Lithium battery health state evaluation method and system

Abstract

The invention discloses a lithium battery health state assessment method and system, and the method comprises the steps: collecting the terminal voltage data and current data of a lithium battery at each moment in a discharge period in real time, so as to obtain a DC impedance estimation value at each moment; obtaining a discharge working condition of the lithium battery at each moment to screen a plurality of direct current impedance effective values, and constructing a plurality of direct current impedance value real-time sequences; calculating a direct-current impedance sequence estimation value of each direct-current impedance value real-time sequence; segmenting the plurality of direct-current impedance sequence estimated values and a preset discharge direct-current impedance data table according to the discharge working condition, and calculating a direct-current impedance average value and a direct-current impedance reference average value in each segment, and determining the health state of the lithium battery according to the DC impedance average value and the DC impedance reference average value, thereby realizing online evaluation of the health state of the lithium battery.

Description

Technical Field

[0001] The present invention relates to the technical field of battery health state estimation, and in particular to a method and system for evaluating the health state of a lithium battery. Background Art

[0002] With the development of science and technology, lithium batteries have been widely used in energy storage fields such as photovoltaics, rail transit, new energy vehicles and communication base stations. At the same time, the health status of the battery characterizes the degree of battery aging, which is of great significance to improving the accuracy of lithium battery state estimation, reducing the risk of battery thermal runaway, and the maintenance and residual value assessment of lithium batteries. How to improve the accuracy of online estimation of the battery's state of health (State Of Health, SOH) has become the core of the battery management system (Battery Management System, BMS). In the prior art, the SOH of on-board lithium batteries is generally evaluated by two aspects: capacity attenuation or internal resistance increase. Among them, the ratio of the current estimated capacity to the maximum available capacity at the time of leaving the factory is used to characterize the SOH, which is defined as SOHQ and is often used in pure electric vehicles that pay more attention to the cruising range; while the DC impedance (Direct Current Resistance, DCR) growth rate is used to characterize the SOH, which is defined as SOHR and is often used in hybrid vehicles that pay more attention to power.

[0003] At present, the mainstream SOHR estimation methods include DC impedance method, AC impedance method and model parameter identification method. In the DC impedance method, a large DC is loaded on the lithium battery as input excitation. It is usually necessary to use the Hybrid Pulse Power Characteristic (HPPC) to obtain its instantaneous voltage and current changes. The DC impedance of the lithium battery is obtained through the ratio of the changes. In offline bench tests, a more accurate DC impedance value can be obtained. However, HPPC is mainly used in laboratory conditions, so it can be used to obtain a reference value of DCR. In real working conditions, there is not enough static working condition to eliminate the polarization effect, which affects the accuracy of the DC impedance DCR; the AC impedance method loads a fixed frequency AC small current on the battery as input excitation, monitors the output response of its terminal voltage, and obtains the ohmic resistance of the lithium battery, the diffusion resistance of the SEI film, and the capacitance value of the SEI film from the AC electrochemical impedance spectroscopy (EIS). This method overcomes many shortcomings of the DC impedance method and can accurately measure the corresponding parameter information. At the same time, due to the small input current, it has almost no damage to the battery. However, the AC impedance method can only be used under non-working conditions and cannot realize the internal resistance measurement of the battery during operation.

[0004] Direct current impedance (DCR) plays an irreplaceable role in estimating the state of health (SOHR) of lithium batteries, estimating the battery power state and diagnosing battery faults. Currently, the DC impedance method is mostly used for DCR estimation. This method cannot obtain relatively accurate and independent ohmic internal resistance and polarization internal resistance, resulting in inaccurate DCR calculation. In addition, the experimental DC impedance method and AC impedance method both require strict test conditions and complex test equipment, and have a long test cycle, and cannot be directly applied to the online estimation of lithium battery SOHR. Summary of the invention

[0005] In order to solve the above technical problems, the present invention discloses a method and system for evaluating the health status of a lithium battery, which are used to estimate the health status of a lithium battery online and improve the accuracy of the estimation.

[0006] In order to achieve the above object, the present invention discloses a method for evaluating the health status of a lithium battery, comprising:

[0007] During a discharge period, terminal voltage data and current data of the lithium battery at each moment are collected in real time to obtain an estimated DC impedance value of the lithium battery at each moment;

[0008] Obtaining the discharge condition of the lithium battery at each moment, and selecting a plurality of DC impedance effective values ​​from the plurality of DC impedance estimation values ​​according to the discharge condition; wherein the DC impedance effective value includes a first DC impedance effective value corresponding to each of the first plurality of moments and a second DC impedance effective value corresponding to each of the last plurality of moments;

[0009] Collecting each of the first DC impedance effective values ​​in real time to construct an initial sequence of DC impedance values ​​corresponding to a preset time period;

[0010] Iteratively updating the initial sequence of DC impedance values ​​in real time according to each of the second DC impedance effective values, and constructing a real-time sequence of DC impedance values ​​corresponding to each of the second DC impedance effective values;

[0011] Calculating a DC impedance sequence estimation value of each of the DC impedance value real-time sequences;

[0012] According to the discharge working condition, the plurality of DC impedance sequence estimation values ​​and the preset discharge DC impedance data table are segmented, and the DC impedance average value and the DC impedance reference average value in each segment are calculated; wherein the discharge DC impedance data table is used to indicate the relationship between the state of charge value and the DC impedance value at different temperatures;

[0013] The health status of the lithium battery is determined according to a plurality of the DC impedance average values ​​and a plurality of the DC impedance reference average values.

[0014] The present invention discloses a method for evaluating the health status of a lithium battery. First, the DC impedance is estimated by collecting the terminal voltage data and current data corresponding to each moment during the discharge period of the lithium battery, so as to realize online detection of the lithium battery when the lithium battery is in a working state. Secondly, the discharge condition corresponding to the lithium battery at each moment is obtained, and the estimated DC impedance estimation value is screened based on the discharge condition, so as to eliminate invalid data according to the screening, thereby improving the efficiency of health status assessment. Furthermore, each effective value of the DC impedance obtained by the screening is collected in real time, and an initial sequence of DC impedance values ​​is first constructed, and then each of the effective values ​​of the DC impedance obtained after the initial sequence is constructed is used to iteratively update the initial sequence of DC impedance values ​​to obtain a real-time sequence of DC impedance values ​​corresponding to each moment, so as to facilitate the estimation of the DC impedance sequence estimation value corresponding to each moment according to the real-time sequence of DC impedance values, and then the accuracy of the DC impedance sequence estimation value is ensured through stable sequence data. Finally, the obtained continuous multiple DC impedance sequence estimation values ​​are segmented using the discharge conditions, and the preset discharge DC impedance data table is segmented based on the discharge conditions, and then the corresponding DC impedance average value and DC impedance reference average value in each same segment are calculated, and then the health status of the lithium battery is determined based on the DC impedance average value and the DC impedance reference average value, thereby realizing online evaluation of the health status of the lithium battery.

[0015] As a preferred example, the real-time acquisition of the terminal voltage data and current data of the lithium battery at each moment to obtain the DC impedance estimation value of the lithium battery at each moment includes:

[0016] Constructing a second-order RC equivalent circuit model, and obtaining the current terminal voltage value and current value of the lithium battery at the current moment and the historical terminal voltage value and historical current value at the previous moment;

[0017] Obtain the state of charge value of the lithium battery at the current moment, and query the preset SOC-U according to the state of charge value OCV Relationship table, obtaining the open circuit voltage value of the lithium battery at the current moment;

[0018] The difference between the current terminal voltage value and the open-circuit voltage value is calculated, and the parameters of the second-order RC equivalent circuit model are identified by a recursive least squares method based on an adaptive forgetting factor according to the current terminal voltage value, the current current value, the historical terminal voltage value, the historical current value and the difference, so as to obtain the ohmic internal resistance, high-frequency polarization internal resistance and low-frequency polarization internal resistance of the lithium battery at the current moment.

[0019] The present invention constructs the second-order RC equivalent circuit model, and based on the real-time voltage data and current data of the battery, uses the recursive least squares method based on the adaptive forgetting factor to perform parameter identification on the circuit model, so as to accurately and independently obtain the ohmic internal resistance, high-frequency polarization internal resistance and low-frequency polarization internal resistance of the lithium battery, and then improve the accuracy of subsequent DC impedance value estimation according to the accurate ohmic internal resistance, high-frequency polarization internal resistance and low-frequency polarization internal resistance.

[0020] As a preferred example, the step of obtaining the estimated DC impedance value of the lithium battery at each moment includes:

[0021] According to the ohmic internal resistance, high-frequency polarization internal resistance and low-frequency polarization internal resistance, the DC impedance estimation value corresponding to the lithium battery at the current moment is estimated by the discrete time observation equation corresponding to the second-order RC equivalent circuit model; wherein the calculation expression of the DC impedance estimation value is:

[0022] R DCR =R0+R f (1-e (-Ts / τf) )+R s (1-e (-Ts / τs) )

[0023] Among them, the R DCR represents the DC impedance estimation value corresponding to the lithium battery at each moment; R0 represents the ohmic internal resistance of the lithium battery; R f represents the concentration polarization resistance of the lithium battery; the R s represents the battery electrochemical polarization internal resistance of the lithium battery; the R f (1-e (-Ts / τf) ) represents the high frequency polarization impedance; the R s (1-e (-Ts / τs) ) represents the low-frequency polarization impedance; the τ f and τ s represents a time constant; and Ts represents a duration.

[0024] The present invention calculates the DC impedance value at each moment based on the parameters obtained by identifying the model and the discrete equation of the second-order equivalent circuit model, thereby realizing online estimation of the DC impedance value.

[0025] As a preferred example, the step of selecting a plurality of effective DC impedance values ​​from a plurality of DC impedance estimation values ​​according to the discharge condition includes:

[0026] Obtaining the state of charge value, temperature value and charge and discharge state of the lithium battery corresponding to each moment;

[0027] When the state of charge value is in a preset state of charge range, the temperature value is in a preset temperature range, and the charge-discharge state is a discharge state, the DC impedance estimation value corresponding to the current moment is determined as the DC impedance effective value.

[0028] The present invention sets a valid data range by setting a state of charge interval, a temperature interval and limiting the lithium battery to be in a discharge state, and then combines the valid data range and the real-time state of charge value, real-time temperature value and real-time charge and discharge state corresponding to the lithium battery at each moment to screen each DC impedance estimation value to obtain a stable DC impedance value, thereby improving the accuracy of lithium battery health status assessment based on stable data.

[0029] As a preferred example, the real-time acquisition of each of the first DC impedance effective values ​​to construct an initial sequence of DC impedance values ​​includes:

[0030] Each of the first DC impedance effective values ​​collected in real time is sequentially saved in a cache array set with a preset dimension to construct an initial sequence of DC impedance values ​​corresponding to the preset dimension.

[0031] The present invention uses a cache array of preset dimensions to store each of the DC impedance effective values ​​estimated in real time, and completes the initial data storage to construct the initial sequence of DC impedance values ​​when the number of stored data reaches the dimension of the cache array. The initial sequence of DC impedance values ​​is then iteratively updated according to each of the DC impedance effective values ​​collected in real time, thereby ensuring that the DC impedance effective values ​​collected at each moment have a corresponding real-time sequence of DC impedance values, so that a stable DC impedance value can be obtained later according to the sequence.

[0032] As a preferred example, the step of calculating the DC impedance sequence estimation value of each of the DC impedance value real-time sequences includes:

[0033] Calculating the mean value and standard deviation of each of the real-time series of DC impedance values, and constructing an abnormal value screening interval for each of the real-time series of DC impedance values ​​according to the mean value and the standard deviation;

[0034] Comparing each of the DC impedance effective values ​​in the DC impedance value real-time sequence with the corresponding abnormal value screening interval, so as to detect abnormal values ​​in the DC impedance value real-time sequence according to the comparison result;

[0035] When the DC impedance effective value is not within the abnormal value screening interval, determining that the current DC impedance effective value is a DC impedance abnormal value, and replacing the DC impedance abnormal value with the average value;

[0036] After replacing the abnormal DC impedance value with the average value, sorting each of the real-time sequences of DC impedance values;

[0037] The intermediate value in the sorted real-time sequence of DC impedance values ​​is obtained, and the intermediate value is determined as the DC impedance sequence estimation value.

[0038] The present invention screens outliers in the sequence data by calculating the mean value and the standard deviation, thereby preventing the outliers from affecting the DCR estimation accuracy. Then, after removing the outliers in the sequence, the sequence is sorted, so that the middle value of the sequence is used as the stable DC impedance estimation value at the current moment, so as to improve the accuracy of the DC impedance estimation value.

[0039] As a preferred example, the method of segmenting the plurality of DC impedance sequence estimation values ​​and the preset discharge DC impedance data table according to the discharge working condition, and calculating the DC impedance average value and the DC impedance reference average value in each segment, includes:

[0040] According to the discharge working condition, a state of charge value and a temperature value corresponding to each of the DC impedance sequence estimation values ​​are obtained;

[0041] According to the state of charge value, segmenting the plurality of DC impedance sequence estimation values ​​at intervals of a preset state of charge change, and obtaining a plurality of DC impedance sequence estimation values ​​corresponding to each segment;

[0042] According to the segment, the state of charge value and the temperature value, by querying SOC-T-DCR Bd Offline data table, to obtain several DC impedance reference values ​​corresponding to each segment;

[0043] According to the DC impedance sequence estimation value, the DC impedance average value in each segment is calculated; wherein the calculation expression of the DC impedance average value is:

[0044] Time k =Time k-1 +Ts

[0045] DCR sum,k =DCR sum,k-1 +DCR k

[0046] DCR Avg,k =DCR sum,k / Time k

[0047] Among them, DCR sum,k The sum of several DC impedance sequence estimates within each segment; DCR kDCR is the estimated value of the DC impedance sequence at time k; Avg,k is the average DC impedance in each segment; Time k is the number of sampling times; Ts is the sampling step length;

[0048] According to the DC impedance reference value, a DC impedance reference average value in each segment is calculated; wherein the calculation expression of the DC impedance average value is:

[0049] Time k =Time k-1 +Ts

[0050] DCR sumBol,k =DCR sumBol,k-1 +DCR k

[0051] DCR AvgBol,k =DCR sumBol,k / Time k

[0052] Among them, DCR sumBol,k It is the sum of several DC impedance reference values ​​of the lithium battery in each segment when it leaves the factory; DCR Bol,k is the DC impedance reference value at time k; DCR AvgBol,k It is the reference average value of DC impedance of each segment.

[0053] The present invention performs segmented processing based on the charge state change amount, and performs average processing on the data in each segment, so as to ensure that a relatively stable DC impedance average value and a DC impedance reference average value are obtained throughout the whole process.

[0054] As a preferred example, the determining the health status of the lithium battery according to a plurality of DC impedance average values ​​and a plurality of DC impedance reference average values ​​includes:

[0055] According to the average value of the DC impedance in each segment, the sum of the average values ​​of the DC impedance in all the segments is calculated, so as to determine the average value of the total DC impedance of the lithium battery during the discharge period by calculating the ratio of the sum of the average values ​​of the DC impedance to the number of segments;

[0056] Calculating the sum of the DC impedance reference average values ​​in all the segments according to the DC impedance reference average value in each segment, so as to determine the total DC impedance reference average value of the lithium battery during the discharge period by calculating the ratio of the sum of the DC impedance reference average values ​​to the number of segments;

[0057] The ratio of the total DC impedance average value to the total DC impedance reference average value is obtained to determine the health state of the lithium battery according to the ratio.

[0058] The present invention evaluates the health status according to the total DC impedance average value and the total DC impedance reference average value that are stable throughout the entire process, thereby ensuring the accuracy of the health status evaluation of the lithium battery.

[0059] On the other hand, the present invention discloses a lithium battery health status assessment system, including an impedance estimation module, a data screening module, a sequence construction module, a sequence update module, a sequence estimation module, a data segmentation module and a health assessment module;

[0060] The impedance estimation module is used to collect terminal voltage data and current data of the lithium battery at each moment in real time during a discharge period to obtain a DC impedance estimation value of the lithium battery at each moment;

[0061] The data screening module is used to obtain the discharge condition of the lithium battery at each moment, and screen a plurality of DC impedance effective values ​​from the plurality of DC impedance estimated values ​​according to the discharge condition; wherein the DC impedance effective value includes a first DC impedance effective value corresponding to each of the first plurality of moments and a second DC impedance effective value corresponding to each of the last plurality of moments;

[0062] The sequence building module is used to collect each of the first DC impedance effective values ​​in real time to build an initial sequence of DC impedance values;

[0063] The sequence updating module is used to iteratively update the DC impedance estimation value sequence according to each of the second DC impedance effective values ​​in real time, and construct a real-time sequence of DC impedance values ​​corresponding to each of the second DC impedance effective values;

[0064] The sequence estimation module is used to calculate the DC impedance sequence estimation value of each of the DC impedance value real-time sequences;

[0065] The data segmentation module is used to segment a plurality of the DC impedance sequence estimation values ​​and the preset discharge DC impedance data table according to the discharge working condition, and calculate the DC impedance average value and the DC impedance reference average value in each segment; wherein the discharge DC impedance data table is used to represent the relationship between the state of charge value and the DC impedance value at different temperatures;

[0066] The health assessment module is used to determine the health status of the lithium battery according to a plurality of DC impedance average values ​​and a plurality of DC impedance reference average values.

[0067] The present invention discloses a lithium battery health status assessment system, which first estimates the DC impedance by collecting the terminal voltage data and current data corresponding to each moment during the discharge period of the lithium battery, so as to realize online detection of the lithium battery when the lithium battery is in a working state. Secondly, the discharge condition corresponding to the lithium battery at each moment is obtained, and the estimated DC impedance estimation value is screened based on the discharge condition, so as to eliminate invalid data according to the screening, thereby improving the efficiency of health status assessment. Furthermore, each DC impedance effective value obtained by screening is collected in real time, and an initial sequence of DC impedance values ​​is first constructed, and then each DC impedance effective value obtained after the initial sequence is constructed is used to iteratively update the initial DC impedance value, so as to obtain a real-time sequence of DC impedance values ​​corresponding to each moment, so as to facilitate the estimation of the DC impedance sequence estimation value corresponding to each moment according to the DC impedance real-time sequence, and then ensure the accuracy of the DC impedance sequence estimation value through stable sequence data. Finally, the obtained continuous multiple DC impedance sequence estimation values ​​are segmented using the discharge conditions, and the preset discharge DC impedance data table is segmented based on the discharge conditions, and then the corresponding DC impedance average value and DC impedance reference average value in each same segment are calculated, and then the health status of the lithium battery is determined based on the DC impedance average value and the DC impedance reference average value, thereby realizing online evaluation of the health status of the lithium battery.

[0068] As a preferred example, the impedance estimation module includes a circuit unit, a data acquisition unit, a parameter identification unit and an estimation unit;

[0069] The circuit unit is used to construct a second-order RC equivalent circuit model, and obtain the current terminal voltage value and current value of the lithium battery at the current moment and the historical terminal voltage value and historical current value at the previous moment;

[0070] The data acquisition unit is used to obtain the state of charge value of the lithium battery at the current moment, and query the preset SOC-U according to the state of charge value. OCV Relationship table, obtaining the open circuit voltage value of the lithium battery at the current moment;

[0071] The parameter identification unit is used to calculate the difference between the current terminal voltage value and the open circuit voltage value, so as to identify the parameters of the second-order RC equivalent circuit model according to the current terminal voltage value, the current current value, the historical terminal voltage value, the historical current value and the difference, and obtain the ohmic internal resistance, high-frequency polarization internal resistance and low-frequency polarization internal resistance of the lithium battery at the current moment by using the recursive least squares method based on the adaptive forgetting factor;

[0072] The estimation unit is used to estimate the DC impedance estimation value corresponding to the lithium battery at the current moment according to the ohmic internal resistance, the high-frequency polarization internal resistance and the low-frequency polarization internal resistance through the discrete time observation equation corresponding to the second-order RC equivalent circuit model; wherein the calculation expression of the DC impedance estimation value is:

[0073] R DCR =R0+R f (1-e (-Ts / τf) )+R s (1-e (-Ts / τs) )

[0074] Among them, the R DCR represents the DC impedance estimation value corresponding to the lithium battery at each moment; R0 represents the ohmic internal resistance of the lithium battery; R f represents the concentration polarization resistance of the lithium battery; the R s represents the battery electrochemical polarization internal resistance of the lithium battery; the R f (1-e (-Ts / τf) ) represents the high frequency polarization impedance; the R s (1-e (-Ts / τs) ) represents the low-frequency polarization impedance; the τ f and τ s represents a time constant; and Ts represents a duration.

[0075] The present invention constructs the second-order RC equivalent circuit model, and based on the real-time voltage data and current data of the battery, uses the recursive least squares method based on the adaptive forgetting factor to perform parameter identification on the circuit model, which can accurately and independently obtain the ohmic internal resistance, high-frequency polarization internal resistance and low-frequency polarization internal resistance of the lithium battery, and then improve the accuracy of subsequent DC impedance value estimation based on the accurate ohmic internal resistance, high-frequency polarization internal resistance and low-frequency polarization internal resistance. Furthermore, the DC impedance value at each moment is calculated based on the parameters obtained by identifying the model and the discrete equation of the second-order equivalent circuit model, so as to realize the online estimation of the DC impedance value. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 : A schematic diagram of a flow chart of a method for evaluating the health status of a lithium battery disclosed in an embodiment of the present invention;

[0077] Figure 2 : A structural schematic diagram of a lithium battery health status evaluation system disclosed in an embodiment of the present invention;

[0078] Figure 3 : A flow chart of a method for evaluating the health status of a lithium battery disclosed in another embodiment of the present invention;

[0079] Figure 4: is a schematic diagram of a second-order RC equivalent circuit model disclosed in another embodiment of the present invention;

[0080] Figure 5 : A schematic diagram of parameter identification based on an adaptive forgetting factor recursive least squares method disclosed in another embodiment of the present invention;

[0081] Figure 6 : A schematic diagram of a DCR abnormal data elimination process based on 2Sigma filtering disclosed in another embodiment of the present invention. DETAILED DESCRIPTION

[0082] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0083] Embodiment 1

[0084] This embodiment discloses a method for evaluating the health status of a lithium battery. For the specific implementation process of the method, please refer to Figure 1 , mainly including steps 101 to 107, wherein the steps are mainly:

[0085] Step 101: During a discharge period, terminal voltage data and current data of the lithium battery at each moment are collected in real time to obtain an estimated value of the DC impedance of the lithium battery at each moment.

[0086] In this embodiment, the steps are mainly as follows: constructing a second-order RC equivalent circuit model, and obtaining the current terminal voltage value and current current value of the lithium battery at the current moment and the historical terminal voltage value and historical current value at the previous moment; obtaining the state of charge value of the lithium battery at the current moment, and querying the preset SOC-U according to the state of charge value. OCV A relationship table is provided to obtain the open circuit voltage value of the lithium battery at the current moment; the difference between the current terminal voltage value and the open circuit voltage value is calculated, so as to identify the parameters of the second-order RC equivalent circuit model according to the current terminal voltage value, the current current value, the historical terminal voltage value, the historical current value and the difference, and obtain the ohmic internal resistance, high-frequency polarization internal resistance and low-frequency polarization internal resistance of the lithium battery at the current moment.

[0087] Further, according to the ohmic internal resistance, high-frequency polarization internal resistance and low-frequency polarization internal resistance, the DC impedance estimation value corresponding to the lithium battery at the current moment is estimated by the discrete time observation equation corresponding to the second-order RC equivalent circuit model; wherein the calculation expression of the DC impedance estimation value is:

[0088] R DCR =R0+R f (1-e (-Ts / τf) )+R s (1-e (-Ts / τs) )

[0089] Among them, the R DCR represents the DC impedance estimation value corresponding to the lithium battery at each moment; R0 represents the ohmic internal resistance of the lithium battery; R f represents the concentration polarization resistance of the lithium battery; the R s represents the battery electrochemical polarization internal resistance of the lithium battery; the R f (1-e (-Ts / τf) ) represents the high frequency polarization impedance; the R s (1-e (-Ts / τs) ) represents the low-frequency polarization impedance; the τ f and τ s represents a time constant; and Ts represents a duration.

[0090] In this embodiment, this step constructs the second-order RC equivalent circuit model, and based on the real-time voltage data and current data of the battery, uses the recursive least squares method based on the adaptive forgetting factor to perform parameter identification on the circuit model, so as to accurately and independently obtain the ohmic internal resistance, high-frequency polarization internal resistance and low-frequency polarization internal resistance of the lithium battery, and then improve the accuracy of subsequent DC impedance value estimation based on the accurate ohmic internal resistance, high-frequency polarization internal resistance and low-frequency polarization internal resistance.

[0091] Step 102: Obtain the discharge condition of the lithium battery at each moment, and select a plurality of DC impedance effective values ​​from the plurality of DC impedance estimation values ​​according to the discharge condition; wherein the DC impedance effective value includes a first DC impedance effective value corresponding to each of the first plurality of moments and a second DC impedance effective value corresponding to each of the last plurality of moments.

[0092] In this embodiment, this step mainly includes: obtaining the state of charge value, temperature value and charge and discharge state corresponding to the lithium battery at each moment; when the state of charge value is in a preset state of charge range, the temperature value is in a preset temperature range and the charge and discharge state is a discharge state, the DC impedance estimation value corresponding to the current moment is determined as the DC impedance effective value.

[0093] In this embodiment, this step sets a valid data range by setting a state of charge interval, a temperature interval and limiting the lithium battery to be in a discharge state, and then combines the valid data range and the real-time state of charge value, real-time temperature value and real-time charge and discharge state corresponding to the lithium battery at each moment to screen each of the DC impedance estimation values ​​to obtain a stable DC impedance value, thereby improving the accuracy of the lithium battery health status assessment based on stable data.

[0094] Step 103: collecting each of the first DC impedance effective values ​​in real time to construct an initial sequence of DC impedance values.

[0095] In this embodiment, this step mainly includes: saving each of the first DC impedance effective values ​​collected in real time in sequence into a cache array set with a preset dimension to construct an initial sequence of DC impedance values ​​corresponding to the preset dimension.

[0096] In this embodiment, this step uses a cache array of preset dimensions to store each of the DC impedance effective values ​​estimated in real time, and completes the initial data storage to construct the initial sequence of DC impedance values ​​when the number of stored data reaches the dimension of the cache array. This allows the initial sequence of DC impedance values ​​to be iteratively updated based on each of the DC impedance effective values ​​collected in real time, thereby ensuring that the DC impedance effective values ​​collected at each moment have a corresponding real-time sequence of DC impedance values, so that a stable DC impedance value can be obtained based on the sequence later.

[0097] Step 104: iteratively update the DC impedance estimation value sequence according to each of the second DC impedance effective values ​​in real time, and construct a real-time sequence of DC impedance values ​​corresponding to each of the second DC impedance effective values.

[0098] Step 105: Calculate a DC impedance sequence estimation value of each of the DC impedance value real-time sequences.

[0099] In this embodiment, this step mainly includes: calculating the average value and standard deviation of each of the real-time sequences of DC impedance values, and constructing an abnormal value screening interval for each of the real-time sequences of DC impedance values ​​according to the average value and the standard deviation; comparing each of the DC impedance effective values ​​in the real-time sequence of DC impedance values ​​with the corresponding abnormal value screening interval, so as to detect abnormal values ​​in the real-time sequence of DC impedance values ​​according to the comparison result; when the DC impedance effective value is not within the abnormal value screening interval, determining that the current DC impedance effective value is a DC impedance abnormal value, and replacing the DC impedance abnormal value with the average value; after replacing the DC impedance abnormal value with the average value, sorting each of the real-time sequences of DC impedance values; obtaining the intermediate value in the sorted real-time sequence of DC impedance values, and determining the intermediate value as the estimated value of the DC impedance sequence.

[0100] In this embodiment, this step screens outliers in the sequence data by calculating the mean value and the standard deviation, so as to prevent the outliers from affecting the DCR estimation accuracy. Then, after removing the outliers in the sequence, the sequence is sorted, so that the middle value of the sequence is used as the stable DC impedance estimation value at the current moment, so as to improve the accuracy of the DC impedance estimation value.

[0101] Step 106: segmenting the plurality of DC impedance sequence estimation values ​​and the preset discharge DC impedance data table according to the discharge condition, and calculating the DC impedance average value and the DC impedance reference average value in each segment; wherein the discharge DC impedance data table is used to represent the relationship between the state of charge value and the DC impedance value at different temperatures.

[0102] In this embodiment, this step mainly includes: according to the discharge condition, obtaining the state of charge value and temperature value corresponding to each of the DC impedance sequence estimation values; according to the state of charge value, segmenting a plurality of the DC impedance sequence estimation values ​​at intervals of a preset state of charge change amount, and obtaining a plurality of DC impedance sequence estimation values ​​corresponding to each segment; according to the segmentation, the state of charge value and the temperature value, by querying SOC-T-DCR Bd Offline data table, to obtain several DC impedance reference values ​​corresponding to each segment;

[0103] Further, according to the DC impedance sequence estimation value, the DC impedance average value in each segment is calculated; wherein the calculation expression of the DC impedance average value is:

[0104] Time k =Time k-1 +Ts

[0105] DCRsum,k =DCR sum,k-1 +DCR k

[0106] DCR Avg,k =DCR sum,k / Time k

[0107] Among them, DCR sum,k The sum of several DC impedance sequence estimates within each segment; DCR k DCR is the estimated value of the DC impedance sequence at time k; Avg,k is the average DC impedance in each segment; Time k is the number of sampling times; Ts is the sampling step length;

[0108] According to the DC impedance reference value, a DC impedance reference average value in each segment is calculated; wherein the calculation expression of the DC impedance average value is:

[0109] Time k =Time k-1 +Ts

[0110] DCR sumBol,k =DCR sumBol,k-1 +DCR k

[0111] DCR AvgBol,k =DCR sumBol,k / Time k

[0112] Among them, DCR sumBol,k It is the sum of several DC impedance reference values ​​of the lithium battery in each segment when it leaves the factory; DCR Bol,k is the DC impedance reference value at time k; DCR AvgBol,k It is the reference average value of DC impedance of each segment.

[0113] In this embodiment, this step performs segmented processing based on the charge state change, and averages the data in each segment to ensure that a relatively stable DC impedance average value and DC impedance reference average value are obtained throughout the entire process.

[0114] Step 107: Determine the health status of the lithium battery according to a plurality of DC impedance average values ​​and a plurality of DC impedance reference average values.

[0115] In this embodiment, this step mainly includes: calculating the sum of the DC impedance average values ​​in all the segments according to the DC impedance average value in each segment, so as to determine the total DC impedance average value of the lithium battery during the discharge period by calculating the ratio of the sum of the DC impedance average values ​​to the number of segments; calculating the sum of the DC impedance reference average values ​​in all the segments according to the DC impedance reference average value in each segment, so as to determine the total DC impedance reference average value of the lithium battery during the discharge period by calculating the ratio of the sum of the DC impedance reference average values ​​to the number of segments; obtaining the ratio of the total DC impedance average value to the total DC impedance reference average value, so as to determine the health status of the lithium battery according to the ratio.

[0116] In this embodiment, this step evaluates the health status according to the total DC impedance average value and the total DC impedance reference average value that are stable throughout the entire process, thereby ensuring the accuracy of the health status evaluation of the lithium battery.

[0117] On the other hand, this embodiment also discloses a lithium battery health status evaluation system. For the specific structure of the evaluation system, please refer to Figure 2 , including an impedance estimation module 201, a data screening module 202, a sequence construction module 203, a sequence update module 204, a sequence estimation module 205, a data segmentation module 206 and a health assessment module 207.

[0118] The impedance estimation module 201 is used to collect the terminal voltage data and current data of the lithium battery at each moment in real time during a discharge period to obtain the DC impedance estimation value of the lithium battery at each moment. The data screening module 202 is used to obtain the discharge condition of the lithium battery at each moment, and screen a plurality of DC impedance effective values ​​from a plurality of DC impedance estimation values ​​according to the discharge condition; wherein the DC impedance effective value includes a first DC impedance effective value corresponding to each moment in the first plurality of moments and a second DC impedance effective value corresponding to each moment in the last plurality of moments.

[0119] The sequence building module 203 is used to collect each of the first DC impedance effective values ​​in real time to build an initial sequence of DC impedance values.

[0120] The sequence updating module 204 is used to iteratively update the initial sequence of DC impedance values ​​according to each of the second DC impedance effective values ​​in real time, and construct a real-time sequence of DC impedance values ​​corresponding to each of the second DC impedance effective values.

[0121] The sequence estimation module 205 is used to calculate the DC impedance sequence estimation value of each of the DC impedance value real-time sequences.

[0122] The data segmentation module 206 is used to segment the plurality of DC impedance sequence estimation values ​​and the preset discharge DC impedance data table according to the discharge working condition, and calculate the DC impedance average value and the DC impedance reference average value in each segment; wherein the discharge DC impedance data table is used to represent the relationship between the state of charge value and the DC impedance value at different temperatures.

[0123] The health assessment module 207 is used to determine the health status of the lithium battery according to a plurality of DC impedance average values ​​and a plurality of DC impedance reference average values.

[0124] In this embodiment, the impedance estimation module 201 includes a circuit unit, a data acquisition unit, a parameter identification unit and an estimation unit;

[0125] The circuit unit is used to construct a second-order RC equivalent circuit model, and obtain the current terminal voltage value and current value of the lithium battery at the current moment and the historical terminal voltage value and historical current value at the previous moment;

[0126] The data acquisition unit is used to obtain the state of charge value of the lithium battery at the current moment, and query the preset SOC-U according to the state of charge value. OCV Relationship table, obtaining the open circuit voltage value of the lithium battery at the current moment;

[0127] The parameter identification unit is used to calculate the difference between the current terminal voltage value and the open circuit voltage value, so as to identify the parameters of the second-order RC equivalent circuit model according to the current terminal voltage value, the current current value, the historical terminal voltage value, the historical current value and the difference, and obtain the ohmic internal resistance, high-frequency polarization internal resistance and low-frequency polarization internal resistance of the lithium battery at the current moment by using the recursive least squares method based on the adaptive forgetting factor;

[0128] The estimation unit is used to estimate the DC impedance estimation value corresponding to the lithium battery at the current moment according to the ohmic internal resistance, the high-frequency polarization internal resistance and the low-frequency polarization internal resistance through the discrete time observation equation corresponding to the second-order RC equivalent circuit model; wherein the calculation expression of the DC impedance estimation value is:

[0129] R DCR =R0+R f (1-e (-Ts / τf) )+R s (1-e (-Ts / τs) )

[0130] Among them, the R DCR represents the DC impedance estimation value corresponding to the lithium battery at each moment; R0 represents the ohmic internal resistance of the lithium battery; R f represents the concentration polarization resistance of the lithium battery; the Rs represents the battery electrochemical polarization internal resistance of the lithium battery; the R f (1-e (-Ts / τf) ) represents the high frequency polarization impedance; the R s (1-e (-Ts / τs) ) represents the low-frequency polarization impedance; the τ f and τ s represents a time constant; and Ts represents a duration.

[0131] The present embodiment discloses a method and system for evaluating the health status of a lithium battery. First, the DC impedance is estimated by collecting the terminal voltage data and current data corresponding to each moment during the discharge period of the lithium battery, so as to realize online detection of the lithium battery when the lithium battery is in a working state. Secondly, the discharge condition corresponding to the lithium battery at each moment is obtained, and the estimated DC impedance estimation value is screened based on the discharge condition, so as to eliminate invalid data according to the screening, thereby improving the efficiency of health status assessment. Furthermore, each effective value of the DC impedance obtained by the screening is collected in real time, and an initial sequence of DC impedance values ​​corresponding to a preset time period is first constructed, and then each of the remaining effective values ​​of the DC impedance is used to iteratively update the initial sequence of DC impedance values ​​to obtain a real-time sequence of DC impedance values ​​corresponding to each moment, so as to facilitate the estimation of the DC impedance sequence estimation value corresponding to each moment according to the real-time sequence of DC impedance values, and then the accuracy of the DC impedance sequence estimation value is ensured through stable sequence data. Finally, the obtained continuous multiple DC impedance sequence estimation values ​​are segmented using the discharge conditions, and the preset discharge DC impedance data table is segmented based on the discharge conditions, and then the corresponding DC impedance average value and DC impedance reference average value in each same segment are calculated, and then the health status of the lithium battery is determined based on the DC impedance average value and the DC impedance reference average value, thereby realizing online evaluation of the health status of the lithium battery.

[0132] Embodiment 2

[0133] Please refer to Figure 3 , is a lithium battery health status assessment method provided in this embodiment, the implementation process of the assessment method includes steps 301 to 306, and the steps are mainly:

[0134] Step 301: construct a second-order equivalent circuit model, collect voltage data and current data of the lithium battery at each moment during the discharge period, and perform parameter identification on the second-order equivalent circuit model based on the voltage data and current data by a recursive least squares method based on an adaptive forgetting factor to obtain an estimated DC impedance value of the lithium battery at each moment.

[0135] Specifically, a second-order RC equivalent circuit model is constructed, and current voltage data and current current data corresponding to the lithium battery at the current moment are obtained, and historical voltage data and historical current data of the lithium battery at the previous moment are obtained; based on the current voltage data and current current data and the historical voltage data and historical current data, parameter identification is performed on the second-order RC equivalent circuit model through a recursive least squares method based on an adaptive forgetting factor to obtain the ohmic internal resistance, high-frequency polarization internal resistance and low-frequency polarization internal resistance of the lithium battery at the current moment; and then the DC impedance estimate of the lithium battery at the current moment is calculated based on the ohmic internal resistance, high-frequency polarization internal resistance and low-frequency polarization internal resistance.

[0136] Optionally, a second-order RC equivalent circuit model is first constructed, which covers the characteristics of battery polarization characteristics (fast / slow) and ohmic internal resistance. The internal composition of the second-order RC equivalent circuit model can be referred to Figure 4 , mainly including the open circuit voltage source, the ohmic internal resistance R0 and the battery terminal voltage U t ; R f and C f They are the concentration polarization resistance and polarization capacitance of the battery, which characterize the rapid electrode reaction inside the battery. The voltage across them is U f ; R s and C s They are the electrochemical polarization internal resistance and polarization capacitance of the battery, representing the slow electrode reaction inside the battery. The voltage across the two ends is U s .

[0137] Will be like Figure 4 The second-order RC equivalent circuit model shown is mathematically expressed, and according to Kirchhoff's law, the system equation and observation equation of the second-order RC equivalent circuit model are obtained. The system equation and observation equation are as follows:

[0138]

[0139] U t =U oc -i L R0-U f -U s

[0140] In the above equation, the discharge current direction is defined as positive, and the charge current direction is defined as negative. Further, the system equation and observation equation are converted into discrete time form, and [U f,k U s,k ] is the state variable, and the following state space equation and observation equation are established:

[0141]

[0142] Ut,k =U OCV,k -U f,k -U s,k -R0i L,k

[0143] Referring to the above formula, these are the state space equation and observation equation corresponding to the second-order RC equivalent circuit model.

[0144] Secondly, a recursive least squares method based on an adaptive forgetting factor (AFFRLS algorithm) is constructed; wherein the construction process of the adaptive forgetting factor recursive least squares method is as follows:

[0145] 1) The least square method is used; wherein the mathematical expression of the least square method is:

[0146]

[0147] In the formula, e Ls,k is a stationary zero-mean white noise.

[0148] 2) The core of the least square method (RLS) is to minimize the sum of square errors between the predicted data and the measured data, and thus the least square method is judged by constructing a cost function J; wherein the expression of the cost function J is:

[0149]

[0150] Among them, the parameter to be identified is θ k As the independent variable, in order to obtain the parameter vector θ k The optimal estimate of the cost function J with respect to the independent variable θ is obtained according to the principle of least squares method. k The minimum value of the construction cost function, that is, when the value of the construction cost function is the minimum, the corresponding θ k The value of is the optimal parameter solution; then the cost function J is minimum and the independent variable θ k The relational expression is:

[0151]

[0152] When the cost function J is minimum, the independent variable θ k When the relational expression of is solved, the expression to be solved is:

[0153]

[0154] 3) The expression to be solved is converted into a recursive form, and a forgetting factor is added in the conversion process to construct a recursive least squares method based on forgetting factor (FFRLS); the expression of the recursive least squares method based on forgetting factor is:

[0155]

[0156] Wherein, λ is the forgetting factor. When the value is 1, the recursive least squares method based on the forgetting factor degenerates into the traditional recursive least squares algorithm. Ls,k is the gain of the algorithm, P Ls,k is the error covariance matrix of the state estimate, and I is the identity matrix. is the parameter value estimated at the previous moment, is the observed value at this moment, y k is the actual observed value of the system, y k and Subtract the prediction error of the system and add the prediction error to the gain matrix K Ls,k The multiplication results in a correction to the estimated parameter value at this time, which is equal to the estimated value K at the previous moment. Ls,k Finally, we get the estimated value

[0157] Fixed forgetting factor λ k When the value is close to 1, the parameter identification result is more stable, the convergence speed is slow, and the identification result is less affected by new data; k When the value is far away from 1, the greater the fluctuation of the parameter identification result, the faster the convergence speed, and the identification result is greatly affected by the new data. The large fluctuation of the identification parameter reflects the time-varying characteristics of the model parameters changing with the working conditions.

[0158] In order to further improve the parameter identification accuracy, the voltage error is introduced to dynamically adjust the forgetting factor λ k The voltage error is the difference between the actual voltage and the identified voltage, which reflects the consistency between the parameter identification equivalent circuit parameters and the actual parameters. When the voltage error is large, the forgetting factor needs to be reduced to improve the tracking performance of the algorithm and achieve the purpose of reducing the voltage error. When the voltage error is small, the forgetting factor needs to be increased to ensure the stability of the algorithm and achieve the purpose of stabilizing the parameters.

[0159] 4) Considering that the frequently changing forgetting factor affects the convergence of the identification parameters, and the strategy of adjusting the forgetting factor of a single data is sensitive to interference, the data sliding window theory is adopted to adjust the size of the forgetting factor by using the mean square value of the multi-step voltage error within the window time instead of the single-step voltage error. The adjustment expression of the forgetting factor is:

[0160]

[0161] In the formula, λ k represents the forgetting factor at time k, λ max and λ min Respectively represent the maximum and minimum values ​​of the forgetting factor, ρ represents the sensitivity coefficient, αk represents the mean square value of the multi-step voltage error in the time window, M represents the time window size, e k Represents the voltage error at time k. The maximum forgetting factor λ of this scheme max =1, minimum forgetting factor λ min =0.98, time window M=30, that is, the average voltage error within 30 consecutive seconds, and the sensitivity coefficient is 500000≤ρ≤1000000.

[0162] Furthermore, the AFFRLS algorithm is applied to Figure 4 The parameter identification of the second-order RC equivalent circuit model of the constructed lithium battery, the process of the AFFRLS algorithm for parameter identification of the second-order RC equivalent circuit model is specifically as follows:

[0163] 1) Based on the system equation and observation equation of the second-order RC equivalent circuit model, the transfer function between the output voltage and input current of the lithium battery is obtained by Kirchhoff's law and Laplace transform; the transfer function is:

[0164]

[0165] 2) Discretize the transfer function of the second-order equivalent circuit model in the frequency domain and transform the battery model into the least squares mathematical form. The expression of the least squares mathematical form is:

[0166] E L (s)=U t (s)-U OCV (s),

[0167]

[0168] 3) Parameter merging is performed on the expression of the least square mathematical form, then:

[0169]

[0170] 4) Define the equivalent intermediate parameter D i (i=1,2,…,5), and perform parameter substitution on the expression of the least square mathematical form after the parameters are combined. Then, the mathematical expression of the equivalent intermediate parameter is defined as:

[0171]

[0172] 5) Simplify the formula in step 3 according to the equivalent intermediate parameters in step 4, and the simplified formula expression is:

[0173]

[0174] 6) Based on the bilinear change law, the frequency domain transfer function can be mapped to the Z plane to perform time domain discretization operations, and the formula: Substituting into the formula in step 5 we get:

[0175]

[0176] Therefore, the discrete expression of the transfer function of the second-order equivalent circuit model of the battery in the z plane is as follows:

[0177]

[0178] Among them, θ i (i=1,2,…,5) are the parameters to be identified.

[0179] 7) The parameter to be identified θ can be obtained from the corresponding relationship of the formula in step 6 i Analytical formula, that is, the parameter to be identified θ i The analytical formula is:

[0180]

[0181] 8) According to the discrete expression of the transfer function of the second-order equivalent circuit model of the battery constructed in step 6 in the z plane, the mathematical analytical expression of the least squares difference equation can be obtained:

[0182] E L (k) = θ1E L (k-1)+θ2E L (k-2)+θ3I L (k)+θ4I L (k-1)+θ5I L (k-2)

[0183] According to the mathematical analytical expression, the parameter matrix (system output) and data matrix (system input) of the second-order RC equivalent circuit model are determined as follows:

[0184]

[0185] 9) After parameter identification, we can get θ at different stages i , through the parameter to be identified in step 7 θ i The analytical expression can be reversed to obtain the equivalent intermediate parameter D i ,Right now

[0186]

[0187] 10) According to the mathematical expression of the equivalent intermediate parameter defined in step 4, the equivalent intermediate parameter D i , and analyze the actual equivalent circuit parameters:

[0188]

[0189] The process of parameter identification of the second-order RC equivalent circuit model based on the above-mentioned AFFRLS algorithm, in one embodiment, refers to Figure 5 , collect the current voltage data and current current data corresponding to the lithium battery at the current moment; wherein, the current voltage data and current current data are set to be the current i at moment k L (k), voltage U at time k t (k). Simultaneously obtain the historical voltage data and historical current data of the lithium battery corresponding to the last moment before the current moment; wherein the historical voltage data and historical current data are set to the current i at moment k-1. L Current i at time (k-1) and k-2 L Voltage U at time (k-2) and time k-1 t Voltage U at time (k-1) and time k-2 t (k-2), and obtain the state of charge SOC(k) of the lithium battery at the current moment, and obtain the U corresponding to the SOC(k) by looking up the table OCV (SOC k );

[0190] The voltage and current data collected above and the state of charge and U OCV (SOC k ) is used as the input parameter of the second-order RC equivalent circuit model, and the adaptive forgetting factor recursive least squares parameter identification is performed, and the output is θ i (i=1,2,…,5).

[0191] Specifically, in the process of parameter identification, the identified algorithm is first initialized, the simulation step is set to Ts = 1s, the initial SOC value is read, the parameter matrix to be identified is initialized to θ = [0 0 0 0 0], and the error covariance matrix P is initialized. Ls,k =10 6 *eye(6); Based on the real-time SOC estimation value, through SOC-U OCV Look up the table to obtain the current moment, combined with the terminal voltage U t Calculate the terminal voltage error at this moment: E L =U t -U OCV , to calculate the adaptive forgetting factor according to the terminal voltage error; according to the current current, voltage, E L The value and the adaptive forgetting factor are used to obtain the identification parameter value θ in real time based on the constructed AFFRLS algorithm. i ; According to the obtained identification parameter value θ iand the formulas constructed in steps 9 and 10 of the parameter identification process to resolve the equivalent circuit parameters R0, R f and R s ; according to the equivalent circuit parameters, calculate the DC impedance of the lithium battery at the current moment K from the discrete-time observation equation of the second-order RC equivalent circuit; then the calculation expression of the DC impedance is:

[0192] R DCR = R0 + R f (1 - e (-Ts / τf) ) + R s (1 - e (-Ts / τs) )

[0193] In the formula, R0 is the ohmic internal resistance, R f (1 - e (-Ts / τf) ) represents the high-frequency polarization impedance, R s (1 - e (-Ts / τs) ) is the low-frequency polarization impedance, R f and R s are the initial values of the polarization impedance, τ f and τ s are the time constants, Ts is the duration, and in this embodiment, Ts = 1s.

[0194] Step 302: Obtain the discharge condition of the lithium battery at each moment to judge the validity of each DC impedance estimated value estimated in real time according to the discharge condition, and screen the DC impedance effective values according to the judgment result.

[0195] Specifically, in this embodiment, obtain the state of charge value, temperature value, and charge and discharge state corresponding to the lithium battery at each moment; when the state of charge value is within a preset state of charge interval, the temperature value is within a preset temperature interval, and the charge and discharge state is the discharge state, determine the DC impedance estimated value corresponding to the current moment as the DC impedance effective value.

[0196] Optionally, test the change law of the internal resistance of the lithium battery at different SOCs and temperatures through the HPPC method, and find that the DC impedance of the lithium battery is relatively stable when the state of charge is in the working interval of 30% < SOC < 80%. At the same time, in order to enable the lithium battery to have better buffering ability, in one embodiment, the state of charge interval [30, 60] is selected as the state of charge interval for screening stable DCR values. In the actual process, the state of charge interval can be adjusted according to the DCR characteristics of the battery cell.

[0197] Furthermore, as the temperature changes, the ohmic impedance and high / low frequency polarization impedance of the lithium battery have different change amplitudes. In order to ensure the stability of the DCR value, in one embodiment, a temperature interval of (20, 50) is selected as the temperature interval for screening a stable DCR value.

[0198] Based on the good parameter identification effect of AFFRLS under discharge conditions, in one embodiment, parameter identification is performed only on batteries in a discharge state, wherein the discharge state is identified by the battery operating state.

[0199] Based on the set temperature range, the state of charge range and the set restriction condition of only evaluating the battery in the discharge state, the validity of the DC impedance estimation value estimated in real time at each moment is determined based on the real-time SOC, temperature and charge and discharge state at each moment; when it is determined that the real-time SOC, temperature and charge and discharge state are within the corresponding range, the DC impedance estimation value corresponding to the current moment is determined to be the DC impedance effective value.

[0200] Step 303: Save each of the DC impedance effective values ​​in real time until the initial sequence of DC impedance values ​​is constructed, and iteratively update the initial sequence of DC impedance values ​​according to each of the DC impedance effective values ​​obtained after the initial sequence of DC impedance values ​​is constructed, so as to construct a real-time sequence of DC impedance values ​​corresponding to each of the DC impedance effective values ​​obtained after the initial sequence of DC impedance values ​​is constructed.

[0201] Specifically, in this embodiment, each of the DC impedance effective values ​​collected in real time is saved in sequence in a cache array set with a preset dimension to construct an initial sequence of DC impedance values ​​corresponding to the preset dimension, and the initial sequence of DC impedance values ​​is updated according to each of the remaining DC impedance effective values ​​collected in real time to construct a real-time sequence of DC impedance values ​​corresponding to each of the DC impedance effective values.

[0202] Optionally, in this embodiment, refer to Figure 6, a Buffer array with a preset dimension of [1,50] is used to store the real-time estimated effective value of DC impedance to construct an initial sequence of DC impedance values ​​with the same dimension. In one embodiment, if the sampling step is set to 1S, the effective value of DC impedance corresponding to the lithium battery in every 1S is calculated. When it is determined that the effective values ​​of DC impedance corresponding to each 1S are all valid data, based on the dimension of the Buffer array of 50, it is necessary to continuously collect the data for the first 50 seconds to complete the storage, so as to construct an initial sequence of DC impedance values ​​based on the effective values ​​of DC impedance estimated every 1S in the first 50 seconds. Then, the initial sequence of DC impedance values ​​is iteratively updated according to the effective values ​​of DC impedance within the valid data range obtained every 1S after 50S, so as to ensure that there is a set of [1,50] real-time sequence of DC impedance values ​​at every moment.

[0203] Step 304: remove abnormal data from each of the real-time sequences of DC impedance values, and after removing the abnormal data, calculate a DC impedance sequence estimation value corresponding to each of the real-time sequences of DC impedance values.

[0204] Specifically, in this embodiment, the mean value and standard deviation of each real-time series of the DC impedance values ​​are calculated, and abnormal data are screened according to the mean value and the standard deviation. After the abnormal data are screened out, the abnormal data are replaced with the mean value, thereby completing the elimination of the abnormal data.

[0205] Optionally, the outlier degree of each effective DC impedance value in the real-time sequence of the DC impedance value is obtained by using the normal distribution principle. When the outlier degree of an effective DC impedance value is too large, it indicates that the effective DC impedance value has an outlier effect, which will affect the estimation accuracy of DCR. In this regard, the 1Sigma criterion, 2sigma criterion or 3sigma criterion in the normal distribution principle can be used to screen abnormal data.

[0206] The 1Sigma criterion defines that approximately 68.27% of the data values ​​fall within the range of the mean (μ) plus or minus one standard deviation (σ), and the mathematical expression is: , then the DCR value not in this interval is an abnormal value.

[0207] The 2Sigma criterion defines that approximately 95.45% of the data values ​​fall within the range of the mean (μ) plus or minus two standard deviations (σ). The mathematical expression is, then the DCR value not in this interval is an outlier.

[0208] The 3Sigma criterion defines that approximately 99.73% of the data values ​​fall within the range of the mean (μ) plus or minus three standard deviations (σ). The mathematical expression is: , then the DCR value not in this interval is an outlier.

[0209] In one embodiment, 2Sigma filtering is used, and the confidence interval is set within the range of the mean plus or minus two standard deviations. The outlier is defined as the deviation between the real-time DCR and the average value of the DCR sequence exceeding two standard deviations. The schematic diagram of the DCR abnormal data elimination process based on 2Sigma filtering is shown in the figure. Figure 6 As shown:

[0210] Reference Figure 6 , calculate the average value DCR of each real-time series of DC resistance values AVG,K and standard deviation DCR Del,K , the K represents the time corresponding to the real-time sequence of the DC impedance value. Traverse each DC impedance effective value in the real-time sequence of the DC impedance value and compare it with the average value DCR AVG,K and standard deviation DCR Del,K The standard value interval is compared to determine whether it is within the interval. When the DC impedance effective value is not within the interval, the average value DCR is calculated based on the average value DCR. AVG,K The DC impedance effective value is replaced to obtain sequence data of each DC impedance value real-time sequence after eliminating abnormal data.

[0211] The DCR outliers are replaced by the DCR sequence average, that is,

[0212]

[0213] Where, DCR k is the calculated value of DCR at time k, DCR k,μ is the average value of the DCR sequence at time k, σ k is the standard deviation of the DCR sequence at time k.

[0214] After removing the outliers of the DCR sequence, the valid DCR sequences in the [1,50] buffer are sorted at the same time, and the middle value of the DCR sequence is sent as the output as the stable DCR estimate at the current moment for subsequent calculations.

[0215] Step 305: segment each DC impedance sequence estimation value and each pre-saved DC impedance reference value based on the SOC variation to calculate an estimated average value corresponding to a number of the DC impedance sequence estimation values ​​and a reference average value corresponding to a number of the DC impedance reference values ​​in each segment.

[0216] Specifically, in this embodiment, the DC impedance reference value is the DC impedance DCR of the battery in the corresponding working condition when it leaves the factory. Bol. In the segmentation process, the effective DCR sequence of the discharge process is segmented based on the SOC change DeltaSOC, where DeltaSOC is defined as 1% / 2% / 5%, and the appropriate interval range is defined according to the needs. This solution takes DeltaSOC=1% as the interval. When the DCR estimation result is valid, several effective DCR estimation values ​​in each DeltaSOC=1% segment are recorded, the DCR average value in the segment is calculated, and the DCR average value is defined as the actual output value of the DCR in the segment. At the same time, in the same segment, based on SOC-T-DCR Bol Offline data table, obtain DCR in this segment through SOC and temperature Bol Value, calculate the DCR in this segment Bol The average value of the DCR is defined as Bol The average value of is the DCR output reference value in this segment.

[0217] Among them, the DCR average value is calculated as follows:

[0218] Time k =Time k-1 +Ts

[0219] DCR sum,k =DCR sum,k-1 +DCR k

[0220] DCR Avg,k =DCR sum,k / Time k

[0221] Where, DCR sum,k The sum of the DCR calculation values ​​in the segment, DCR k is the effective DCR calculated value at time k, DCR Avg,k is the effective DCR average value in the segment, Time k is the sampling times, Ts is the sampling step, that is, how often the DCR value is sampled.

[0222] Among them, DCR Bol The average value of is calculated as follows:

[0223] Time k =Time k-1 +Ts

[0224] DCR sumBol,k =DCR sumBol,k-1 +DCR k

[0225] DCR AvgBol,k =DCR sumBol,k / Time k

[0226] Where, DCR sumBol,k The sum of the DCR reference values ​​in the segment when leaving the factory. Bol,k is the DCR table value when leaving the factory at time k, DCR AvgBol,k It is the average reference value of DCR in the segment when it leaves the factory.

[0227] Step 306: Determine the health status of the lithium battery according to the multiple estimated average values ​​and the multiple reference average values.

[0228] Specifically, in this embodiment, according to the DC impedance average value in each segment, the DC impedance reference average value and the number of segments, the total DC impedance average value and the total DC impedance reference average value corresponding to the lithium battery during the entire discharge period are determined, and the health status of the lithium battery is determined according to the ratio of the total DC impedance average value to the total DC impedance reference average value.

[0229] Optionally, as the battery discharges, the average value of the DCR in each segment is gradually averaged, and at the same time, the DCR in each segment is gradually averaged. Bol The average value is averaged based on the average DCR and average DCR during the discharge cycle Bol The ratio of the DC impedance to the health status SOHR of the lithium battery is obtained.

[0230] SOHR=DCR Avg,k / DCR AvgBol,k

[0231] In a specific implementation manner, different manufacturers define the degree of battery aging differently. Preferably, it can be set that when the SOHR is stably greater than 2, the battery life is considered to have ended.

[0232] The present embodiment discloses a method for evaluating the health status of a lithium battery. Under dynamic discharge conditions, first, the terminal voltage and total current of the lithium battery are collected to obtain the current operating status value of the battery. On the basis of continuous data collection, the ohmic internal resistance, high-frequency polarization internal resistance and low-frequency polarization internal resistance of the lithium battery when the "terminal voltage square error is minimum" are obtained based on the adaptive forgetting factor recursive least squares identification algorithm, and the DCR estimation value is calculated based on the second-order equivalent circuit discrete equation, thereby obtaining a continuously changing DCR sequence. Secondly, based on the battery SOC, temperature and charge and discharge state, the valid data range is divided. Within the valid data range, the DCR sequence is preprocessed, and abnormal data is eliminated through 2Sigma principle filtering. Then, segmented processing is performed based on the SOC change amount to obtain the relatively stable DCR estimation value and the average value of the DCR reference value of the whole process; finally, the lithium battery SOHR is obtained through the health status estimation model, and the online health status assessment of the lithium battery is realized.

[0233] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for evaluating the health status of a lithium battery, characterized in that: include: During a discharge period, terminal voltage data and current data of the lithium battery at each moment are collected in real time to obtain an estimated DC impedance value of the lithium battery at each moment; Obtaining the discharge condition of the lithium battery at each moment, and selecting a plurality of DC impedance effective values ​​from the plurality of DC impedance estimation values ​​according to the discharge condition; wherein the DC impedance effective value includes a first DC impedance effective value corresponding to each of the first plurality of moments and a second DC impedance effective value corresponding to each of the last plurality of moments; Collecting each of the first DC impedance effective values ​​in real time to construct an initial sequence of DC impedance values; Iteratively updating the initial sequence of DC impedance values ​​in real time according to each of the second DC impedance effective values, and constructing a real-time sequence of DC impedance values ​​corresponding to each of the second DC impedance effective values; Calculating a DC impedance sequence estimation value of each of the DC impedance value real-time sequences; According to the discharge working condition, the plurality of DC impedance sequence estimation values ​​and the preset discharge DC impedance data table are segmented, and the DC impedance average value and the DC impedance reference average value in each segment are calculated; wherein the discharge DC impedance data table is used to indicate the relationship between the state of charge value and the DC impedance value at different temperatures; The health status of the lithium battery is determined according to a plurality of the DC impedance average values ​​and a plurality of the DC impedance reference average values.

2. A method for evaluating the health status of a lithium battery according to claim 1, characterized in that: The real-time collection of the terminal voltage data and current data of the lithium battery at each moment to obtain the DC impedance estimation value of the lithium battery at each moment includes: Constructing a second-order RC equivalent circuit model, and obtaining the current terminal voltage value and current value of the lithium battery at the current moment and the historical terminal voltage value and historical current value at the previous moment; Obtain the state of charge value of the lithium battery at the current moment, and query the preset SOC-U according to the state of charge value OCV Relationship table, obtaining the open circuit voltage value of the lithium battery at the current moment; The difference between the current terminal voltage value and the open-circuit voltage value is calculated, and the parameters of the second-order RC equivalent circuit model are identified by a recursive least squares method based on an adaptive forgetting factor according to the current terminal voltage value, the current current value, the historical terminal voltage value, the historical current value and the difference, so as to obtain the ohmic internal resistance, high-frequency polarization internal resistance and low-frequency polarization internal resistance of the lithium battery at the current moment.

3. A method for evaluating the health status of a lithium battery according to claim 2, characterized in that: The obtaining of the estimated DC impedance value of the lithium battery at each moment includes: According to the ohmic internal resistance, high-frequency polarization internal resistance and low-frequency polarization internal resistance, the DC impedance estimation value corresponding to the lithium battery at the current moment is estimated by the discrete time observation equation corresponding to the second-order RC equivalent circuit model; wherein the calculation expression of the DC impedance estimation value is: R DCR NR0+R f (1-e (-Ts / τf) )+R s (1-e (-Ts / τs) ) Among them, the R DCR represents the DC impedance estimation value corresponding to the lithium battery at each moment; R0 represents the ohmic internal resistance of the lithium battery; R f represents the concentration polarization resistance of the lithium battery; the R s represents the battery electrochemical polarization internal resistance of the lithium battery; the R f (1-e (-Ts / τf) ) represents the high frequency polarization impedance; the R s (1-e (-Ts / τs) ) represents the low-frequency polarization impedance; the τ f and τ s represents a time constant; and Ts represents a duration.

4. A method for evaluating the health status of a lithium battery according to claim 1, characterized in that: The step of selecting a plurality of DC impedance effective values ​​from a plurality of DC impedance estimated values ​​according to the discharge working condition comprises: Obtaining the state of charge value, temperature value and charge and discharge state of the lithium battery corresponding to each moment; When the state of charge value is in a preset state of charge range, the temperature value is in a preset temperature range, and the charge-discharge state is a discharge state, the DC impedance estimation value corresponding to the current moment is determined as the DC impedance effective value.

5. A method for evaluating the health status of a lithium battery according to claim 1, characterized in that: The real-time acquisition of each of the first DC impedance effective values ​​to construct an initial sequence of DC impedance values ​​includes: Each of the first DC impedance effective values ​​collected in real time is sequentially saved in a cache array set with a preset dimension to construct an initial sequence of DC impedance values ​​corresponding to the preset dimension.

6. A method for evaluating the health status of a lithium battery according to claim 1, characterized in that: The step of calculating a DC impedance sequence estimation value of each DC impedance value real-time sequence comprises: Calculating the mean value and standard deviation of each of the real-time series of DC impedance values, and constructing an abnormal value screening interval for each of the real-time series of DC impedance values ​​according to the mean value and the standard deviation; Comparing each of the DC impedance effective values ​​in the DC impedance value real-time sequence with the corresponding abnormal value screening interval, so as to detect abnormal values ​​in the DC impedance value real-time sequence according to the comparison result; When the DC impedance effective value is not within the abnormal value screening interval, determining that the current DC impedance effective value is a DC impedance abnormal value, and replacing the DC impedance abnormal value with the average value; After replacing the abnormal DC impedance value with the average value, sorting each of the real-time sequences of DC impedance values; The intermediate value in the sorted real-time sequence of DC impedance values ​​is obtained, and the intermediate value is determined as the DC impedance sequence estimation value.

7. A method for evaluating the health status of a lithium battery according to claim 1, characterized in that: The step of segmenting the plurality of DC impedance sequence estimation values ​​and the preset discharge DC impedance data table according to the discharge working condition, and calculating the DC impedance average value and the DC impedance reference average value in each segment, comprises: According to the discharge working condition, a state of charge value and a temperature value corresponding to each of the DC impedance sequence estimation values ​​are obtained; According to the state of charge value, segmenting the plurality of DC impedance sequence estimation values ​​at intervals of a preset state of charge change, and obtaining a plurality of DC impedance sequence estimation values ​​corresponding to each segment; According to the segment, the state of charge value and the temperature value, by querying SOC-T-DCR Bd Offline data table, to obtain several DC impedance reference values ​​corresponding to each segment; According to the DC impedance sequence estimation value, the DC impedance average value in each segment is calculated; wherein the calculation expression of the DC impedance average value is: Time k =Time k-1 +Ts DCR sum,k =DCR sum,k-1 +DCR k DCR Avg,k =DCR sum,k / Time k Among them, DCR sum,k The sum of several DC impedance sequence estimates within each segment; DCR k DCR is the estimated value of the DC impedance sequence at time k; Avg,k is the average DC impedance in each segment; Time k is the number of sampling times; Ts is the sampling step length; According to the DC impedance reference value, a DC impedance reference average value in each segment is calculated; wherein the calculation expression of the DC impedance reference average value is: Time k =Time k-1 +Ts DCR sumBol,k =DCR sumBol,k-1 +DCR k DCR AvgBol,k =DCR sumBol,k / Time k Among them, DCR sumBol,k It is the sum of several DC impedance reference values ​​of the lithium battery in each segment when it leaves the factory; DCR Bol,k is the DC impedance reference value at time k; DCR AvgBol,k It is the reference average value of DC impedance of each segment.

8. A method for evaluating the health status of a lithium battery according to claim 7, characterized in that: The determining the health status of the lithium battery according to a plurality of DC impedance average values ​​and a plurality of DC impedance reference average values ​​comprises: According to the average value of the DC impedance in each segment, the sum of the average values ​​of the DC impedance in all the segments is calculated, so as to determine the average value of the total DC impedance of the lithium battery during the discharge period by calculating the ratio of the sum of the average values ​​of the DC impedance to the number of segments; Calculating the sum of the DC impedance reference average values ​​in all the segments according to the DC impedance reference average value in each segment, so as to determine the total DC impedance reference average value of the lithium battery during the discharge period by calculating the ratio of the sum of the DC impedance reference average values ​​to the number of segments; The ratio of the total DC impedance average value to the total DC impedance reference average value is obtained to determine the health state of the lithium battery according to the ratio.

9. A lithium battery health status assessment system, characterized in that: It includes an impedance estimation module, a data screening module, a sequence construction module, a sequence update module, a sequence estimation module, a data segmentation module and a health assessment module; The impedance estimation module is used to collect terminal voltage data and current data of the lithium battery at each moment in real time during a discharge period to obtain a DC impedance estimation value of the lithium battery at each moment; The data screening module is used to obtain the discharge condition of the lithium battery at each moment, and screen a plurality of DC impedance effective values ​​from the plurality of DC impedance estimated values ​​according to the discharge condition; wherein the DC impedance effective value includes a first DC impedance effective value corresponding to each of the first plurality of moments and a second DC impedance effective value corresponding to each of the last plurality of moments; The sequence building module is used to collect each of the first DC impedance effective values ​​in real time to build an initial sequence of DC impedance values; The sequence updating module is used to iteratively update the initial sequence of DC impedance values ​​according to each of the second DC impedance effective values ​​in real time, and construct a real-time sequence of DC impedance values ​​corresponding to each of the second DC impedance effective values; The sequence estimation module is used to calculate the DC impedance sequence estimation value of each of the DC impedance value real-time sequences; The data segmentation module is used to segment a plurality of the DC impedance sequence estimation values ​​and the preset discharge DC impedance data table according to the discharge working condition, and calculate the DC impedance average value and the DC impedance reference average value in each segment; wherein the discharge DC impedance data table is used to represent the relationship between the state of charge value and the DC impedance value at different temperatures; The health assessment module is used to determine the health status of the lithium battery according to a plurality of DC impedance average values ​​and a plurality of DC impedance reference average values.

10. A lithium battery health status assessment system according to claim 9, characterized in that: The impedance estimation module includes a circuit unit, a data acquisition unit, a parameter identification unit and an estimation unit; The circuit unit is used to construct a second-order RC equivalent circuit model, and obtain the current terminal voltage value and current value of the lithium battery at the current moment and the historical terminal voltage value and historical current value at the previous moment; The data acquisition unit is used to obtain the state of charge value of the lithium battery at the current moment, and query the preset SOC-U according to the state of charge value. OCV Relationship table, obtaining the open circuit voltage value of the lithium battery at the current moment; The parameter identification unit is used to calculate the difference between the current terminal voltage value and the open circuit voltage value, so as to identify the parameters of the second-order RC equivalent circuit model according to the current terminal voltage value, the current current value, the historical terminal voltage value, the historical current value and the difference, and obtain the ohmic internal resistance, high-frequency polarization internal resistance and low-frequency polarization internal resistance of the lithium battery at the current moment by using the recursive least squares method based on the adaptive forgetting factor; The estimation unit is used to estimate the DC impedance estimation value corresponding to the lithium battery at the current moment according to the ohmic internal resistance, the high-frequency polarization internal resistance and the low-frequency polarization internal resistance through the discrete time observation equation corresponding to the second-order RC equivalent circuit model; wherein the calculation expression of the DC impedance estimation value is: R DCR NR0+R f (1-e (-Ts / τf) )+R s (1-e (-Ts / τs) ) Among them, the R DCR represents the DC impedance estimation value corresponding to the lithium battery at each moment; R0 represents the ohmic internal resistance of the lithium battery; R f represents the concentration polarization resistance of the lithium battery; the R s represents the battery electrochemical polarization internal resistance of the lithium battery; the R f (1-e (-Ts / τf) ) represents the high frequency polarization impedance; the R s (1-e (-Ts / τs) ) represents the low-frequency polarization impedance; the τ f and τ s represents a time constant; and Ts represents a duration.

Images