A sparse data battery state of health assessment method

Abstract

The application discloses a kind of sparse data battery health state evaluation method, it is related to new energy battery application technical field, the present application is based on the charging data and discharge data of sparse data acquisition battery system, using the characteristics that current fluctuation is smaller in charging process obtains the ampere-hour integral capacity SOC0 of charging process under sparse data condition, and using the characteristics that current fluctuation is larger in discharge process, open-circuit voltage is identified using battery model, the real SOC1 corresponding before charging is obtained by the open-circuit voltage-state of charge relationship conversion, then the results of two parts are added to obtain SOH.The technical scheme provided by the application can solve the problem that SOH cannot be estimated by sparse data in prior art, and the condition constraint on standing time before charging can be avoided, and the application value of sparse data is effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of new energy battery application technology, and in particular to a sparse data battery health status assessment method. Background Art

[0002] Battery State of Health (SOH) refers to the ratio of a battery's current available capacity after a full charge or discharge to its capacity when new. It's a key indicator of battery aging and influences its continued serviceability and retirement disposal. Using cloud-based battery data to diagnose battery health quickly and easily provides this information, playing a crucial role in battery health testing, warranty maintenance, and end-of-life assessments.

[0003] Existing battery health estimation methods fall into two main categories: one that relies on battery discharge data and the other on charging data. Methods that rely on discharge data primarily estimate SOH based on the simultaneous processing of open-circuit voltage and ampere-hour integrals, while methods that rely on charging data primarily use capacity differential curves or post-charge voltage curves to estimate SOH. While both methods offer high accuracy, they require a high data acquisition frequency and a long rest period before charging.

[0004] However, most companies currently upload battery cloud data at a low frequency (e.g., one record every 15 seconds or 10 seconds), failing to meet the application requirements of ampere-hour integration, significantly reducing the application value of sparse data. Furthermore, vehicles are often charged immediately after driving, making it difficult to obtain an accurate static open-circuit voltage before charging. This makes it difficult to accurately estimate the SOC before charging, making it difficult to accurately assess the SOH using sparse data. Summary of the Invention

[0005] The present invention provides a sparse data battery health status assessment method, the main purpose of which is to solve the problems of existing battery health estimation methods having high requirements on sparse data collection frequency and high requirements on pre-charging rest time.

[0006] The present invention adopts the following technical solutions:

[0007] A sparse data battery health status assessment method includes the following steps:

[0008] Step S1: Search the vehicle's historical storage data for the minimum state of charge (SOC) of the battery system within a set time. min , record this time as t2, search forward from t2 to the last charging end time, record it as t1, search backward from t2 to the charging end time, record it as t3;

[0009] Step S2: Collect the current I and time t during the time period from t2 to t3, and use the ampere-hour integration algorithm to calculate the SOC0 during this time period:

[0010]

[0011] Where: Q n is the rated capacity of the battery;

[0012] Step S3: Collect the current I and voltage V during the time period from t1 to t2, build a battery model, and use the parameter identification algorithm to calculate the open circuit voltage U corresponding to each voltage V. oc , take the last value as the OCV1 within the time period, and convert OCV1 into SOC1 based on the open circuit voltage-state of charge relationship established in advance;

[0013] Step S4: Add the SOC0 obtained in step S2 and the SOC1 obtained in step S3 to obtain the available capacity retention rate SOH.

[0014] Furthermore, the following steps are included: Step S5, changing the value target of the voltage V, repeating steps S3 and S4 to respectively obtain the available capacity retention rate SOH corresponding to different value targets, and thereby evaluating the capacity consistency of the battery system monomers.

[0015] Furthermore, the target value of voltage V includes the highest cell voltage value V max and the minimum single cell voltage V min ; When the voltage V in step S3 is the highest single voltage value V max When the SOH obtained in step S4 is the optimal cell available capacity retention rate SOH max ; When the voltage V in step S3 is the lowest single cell voltage value V min When the SOH obtained in step S4 is the battery system available capacity retention rate SOH sys .

[0016] Further, in step S5, according to the best cell available capacity retention rate SOH max and battery system available capacity retention rate SOH sys Calculate the battery system single cell capacity consistency.

[0017] Furthermore, the target value of voltage V includes the voltage V of different single cells. i , when the voltage V in step S3 is the voltage V of a single cell i When the SOH obtained in step S4 is the available capacity retention rate SOH of the battery cell i .

[0018] Further, in step S5, according to different cell available capacity retention rates SOHi Calculate the battery system single cell capacity consistency.

[0019] Furthermore, the battery system single cell capacity consistency is calculated based on the absolute value or dispersion of the difference in available capacity retention rate SOH corresponding to different voltage value targets.

[0020] Further, in step S1, it is determined whether the battery is fully charged at time t3. If so, the subsequent steps are executed. Otherwise, the minimum state of charge value SOC outside the time period from t1 to t3 is searched again. min , and redefine t1, t2, and t3 according to step S1.

[0021] Furthermore, in step S3, the parameter identification algorithm can be based on a battery equivalent circuit model, a fractional order model or an electrochemical model, and can be combined with any one of the least squares identification algorithm, the Kalman filter algorithm, the H-infinity algorithm, and the intelligent machine learning optimization algorithm.

[0022] Furthermore, in step S3, the open circuit voltage-state of charge relationship may be fitted using a Gaussian function, a polynomial function, or a hyperbolic tangent function, or may be fitted using a linear difference.

[0023] Compared with the prior art, the present invention has the following beneficial effects:

[0024] 1. The present invention collects charging and discharging data of a battery system based on sparse data, and utilizes the characteristic of small current fluctuation in the charging process to obtain the ampere-hour integrated capacity SOC0 of the charging process under sparse data conditions. It also utilizes the characteristic of large current fluctuation in the discharge process to identify the open circuit voltage using a battery model, and converts the open circuit voltage-state of charge relationship to obtain the corresponding actual SOC1 before charging, thereby realizing the estimation of the battery SOH, avoiding the conditional constraint on the rest time before charging, improving the versatility of the algorithm, and enhancing the utilization value of sparse data.

[0025] 2. The algorithm provided by the present invention can estimate multiple available capacity retention rates (SOH) of the battery system by simply changing different voltage values ​​during the discharge process, which is conducive to a preliminary assessment of the battery health status. After further processing of multiple available capacity retention rates (SOH), the consistency of the single cell capacity of the battery system can be obtained, thereby comprehensively and accurately conducting a systematic assessment of the battery health status. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a control flow chart of the present invention. DETAILED DESCRIPTION

[0027] The specific embodiments of the present invention are described below with reference to the accompanying drawings. In order to fully understand the present invention, many details are described below, but for those skilled in the art, the present invention can be implemented without these details.

[0028] Reference Figure 1 , a sparse data battery health status assessment method, comprising the following steps:

[0029] Step S1: Search the vehicle's historical storage data for the minimum state of charge (SOC) of the battery system within a set time. min , record this moment as t2, search forward from t2 to the last charging end time, record it as t1, search backward from t2 to the charging end time, record it as t3.

[0030] The battery system mentioned in this step can be a new energy vehicle battery system or an energy storage system. As a preferred solution, the time is set to within the last month. In order to obtain more accurate data, after searching for t3, it should first be determined whether the battery is fully charged at t3. If it is fully charged, it means that t3 meets the requirements and the subsequent steps can be performed. If it is not fully charged, re-search the minimum state of charge value SOC outside the time period from t1 to t3. min , and redefine t1, t2, and t3 according to step S1. Specifically, full charge means that the state of charge value SOC of the battery system is 100%.

[0031] Step S2: Collect the current I and time t from time t2 to t3, and use the ampere-hour integration algorithm to calculate the SOC0 during this time period. The data collected in this step is charging data, and its purpose is to use the characteristic of small current fluctuations during the charging process to obtain the ampere-hour integrated capacity SOC0 of the charging process under sparse data. The specific calculation formula is:

[0032]

[0033] Where: Q n is the rated capacity of the battery.

[0034] Step S3: Collect the current I and voltage V during the time period from t1 to t2, build a battery model, and use the parameter identification algorithm to calculate the open circuit voltage U corresponding to each voltage V. oc , take the last value as OCV1 in this time period, and convert OCV1 into SOC1 according to the open circuit voltage-state of charge relationship established in advance. The data collected in this step is discharge data, and its purpose is to use the battery model to calculate the open circuit voltage U by taking advantage of the large current fluctuation in the discharge process. oc Identification is performed, and then the actual SOC1 corresponding to before charging is obtained through the open circuit voltage-state of charge relationship.

[0035] The parameter identification algorithm can be based on a battery equivalent circuit model, fractional-order model, or electrochemical model, and can be combined with any of the following algorithms: least squares identification, Kalman filtering, H-infinity, or intelligent machine learning optimization. The open-circuit voltage-state-of-charge relationship can be fitted using a Gaussian, polynomial, or hyperbolic tangent function, or using linear interpolation.

[0036] Based on the characteristics of the parameter identification algorithm, it is also possible to collect only about 600 data points to ensure that there is a certain step size from the initial value to the true value, without collecting all the data in the time period from t1 to t2 for calculation.

[0037] In order to achieve a multi-dimensional evaluation of the battery health status, the voltage V in this step can be the highest single cell voltage V corresponding to each current I1 in the battery system during the time period from t1 to t2. max Or the lowest single voltage value V min , or you can specify the voltage V of a single cell i .

[0038] Step S4: Add the SOC0 obtained in step S2 and the SOC1 obtained in step S3 to obtain the available capacity retention rate SOH, that is:

[0039] SOH=SOC0+SOC1

[0040] Specifically, when the voltage V in step S3 is the highest cell voltage value V max When the SOH obtained in step S4 is the optimal cell available capacity retention rate SOH max ; When the voltage V in step S3 is the lowest single cell voltage value V min When the SOH obtained in step S4 is the battery system available capacity retention rate SOH sys ; When the voltage V in step S3 is the voltage V of a single cell i When the SOH obtained in step S4 is the available capacity retention rate SOH of the battery cell i It can be seen that the algorithm provided by the present invention can estimate multiple available capacity retention rates SOH of the battery system by only changing the target value of the voltage V, which is conducive to a preliminary assessment of the battery health status.

[0041] Step S5: Change the target value of voltage V, repeat steps S3 and S4 to respectively obtain the available capacity retention rate SOH corresponding to different target values, and thereby evaluate the capacity consistency of the battery system cells.

[0042] Specifically, this step calculates the battery system's cell capacity consistency based on the absolute value or dispersion of the difference in available capacity retention (SOH) corresponding to different voltage targets, thereby comprehensively evaluating the battery's state of health. This algorithm solves the problem of sparse data in existing solutions that prevent SOH estimation, improving the application value of sparse data and simultaneously obtaining battery system SOH estimation and capacity consistency diagnosis results.

[0043] When the voltage V is set to the highest single cell voltage V max and the minimum single cell voltage V min When the battery is used, the best battery capacity retention rate SOH can be used. max and battery system available capacity retention rate SOH sys Calculate the battery system single cell capacity consistency. When the voltage V is taken as the voltage V of different single cells i When the available capacity retention rate SOH of each battery cell is i Calculate the battery system single cell capacity consistency.

[0044] 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 sparse data battery health status assessment method, characterized by: The steps include: Step S1: Search the vehicle's historical storage data for the minimum state of charge (SOC) of the battery system within a set time. min , record this time as t2, search forward from t2 to the last charging end time, record it as t1, search backward from t2 to the charging end time, record it as t3; in this step, determine whether the battery is fully charged at t3. If so, execute the subsequent steps. Otherwise, re-search for the minimum state of charge value SOC outside the time period from t1 to t3 min , and redefine t1, t2 and t3 according to step S1; Step S2: Collect the current I and time t during the time period from t2 to t3, and use the ampere-hour integration algorithm to calculate the SOC0 during this time period: Where: Q n is the rated capacity of the battery; Step S3: Collect the current I and voltage V during the time period from t1 to t2, build a battery model, and use the parameter identification algorithm to calculate the open circuit voltage U corresponding to each voltage V. oc , take the last value as the OCV1 within the time period, and convert OCV1 into SOC1 based on the open circuit voltage-state of charge relationship established in advance; Step S4, adding the SOC0 obtained in step S2 and the SOC1 obtained in step S3 to obtain the available capacity retention rate SOH; Step S5: Change the target value of voltage V, repeat steps S3 and S4 to respectively obtain the available capacity retention rate SOH corresponding to different value targets, and thereby evaluate the battery system single cell capacity consistency; calculate the battery system single cell capacity consistency based on the absolute value or discreteness of the difference between the available capacity retention rates SOH corresponding to different voltage value targets.

2. The sparse data battery health status assessment method according to claim 1, characterized in that: The target value of voltage V includes the highest single cell voltage value V max and the minimum single cell voltage V min ; When the voltage V in step S3 is the highest single voltage value V max When the SOH obtained in step S4 is the optimal cell available capacity retention rate SOH max ; When the voltage V in step S3 is the lowest single cell voltage value V min When the SOH obtained in step S4 is the battery system available capacity retention rate SOH sys .

3. The sparse data battery health status assessment method according to claim 2, characterized in that: In step S5, according to the best cell available capacity retention rate SOH max and battery system available capacity retention rate SOH sys Calculate the battery system single cell capacity consistency.

4. The sparse data battery health status assessment method according to claim 1, characterized in that: The target value of voltage V includes the voltage V of different single cells i , when the voltage V in step S3 is the voltage V of a single cell i When the SOH obtained in step S4 is the available capacity retention rate SOH of the battery cell i .

5. The sparse data battery health status assessment method according to claim 3, characterized in that: In step S5, according to different cell available capacity retention rates SOH i Calculate the battery system single cell capacity consistency.

6. The sparse data battery health status assessment method according to claim 1, characterized in that: In step S3, the parameter identification algorithm can be based on a battery equivalent circuit model, a fractional order model or an electrochemical model, and can be combined with any one of the least squares identification algorithm, the Kalman filter algorithm, the H-infinity algorithm, and the intelligent machine learning optimization algorithm.

7. The sparse data battery health status assessment method according to claim 1, characterized in that: In step S3, the open circuit voltage-state of charge relationship is fitted by a Gaussian function, a polynomial function, or a hyperbolic tangent function formula fitting method, or a linear difference fitting method.

Images