A method for online quantitative evaluation of the micro-short circuit degree of lithium iron phosphate batteries

Abstract

The present invention provides a method for online quantitative assessment of the degree of micro-short circuit in lithium iron phosphate batteries. By collecting the current and voltage data of two cells during the vehicle's driving conditions on a certain date, the open-circuit voltage of the two cells is identified using a constructed battery model. After fitting or smoothing, the difference in the open-circuit voltage of the two cells is integrated with a fixed point a as the cutoff point. By controlling parameters such as b and c, an S value that can stably reflect the characteristics of the micro-short circuit is obtained. By calculating the rate of change of S with the date, the self-discharge rate value of the lithium iron phosphate battery can be obtained. Because the S value is not affected by the charge rate, temperature, model error, depth of discharge, etc., the present invention can sensitively reflect the occurrence of micro-short circuits, achieve quantitative and high-precision estimation of the degree of short circuit, and solve the problem of low accuracy of online estimation of the self-discharge rate of lithium iron phosphate batteries in existing technical solutions, reserve more time for repairing problem batteries, and ensure the safety of vehicle property.

Description

Technical Field

[0001] The present invention relates to the field of new energy battery application technology, and more specifically to a method for online quantitative evaluation of the micro-short circuit degree of a lithium iron phosphate battery. Background Art

[0002] The safety of new energy vehicles has always been a major concern for consumers. Accidents such as fires and smoke are often caused by internal battery short circuits. While onboard battery management systems or fire extinguishing systems can provide an alert 5-30 minutes in advance of a fire, allowing passengers time to escape, they cannot prevent property damage resulting from vehicle damage. Identifying micro-short circuits—that is, detecting even the smallest battery short circuits—is an effective way to prevent property damage.

[0003] At present, there are many online battery micro-short circuit detection methods, but they are effective for battery systems with a large slope of change in the battery open-circuit voltage platform. For battery systems with a relatively flat voltage platform, such as lithium iron phosphate, the voltage does not change significantly with the change of the discharge and charge state, resulting in poor algorithm effect and even inability to effectively identify slight internal short circuits.

[0004] Application publication number CN 113848495A discloses a method for diagnosing internal micro-short circuit faults based on charging curves. The method includes: capacity increment method (ICA) analysis, battery capacity change rate monitoring, and voltage release curve tracking. Based on the IC curve obtained from the battery charging curve, aging characteristics are extracted to understand the current aging state of the battery. The characteristics of the battery internal short circuit fault are obtained by comparing the capacity decay rates of two adjacent charging cycles obtained through long-term monitoring. Finally, the voltage curve shortly after charging is compared with the voltage curve in a brand new state to determine whether a micro-short circuit has occurred. This patent primarily utilizes battery charging curves, but in actual use of new energy vehicles, constant current charging is difficult to achieve for many vehicles, reducing its versatility.

[0005] The Chinese invention patent application, CN 11929602A, discloses a quantitative diagnostic method for single-cell battery leakage or micro-short circuits based on capacity estimation. The method includes the following steps: S1: obtaining charge and discharge data of the battery cell; S2: using traditional capacity estimation methods to estimate the battery's charge capacity CC and discharge capacity CD; S3: calculating the ratio of discharge capacity to charge capacity. When the ratio is less than a threshold, a leakage fault is determined; S4: calculating the leakage current estimate based on the ratio of discharge capacity to charge capacity. However, the charging capacity on a real vehicle is significantly affected by temperature and current rate; the discharge capacity on a real vehicle is also significantly affected by ambient temperature and vehicle operating conditions. Therefore, this method cannot be applied on a real vehicle. Summary of the Invention

[0006] The present invention provides a method for online quantitative evaluation of the micro-short circuit degree of lithium iron phosphate batteries, so as to overcome the shortcomings of existing online battery micro-short circuit detection methods, which are affected by temperature, current rate, etc. during the charging and discharging process, resulting in low online estimation accuracy of the self-discharge rate of lithium iron phosphate batteries.

[0007] The present invention adopts the following technical solutions:

[0008] A method for online quantitative evaluation of the micro-short circuit degree of a lithium iron phosphate battery comprises the following steps:

[0009] Step 1: Collect battery parameters during vehicle operation on a certain date D(i), including the highest cell voltage V max , Minimum single cell voltage V min , current I, time t.

[0010] Step 2: Calculate and process the current I and time t array collected in step 1 to obtain the capacity value Q(k) at each moment. Then, by building a battery model and parameter identification, obtain the open circuit voltage data corresponding to each voltage data of D(i) on the day. max The processed result is recorded as U oc1 (k), by V min The processed result is recorded as U oc2 (k).

[0011] Step 3. Solve L: Use Gaussian function to solve U oc1 (k) Fit or perform smoothing and use the formula dU oc1 (k)=U oc1 (k)-U oc1 (k-1) Solve for dU oc1 ; Starting from k=1, dU oc1 (k) performs a threshold judgment. When it is less than or equal to a, it proceeds to k+1 to continue the judgment. When it is greater than a, L=k is specified and the judgment ends. a is the control parameter for ending the difference calculation, which is set according to the characteristics of the battery open circuit voltage curve.

[0012] Step 4. Solve S(i): Substitute the corresponding values solved in steps 2 and 3 into the following formula Calculate S(i); where b is the U solved in step 2 oc1 (k), U oc2 The difference between the two errors at (k) is set according to the characteristics of the battery model and can also be simplified to 0; c is the conversion coefficient corresponding to the characteristics of the battery open circuit voltage.

[0013] Step 5. Calculate SDR: Use S(i) and D(i) obtained in step 4 to perform a linear fit. The resulting slope is the SDR.

[0014] In a preferred embodiment, the above step 1 of collecting battery-related parameter data begins when the battery system is fully charged, that is, SOC ≥ 99%, and the data collection ends when the vehicle ends operation and before charging begins.

[0015] In a preferred embodiment, the highest cell voltage V max It can be the highest cell voltage value of all cells in the battery system at any moment, the voltage value of the cell with the highest state of charge or the best state of health (SOH), the voltage value of a specific cell, the average voltage value of all cells, or the average voltage value of a group of normal cells; the lowest cell voltage V min It can be the lowest cell voltage value of all cells in the battery system at any moment, the voltage value of the cell with the lowest state of charge or the worst state of health (SOH), or the voltage value of a specific cell.

[0016] In a preferred embodiment, the capacity value Q(k) at each moment in the above step 2 is obtained by processing according to the following two formulas: Q(1) = Q0-(1), Q(k) = Q(k-1)-current(k)*[t(k)-t(k-1)] / 3600-(2); wherein k is a serial number from 1 to N, N is the total number of time t arrays, and Q0 is the rated capacity of the battery system.

[0017] In a preferred embodiment, the battery model in the above step 2 can be a battery equivalent circuit model or an electrochemical model; the parameter identification is any algorithm that can identify the OCV, which can be any one of the least squares identification algorithm, Kalman filter algorithm, H infinity algorithm, and intelligent machine learning optimization algorithm.

[0018] In a preferred embodiment, the above step 3 is used to fit U oc1 The function of (k) can be a Gaussian function; it can also be a polynomial, a hyperbolic tangent function; it can also be no function fitting, and only the original array is smoothed.

[0019] In a preferred embodiment, the purpose of solving L in step 3 above is to locate U oc1 (k) The position where the first platform ends or the second platform begins can also be designed as dU(k) = U oc1 (k)-U oc2 (k) is then obtained by finding the trough or foot position after the peak of dU appears.

[0020] In a preferred embodiment, the values of a, b, and c in steps three and four can be obtained through experimental testing methods. Alternatively, the SDR model of steps one to four can be constructed through parameter optimization and solving methods, and the known SDR values can be obtained and the corresponding error objective function can be established and solved using an intelligent algorithm. Alternatively, a neural network can be constructed using a deep learning method and solved through data training. The intelligent algorithm is any one of a traversal algorithm, a genetic algorithm, a particle swarm algorithm, and an intelligent machine learning optimization algorithm. A, b, and c are all parameters with mV as the dimension, and their values are all greater than 0. For lithium iron phosphate battery systems, the value of c is between 0.3 and 0.4 based on the characteristics of the battery model.

[0021] It can be seen from the above description of the present invention that, compared with the prior art, the present invention has the following advantages:

[0022] First, the present invention collects the current and voltage data of two cells during vehicle driving on a specific date, builds a battery model, and identifies the open-circuit voltage of the two cells. After fitting or smoothing, the difference between the two cells' open-circuit voltages is integrated with a fixed point a as the cutoff. By controlling parameters such as b and c, an S value is obtained that stably reflects the characteristics of micro-short circuits. By calculating the rate of change of S with date, the self-discharge rate of the lithium iron phosphate battery can be obtained. Because the S value is not affected by factors such as charge rate, temperature, model error, and depth of discharge, the present invention can sensitively reflect the occurrence of micro-short circuits, achieving quantification and high-precision estimation of the short circuit extent. This allows for accurate identification of micro-short circuits, resolving the low accuracy of online self-discharge rate estimation for lithium iron phosphate batteries in existing solutions, allowing more time for repair of problematic batteries and protecting vehicle property.

[0023] 2. The present invention is not limited by working conditions. There is no need to wait for the battery to be discharged to a certain depth or to meet certain static conditions. It does not rely on charging data, does not require disassembly of the battery box, and does not require long-term static storage. It is suitable for battery systems with relatively flat charge and discharge voltage platforms, such as battery systems with positive or negative electrodes containing materials such as lithium iron phosphate and lithium titanate. In addition, the calculation process is less time-consuming and suitable for online estimation.

[0024] 3. The present invention can realize micro short circuit monitoring by extracting the highest and lowest cell voltages (or the voltage of the cell with the lowest SOC) from the operating data of the day, without collecting the voltage value of each cell, reducing the data transmission load and simplifying the calculation. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 Flowchart of the present invention. DETAILED DESCRIPTION

[0026] Refer to the following Figure 1The following describes specific embodiments of the present invention. Numerous details are provided below to provide a comprehensive understanding of the present invention, but those skilled in the art will appreciate that the present invention can be implemented without these details. Well-known components, methods, and processes are not described in detail below.

[0027] This embodiment provides a method for online quantitative assessment of the degree of micro-short circuit of a lithium iron phosphate battery. By collecting the current and voltage data of two cells during vehicle driving conditions on a certain date, the open-circuit voltage of the two cells is identified using a constructed battery model. After fitting or smoothing, the difference between the open-circuit voltages of the two cells is integrated with a fixed point a as the cutoff point. By controlling parameters such as b and c, an S value that can stably reflect the micro-short circuit characteristics is obtained. By calculating the rate of change of S with date, the self-discharge rate value of the lithium iron phosphate battery can be obtained.

[0028] The method for online quantitative evaluation of the micro-short circuit degree of lithium iron phosphate batteries of the present invention comprises the following specific steps:

[0029] Step 1: Select a date D(i) and collect battery parameters during vehicle operation on that day, including the highest cell voltage V max , Minimum single cell voltage V min , current I, time t, etc.; among them, the time when data collection starts requires the battery system to be fully charged, that is, SOC>=99%, and the time when data collection ends is the time when the vehicle ends operation and starts to be charged.

[0030] Step 2: Process the current I and time t arrays in the data collected in step 1 according to formulas (1)-(2) to obtain the capacity value Q(k) at each moment, where: k is a serial number from 1 to N, N is the total number of time t arrays, and Q0 is the rated capacity of the battery system.

[0031] Q(1)=Q0 (1)

[0032] Q(k)=Q(k-1)-current(k)*[t(k)-t(k-1)] / 3600 (2)

[0033] Step 3: The data obtained in steps 1 and 2 are used to build a battery model and parameter identification algorithm to obtain the open circuit voltage data corresponding to each voltage data of D(i). max The processed result is recorded as U oc1 (k), by V min The processed result is recorded as U oc2 (k).

[0034] Step 4. Solve L: Use Gaussian function to solve U oc1 (k) Fit or perform smoothing and use formula (3) to solve dUoc1 .

[0035] dU oc1 (k)=U oc1 (k)-U oc1 (k-1) (3)

[0036] Starting from k=1, oc1 (k) performs a threshold judgment. When it is less than or equal to a, the judgment is continued at k+1. When it is greater than a, L=k is set and the judgment ends. a is the control parameter for ending the difference calculation, which is set according to the characteristics of the battery open circuit voltage curve.

[0037] Step 5. Solve S(i): Substitute the corresponding values solved in steps 2 to 4 into formula (4) to calculate S(i);

[0038]

[0039] Among them, b is the U solved in step 3 oc1 (k), U oc2 The difference between the two errors at (k) is set according to the characteristics of the battery model and can also be simplified to 0; c is the conversion coefficient corresponding to the characteristics of the battery open circuit voltage.

[0040] Step 6. Calculate SDR: Use S(i) and D(i) obtained in step 5 to perform a linear fit. The resulting slope is the SDR.

[0041] Step 7: Set the corresponding short circuit warning level according to the SDR size to implement short circuit warning.

[0042] The battery system mentioned above can be a new energy vehicle battery system or an energy storage system.

[0043] The highest cell voltage V in step 1 above max It can be the highest cell voltage value of all cells in the battery system at any moment, or the voltage value of the cell with the highest state of charge or the best state of health (SOH), the voltage value of a specific cell, or the average voltage value of all cells or the average voltage value of a group of normal cells. The lowest cell voltage V min It can be the lowest cell voltage value of all cells in the battery system at any moment, or the voltage value of the cell with the lowest state of charge or the worst state of health (SOH), or the voltage value of a specific cell.

[0044] The current value I in the above step 1 is set to be a negative value for charging and a positive value for discharging; the current value I can also be set to be a positive value for charging and a negative value for discharging, then the formula (1) is adjusted to Q(1)=-Q0.

[0045] The initial value of Q in step 2 is set to the rated capacity value, or it can be set to 0 or other values, and the associated formulas are changed accordingly. For example, when the initial value of Q is 0, all Qs in the above formulas are replaced by (Q0-Q).

[0046] The battery model in the above step three can be a battery equivalent circuit model or an electrochemical model; the parameter identification is any algorithm that can identify the OCV, which can be any one of the least squares identification algorithm, Kalman filter algorithm, H infinity algorithm, and intelligent machine learning optimization algorithm.

[0047] The above step 4 is used to fit U oc1 The function of (k) can be a Gaussian function, or other functions such as a polynomial, hyperbolic tangent, etc.; it can also be no function fitting, and only the original array is smoothed.

[0048] The purpose of solving L in step 4 is to locate U oc1 (k) The position where the first platform ends or the second platform begins can also be designed by dU(k) = U oc1 (k)-U oc2 (k) is obtained by finding the trough or peak foot position after the peak of dU appears.

[0049] The values of a, b, and c in steps 4 and 5 above can be obtained through experimental testing methods; they can also be solved through parameter optimization methods by building an SDR model from steps 1 to 5, obtaining a certain amount of known SDR values, and establishing a corresponding error objective function using an intelligent algorithm, or by building a neural network using methods such as deep learning, and solving them through data training. Among them, the intelligent algorithm can be any method that can be solved, such as a traversal algorithm, a genetic algorithm, a particle swarm algorithm, an intelligent machine learning optimization algorithm, etc., where a, b, and c are all parameters with mV as the dimension, and their values are all greater than 0. For lithium iron phosphate battery systems, the value of c is between 0.3 and 0.4 based on the characteristics of the battery model.

[0050] The above is only a specific implementation of the present invention, but the design concept of the present invention is not limited to this. Any non-substantial changes to the present invention using this concept shall be deemed as an infringement of the protection scope of the present invention.

Claims

1. A method for online quantitative evaluation of the micro-short circuit degree of lithium iron phosphate batteries, characterized in that: The following steps are involved: Step 1: Collect battery parameters during vehicle operation on a certain date D(i), including the highest cell voltage V max , Minimum single cell voltage V min , current I, time t; Step 2: Calculate and process the current I and time t array collected in step 1 to obtain the capacity value Q(k) at each moment. Then, by building a battery model and parameter identification, obtain the open circuit voltage data corresponding to each voltage data of D(i) on the day. max The processed result is recorded as U oc1 (k), by V min The processed result is recorded as U oc2 (k); The capacity value Q(k) at each moment is obtained by processing according to the following two formulas: Q(1)=Q0-(1), ; Where k is a serial number from 1 to N, N is the total number of arrays at time t, and Q0 is the rated capacity of the battery system; Step 3. Solve L: Use Gaussian function to solve U oc1 (k) Fit or perform smoothing and use the formula dU oc1 (k)=U oc1 (k)-U oc1 (k-1) Solve for dU oc1 ; Starting from k=1, dU oc1 (k) performs a threshold judgment. When it is less than or equal to a, the process proceeds to k+1 to continue the judgment. When it is greater than a, L=k is specified and the judgment ends. a is the control parameter for ending the difference calculation, which is set according to the characteristics of the battery open circuit voltage curve. Step 4. Solve S(i): Substitute the corresponding values solved in steps 2 and 3 into the following formula , calculate S(i); where b is the U solved in step 2 oc1 (k), U oc2 The difference between the two errors at (k) is set according to the characteristics of the battery model; c is the conversion coefficient corresponding to the characteristics of the battery open circuit voltage; Step 5. Calculate SDR: Use S(i) and D(i) obtained in step 4 to perform a linear fit. The resulting slope is the SDR.

2. The method for online quantitative evaluation of the micro-short circuit degree of a lithium iron phosphate battery according to claim 1, characterized in that: The step 1 of collecting battery-related parameter data requires that the battery system is fully charged, that is, SOC ≥ 99%, and the data collection ends when the vehicle ends operation and before charging begins.

3. The method for online quantitative evaluation of the micro-short circuit degree of a lithium iron phosphate battery according to claim 1, characterized in that: The highest cell voltage V in step 1 max , is the highest cell voltage value of all cells in the battery system at any moment, or the voltage value of the cell with the highest state of charge or the best state of health (SOH), or the average voltage value of all cells or the average voltage value of the normal cell group; the lowest cell voltage V min , which is the lowest cell voltage value of all cells in the battery system at any moment, or the voltage value of the cell with the lowest state of charge or the worst state of health (SOH).

4. The method for online quantitative evaluation of the micro-short circuit degree of a lithium iron phosphate battery according to claim 1, characterized in that: The battery model in the step 2 is a battery equivalent circuit model or an electrochemical model, and the parameter identification is any one of a least squares identification algorithm, a Kalman filter algorithm, an H-infinity algorithm, and an intelligent machine learning optimization algorithm.

5. The method for online quantitative evaluation of the micro-short circuit degree of a lithium iron phosphate battery according to claim 1, characterized in that: The step three is used to fit U oc1 The function of (k) is a Gaussian function; or a polynomial; or a hyperbolic tangent function.

6. The method for online quantitative evaluation of the micro-short circuit degree of a lithium iron phosphate battery according to claim 1, characterized in that: The purpose of solving L in step 3 is to locate U oc1 (k) The position where the first platform ends or the second platform begins, using the design dU(k)=U oc1 (k)-U oc2 (k) is then obtained by finding the trough or foot position after the peak of dU appears.

7. The method for online quantitative evaluation of the micro-short circuit degree of a lithium iron phosphate battery according to claim 1, characterized in that: The appropriate values of a, b and c in steps three and four are obtained through experimental testing methods; or the SDR model of steps one to four is constructed through parameter optimization and solution methods, the known SDR value is obtained and the corresponding error objective function is established and solved using an intelligent algorithm; or a neural network is constructed using a deep learning method and solved through data training.

8. The method for online quantitative evaluation of the micro-short circuit degree of a lithium iron phosphate battery according to claim 7, characterized in that: The intelligent algorithm is any one of a traversal algorithm and an intelligent machine learning optimization algorithm, wherein a, b and c are all parameters with mV as the dimension, and their values are all greater than 0. For the lithium iron phosphate battery system, the value of c is between 0.3 and 0.4 based on the characteristics of the battery model.

9. The method for online quantitative evaluation of the micro-short circuit degree of a lithium iron phosphate battery according to claim 1, characterized in that: b in the step 4 is simplified to 0.

Images