are set to build a full-scenario calibration system covering from ideal boundary conditions to sustained dynamic working conditions. The absolute zero-point calibration mode uses the physical boundaries of full charge or full discharge to provide a calibration reference based on physical boundaries, and is suitable for working conditions with complete charge and discharge cycles. The floating anchor calibration mode solves the problem of not being able to reach the boundary point when running for a long time in the partial capacity interval by identifying electrochemical characteristics in the stable charging and discharging process. The quasi-static voltage calibration mode provides a model-based regular calibration opportunity for intermittent running working conditions. The on-demand dynamic impedance calibration mode as an active intervention means ensures that the system can still enforce calibration to prevent the accumulation of errors in sustained dynamic applications without any passive calibration opportunities.

[0069] Specifically, in an embodiment, the S2 includes steps S21-S26.

[0070] S21. With the highest priority, it is judged whether the real-time running parameters of the battery reach a preset charge and discharge physical boundary condition, and if so, the absolute zero-point calibration mode is selected.

[0071] S22. If the physical boundary condition is not reached, with the second highest priority, it is judged whether a preset electrochemical characteristic representing the internal state of the battery can be identified in the running data of the battery, and if so, the floating anchor calibration mode is selected.

[0072] S23. If the electrochemical characteristic cannot be identified, with the third priority, it is judged whether the battery meets a preset quasi-static state defined by the combination of duration and current threshold, and if so, the quasi-static voltage calibration mode is selected.

[0073] S24. In parallel with S21-S23, a counter is maintained to track the cumulative running amount since the last successful calibration.

[0074] S25. With the lowest priority, it is judged whether the value of the counter reaches a preset threshold, and if so, and none of the aforementioned physical boundary condition, electrochemical characteristic and quasi-static state is met in the current running period, the on-demand dynamic impedance calibration mode is selected, wherein the on-demand dynamic impedance calibration mode is characterized by actively injecting a probe signal into the battery for calibration.

[0075] S26. After any calibration mode is selected and successfully executed, the counter is reset.

[0076] Real-time operating parameters are physical quantities measured directly by sensors without deep processing, mainly including the terminal voltage and the loop current of the battery. The charge-discharge physical boundary condition is a combination of parameter thresholds predefined by the system, indicating that the battery has reached a full charge or a full discharge state. The electrochemical feature is a specific pattern caused by the phase change of the electrode material in the battery during the charge-discharge process, which can be stably identified on the voltage or its derivative curve. The quasi-static state is a state in which the load current of the battery remains in a small range close to zero for a certain duration. The counter is a software timer or an electric quantity accumulator, which tracks the cumulative operating quantity, which can be the operating time since the last calibration or the cumulative ampere-hour processed.

[0077] Steps S21 to S23 prioritize the triggering conditions of passive calibration modes, which rely on specific states naturally reached by the battery without intervention to the system. Only when all passive calibration opportunities do not occur within a cumulative operating quantity, step S24 triggers the active intervention calibration mode to ensure that the error is finally suppressed.

[0078] The preset charge-discharge physical boundary condition specifically defines the full charge and full discharge states of the battery. In an embodiment, the boundary condition of the full charge state is defined as: in the constant voltage charging stage, the total voltage of the battery pack stabilizes at the charging cutoff voltage, while the charging current decreases and continues to be less than a first current threshold, for example, two percent of the rated capacity. The boundary condition of the full discharge state is defined as: in the discharging process, the total voltage of the battery pack decreases to reach the discharge cutoff voltage threshold. In the above example of an electric vehicle, when the vehicle is charging at a charging pile, the battery pack voltage stabilizes at 400 volts and the charging current decreases to less than 5 amperes, which reaches the full charge physical boundary condition.

[0079] The role of the counter is to quantify the risk of inaccuracy of the reference SOC value and serve as the basis for triggering the lowest priority calibration mode. By resetting the counter after each successful calibration, the counter starts tracking the cumulative operating quantity. When the cumulative operating quantity reaches a preset threshold, for example, 24 hours of cumulative operation or 5 equivalent full cycles of cumulative completion, it indicates that the reference SOC value may have accumulated a large error. At this time, if there is no other higher priority calibration opportunity, the judgment condition of step S24 is met, thereby forcibly starting the on-demand dynamic impedance calibration to eliminate the possible significant error. After calibration, step S25 performs resetting to start a new round of tracking, forming a closed-loop control.

[0080] Specifically, in an embodiment, the step of selecting the absolute zero point calibration mode to generate the calibrated SOC value includes S211-S213.

[0081] S211. Monitor whether the battery reaches a preset full charge state or a preset full discharge state.

[0082] S212. When the battery reaches the full charge state, the calibration SOC value is generated as 100%.

[0083] S213. When the battery reaches the full discharge state, the calibration SOC value is generated as a preset minimum safe SOC value.

[0084] Full charge state indicates a state that the battery has received charge up to its current maximum capacity. Full discharge state indicates a state that the battery has discharged down to its safe discharge lower limit. Minimum safe SOC value is a non-zero charge threshold, for example 5%, set to prevent irreversible damage to the battery due to over-discharge.

[0085] The definition of preset full charge state or full discharge state is based on the judgment of battery charge cutoff condition and discharge cutoff condition. In one embodiment, full charge state is defined as: the battery is in constant voltage charging phase, its charging current is less than a first preset current threshold, for example 2% of rated capacity, and lasts more than a first preset time length. Full discharge state is defined as: the battery is in discharging process, its terminal voltage reaches and lasts below a preset discharge cutoff voltage under load condition.

[0086] The principle of this calibration mode is to eliminate the cumulative error of SOC estimation by using the physical operating boundaries determined by the battery. The battery management system continuously monitors voltage and current in S211. In the above electric vehicle example, when the vehicle is charging at night, the system monitors that the battery pack enters the constant voltage charging phase and the charging current drops below 5 amperes and maintains for 5 minutes, then it is judged that the battery has reached the full charge state. In S212, the system generates a calibration SOC value determined as 100%. Conversely, if the vehicle continues to travel, the system monitors that the battery pack voltage reaches the 300-volt discharge cutoff voltage under load, it is judged that the battery has reached the full discharge state, then a minimum safe SOC value of 5%, for example, is generated in S213.

[0087] Specifically, in one embodiment, the step of selecting the floating anchor calibration mode to generate the calibration SOC value includes S221-S223.

[0088] S221. Determine whether the battery is in a continuous and current stable charging or discharging process.

[0089] S222. In the process, an online voltage change rate characteristic curve is calculated based on the trajectory of voltage change with cumulative charge.

[0090] S223. One or more preset characteristic peaks are identified from the characteristic curve, and the current SOC value is determined as the calibration SOC value according to a pre-stored correspondence relationship between the characteristic peaks and SOC.

[0091] The voltage rate-of-change characteristic curve, also technically known as incremental capacity analysis curve or dQ / dV curve, refers to the rate curve of the battery terminal voltage changing with the amount of charge being charged or discharged. The characteristic peak is a local extremum point appearing on the voltage rate-of-change characteristic curve, corresponding to the phase transition of the battery electrode material at a specific electrochemical potential.

[0092] The criterion for judging a continuous and current-smooth charging or discharging process is, in one embodiment, defined as: the duration of the battery's charging and discharging process exceeds a second preset time length, for example 20 minutes, and the fluctuation amplitude of the current during this period is less than a certain percentage of its average value, for example 10%. The preset characteristic peak is a voltage rate-of-change peak value with a stable position and shape, which is obtained by offline experiment calibration according to the battery design model. For example, a lithium nickel-manganese-cobalt oxide battery will stably appear a peak value with a specific shape when the SOC reaches 55% in an environment of 25 degrees Celsius.

[0093] The calibration mode calibrates an accurate anchor point in the middle SOC by capturing the intrinsic electrochemical fingerprint of the battery which does not easily change with use. In the context of the electric vehicle example, when the vehicle is stably charged at a current of 32 amperes for more than 20 minutes on a public alternating current charging pile, the judgment condition of S221 is satisfied. The system then starts recording the change of voltage with the amount of charge in step S222, and calculates the voltage rate-of-change characteristic curve in real time. When the system identifies a peak value on the curve that matches the 55% SOC anchor point characteristic peak pre-stored in the memory, S223 is triggered, and the system generates a calibration SOC value determined as 55%. In this way, the problem of the vehicle being unable to obtain absolute zero calibration due to long-term use in the partial SOC interval is solved.

[0094] Specifically, in one embodiment, the step of selecting the quasi-static voltage calibration mode to generate the calibration SOC value includes S231-S232.

[0095] S231. Determine whether the battery is in a quasi-static state in which the absolute value of its current is less than a preset threshold.

[0096] S232. When the battery is in the quasi-static state, use the dynamic SOC value output by the battery model-based filter as the calibration SOC value.

[0097] The quasi-static state refers to a stable state in which the external load current of the battery remains at a sufficiently low level for a certain duration, so that the electrochemical polarization phenomenon inside the battery is significantly alleviated, thereby enabling its terminal voltage to converge and approach its true open circuit voltage.

[0098] The definition of the preset threshold value includes both the current amplitude and the duration. In one embodiment, the threshold value is defined as: the absolute current value of the battery is less than one percent of its rated capacity, for example, for a 100 ampere-hour battery, the current is less than 1 ampere; and the duration of the low current state exceeds a preset time, for example, 5 minutes.

[0099] The calibration mode takes advantage of the feature that the dynamic SOC estimate will converge to the true value under static conditions. In S231, the system continuously monitors the current. In the context of the electric vehicle example, the vehicle is parked in a destination parking lot for 10 minutes, during which only weak static power consumption occurs. The system determines that this working condition meets the quasi-static state condition of an absolute current value less than 1 ampere and a duration of more than 5 minutes. During this period, the filter based on the battery model will correct its internal state only according to the voltage observation value, because there is no dynamic current input, so that the dynamic SOC value as its output gradually converges to a stable value reflecting the current open circuit voltage. In S232, the system accepts this dynamically stable SOC value, for example, 45.3%, and outputs it as the calibration SOC value.

[0100] Specifically, in some embodiments, the step of selecting the on-demand dynamic impedance calibration mode to generate the calibration SOC value includes S241-S244.

[0101] S241. Determine whether other higher priority calibration modes have failed to be triggered within a preset time period or a preset cumulative charge period.

[0102] S242. If not, actively request a power conversion system connected to the battery to superimpose a detection current signal.

[0103] S243. Calculate a dynamic impedance characteristic of the battery according to its voltage response to the detection current signal.

[0104] S244. Determine the current SOC value based on the dynamic impedance characteristic and a pre-stored correspondence between the dynamic impedance characteristic and the SOC, and output it as the calibration SOC value.

[0105] The power conversion system refers to a power electronic device that controls the charging or discharging of the battery, such as an on-board charger or an inverter. The detection current signal is a small perturbation current with a specific waveform, such as a sine wave or a square wave, generated by the power conversion system according to the instructions of the battery management system. The dynamic impedance characteristic is the complex impedance exhibited by the battery at a specific frequency, which is a key indicator for evaluating the internal electrochemical state of the battery.

[0106] The active request of superimposed probing current signal in S242 is because the system cannot wait for the battery to naturally enter a certain stable state which is easy to analyze when the battery is continuously in dynamic working condition. By actively applying a probing signal with a known waveform, the system can create a controllable measurement condition in a complex operating environment.

[0107] The calculation of dynamic impedance characteristics in S243 is to obtain an electrochemical fingerprint which can establish a stable mapping relationship with SOC. The dynamic impedance of the battery is closely related to the ion concentration and electrode state inside the battery, which is directly related to SOC. By calculating the voltage response of the battery to the known probing current, the dynamic impedance can be accurately solved, thereby obtaining a SOC calculation basis independent of open circuit voltage.

[0108] The active probing method can be used as the final calibration method when there is no opportunity for passive calibration. Based on the above example of electric vehicles, assume that the vehicle is used for vehicle-to-grid (V2G) applications and continuously charges and discharges under grid instructions for 48 hours without meeting any higher-priority calibration conditions. At this time, the judgment condition of S241 is met. In S242, the system sends instructions to the vehicle's power conversion system through the vehicle's communication network, requesting it to superimpose a small sinusoidal current with a frequency of 1 Hz on the main current for 10 seconds. In S243, the system synchronously collects the voltage response and calculates the dynamic impedance characteristic value of the battery at 1 Hz. Finally, in S244, the system queries the pre-existing impedance-SOC correspondence graph in the memory, and obtains the current SOC as 42.8% from the impedance value, and outputs it as the calibrated SOC value.

[0109] S3. According to the source mode of the calibrated SOC value and its corresponding confidence level, a pre-set correction strategy is adopted to correct the reference SOC value using the calibrated SOC value to obtain a final SOC estimation value.

[0110] The confidence level is a reliability evaluation of the calibrated SOC value, which is directly related to the specific calibration mode that generates the calibrated SOC value. In this method, different calibration modes have different intrinsic confidence levels due to their different principles, for example, the absolute zero point calibration mode based on physical boundaries has the highest confidence level, while the calibration mode based on the model has relatively lower confidence level.

[0111] The correction strategy is a set of specific algorithms for performing calibration actions. The reason for correction is that the reference SOC value, which is a long-term cumulative reference for electric quantity, will produce cumulative errors due to sensor deviation and other factors. That is, the correction strategy uses the calibrated SOC value with higher confidence level to periodically correct the drift of the reference SOC value, ensuring the long-term estimation accuracy of the method.

[0112] Specifically, in some embodiments, the S3 comprises steps S31-S33.

[0113] S31. Determine which specific calibration mode in the hierarchical calibration arbitrator generates the calibration SOC value.

[0114] S32. If the calibration SOC value is derived from the absolute zero-point calibration mode, instantaneously reset the reference SOC value to the calibration SOC value as the final SOC estimate.

[0115] S33. If the calibration SOC value is derived from a calibration mode other than the absolute zero-point calibration mode, gradually and smoothly modify the reference SOC value towards the calibration SOC value over a pre-set period of time or number of operating cycles using a pre-set convergence algorithm, and output the reference SOC value during the dynamic modification process as the final SOC estimate in real time.

[0116] Instantaneous reset is a direct assignment operation that immediately overwrites the current reference SOC value with the calibration SOC value. Smooth convergence is a gradual modification process designed to avoid abrupt changes in the SOC estimate. Convergence algorithm is a specific mathematical method to achieve smooth convergence, an example of which is the first-order lag filter algorithm. In this algorithm, the modified reference SOC value for each cycle is obtained by weighted averaging of the reference SOC value of the previous cycle and the current calibration SOC value. By setting a small weight factor, the reference SOC value can be smoothly approached to the calibration SOC value over multiple cycles, and the convergence speed is determined by the size of the weight factor.

[0117] The principle of this modification strategy is to take different intensities of modification actions according to the confidence level of the calibration event, in order to balance the accuracy of calibration and the smoothness of output. The system first identifies the source of the calibration value. In the electric vehicle example above, if the vehicle has completed charging, the arbitrator triggers the absolute zero-point calibration mode and generates a calibration SOC value of 100%. Since this value has the highest confidence level, the system performs instantaneous reset in S32, directly and immediately updating the reference SOC value that may have errors, for example 98.7%, to 100%, and using it as the final SOC estimate.

[0118] In another scenario, if the vehicle is parked in a parking lot for 10 minutes, the arbitrator triggers the quasi-static voltage calibration mode and generates a calibration SOC value of 45.3%. At this time, the reference SOC value is 43.1%. Since the calibration value has high confidence but does not belong to the physical boundary, the system performs the smoothing convergence strategy in step S33. Through the convergence algorithm, the reference SOC value is gradually modified from 43.1% to 45.3% in the next few operating cycles. The final SOC estimation value output externally will show a continuous process from 43.1% to 45.3%, thereby avoiding the sudden change in the range display caused by the jump in the SOC value.

[0119] It should be understood that the size of the serial number of each step in the above embodiments does not mean the order of execution. The execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0120] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the above-mentioned division of each functional unit and module is exemplified. In actual application, the above-mentioned functions can be completed by different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above.

[0121] The above embodiments are only used to illustrate the technical solutions of the present application, but not limit it. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features. The modification or replacement does not make the essence of the corresponding technical solution deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method of estimating a state of charge of a battery, characterized by, The method comprises the following steps: S1. Real-time operation data of voltage, current and temperature of the battery is obtained, and a reference SOC value as a long-term power accumulation reference is obtained in parallel through ampere-hour integration based on the real-time operation data, and a dynamic SOC value reflecting the current voltage characteristics of the battery in real time is obtained through a filter based on a battery model; S2. A hierarchical calibration arbitrator comprising multiple calibration modes is established, the hierarchical calibration arbitrator judges and selects a highest priority calibration mode whose trigger condition is met according to a preset priority from high to low based on the reference SOC value, the dynamic SOC value and the real-time operation data, to generate a calibration SOC value; wherein the calibration modes of the hierarchical calibration arbitrator include: absolute zero point calibration mode, floating anchor point calibration mode, quasi-static voltage calibration mode, on-demand dynamic impedance calibration mode; S3. According to the source mode of the calibration SOC value and the corresponding confidence, a preset correction strategy is adopted to correct the reference SOC value using the calibration SOC value to obtain a final SOC estimation value; The S2 comprises the following steps: S21. With the highest priority, it is judged whether the real-time operation parameters of the battery reach a preset charge-discharge physical boundary condition, if so, the absolute zero point calibration mode is selected; S22. If the physical boundary condition is not reached, with the second highest priority, it is judged whether a preset electrochemical feature representing the internal state of the battery can be identified in the operation data of the battery, if so, the floating anchor point calibration mode is selected; S23. If the electrochemical feature cannot be identified, with the third highest priority, it is judged whether the battery meets a preset quasi-static state defined by the duration and current threshold, if so, the quasi-static voltage calibration mode is selected; S24. In parallel with S21-S23, a counter for tracking the cumulative operation amount since the last successful calibration is maintained; S25. With the lowest priority, it is judged whether the value of the counter reaches a preset threshold, if so, and the aforementioned physical boundary condition, electrochemical feature and quasi-static state are not met in the current operation period, the on-demand dynamic impedance calibration mode is selected, wherein the on-demand dynamic impedance calibration mode is characterized by actively injecting a detection signal into the battery for calibration; S26. After any calibration mode is selected and successfully executed, the counter is reset.

2. The method of estimating a state of charge of a battery according to claim 1, characterized by, The S1 comprises the following steps: S11. Based on the reference SOC value of the last period, the updated reference SOC value is calculated by time integrating the current in the real-time operation data of the current period, and combining the rated capacity and coulomb efficiency of the battery; S12. An equivalent circuit model of the battery is established as a state space equation, and the current in the real-time operation data is used as the input of the state space equation to obtain a predicted SOC value; S13. Using the voltage in real-time running data as an observation value of the state space equation, the predicted SOC value is corrected by comparing the error between the observation value and the predicted output voltage of the equivalent circuit model to determine the dynamic SOC value.

3. The method of estimating the state of charge of a battery according to claim 2, characterized by, The S3 includes the following steps: S31. Determine which specific calibration mode in the hierarchical calibration arbitrator generates the calibration SOC value; S32. If the calibration SOC value is derived from the absolute zero-point calibration mode, reset the reference SOC value to the calibration SOC value instantaneously as the final SOC estimation value; S33. If the calibration SOC value is derived from a calibration mode other than the absolute zero-point calibration mode, gradually and smoothly correct the reference SOC value to the calibration SOC value within a preset period of time or multiple operating cycles through a preset convergence algorithm, and output the reference SOC value during the dynamic correction process as the final SOC estimation value in real time.

4. The method of estimating a state of charge of a battery according to claim 1, characterized by, The steps of selecting the absolute zero-point calibration mode to generate the calibration SOC value include: S211. Monitor whether the battery reaches a preset full charge state or a preset full discharge state; S212. When the battery reaches the full charge state, generate the calibration SOC value as 100%; S213. When the battery reaches the full discharge state, generate the calibration SOC value as a preset minimum safe SOC value.

5. The method of estimating the state of charge of a battery according to claim 4, characterized by, The steps of selecting the floating anchor point calibration mode to generate the calibration SOC value include: S221. Determine whether the battery is in a continuous and current-stable charging or discharging process; S222. During the process, calculate an online voltage change rate characteristic curve based on the change trajectory of the voltage with respect to the cumulative charge; S223. Identify one or more preset characteristic peaks from the characteristic curve, and determine the current SOC value according to a pre-stored correspondence between the characteristic peaks and SOC, and take it as the calibration SOC value.

6. The method of estimating the state of charge of a battery according to claim 5, wherein, The steps of selecting the quasi-static voltage calibration mode to generate the calibration SOC value include: S231. Determine whether the battery is in a quasi-static state where the absolute value of its current is less than a preset threshold; S232. When the battery is in the quasi-static state, take the dynamic SOC value output by the filter based on the battery model as the calibration SOC value.

7. The method of estimating the state of charge of a battery according to claim 6, wherein, The steps of selecting the on-demand dynamic impedance calibration mode to generate the calibration SOC value include: S241. Determine whether other calibration modes of higher priority have failed to be triggered within a preset period of time or a preset cumulative charge period; S242. If not, actively request a superimposed detection current signal from a power conversion system connected to the battery; S243. Calculate a dynamic impedance characteristic of the battery according to its voltage response to the detection current signal; S244. Determine the current SOC value according to a pre-stored correspondence between the dynamic impedance characteristic and SOC, and take it as the calibration SOC value.

Images