Lithium battery pack multi-fault diagnosis method, system and device based on local weighted Manhattan distance and medium

Abstract

The invention discloses a lithium battery pack multi-fault diagnosis method, system and device based on a local weighted Manhattan distance and a medium, and relates to the technical field of battery application. The method comprises the following steps: acquiring voltage curves of each single battery in a target battery pack in a charging stage and a discharging stage; calculating each voltage curve in the charging stage by using a traditional Manhattan distance algorithm to obtain a first deviation value; calculating each voltage curve in the discharge stage by using a local weighted Manhattan distance algorithm to obtain a second deviation value; wherein the deviation value is deviation data between each voltage curve and a normal voltage curve in the charging or discharging stage; and comparing the first deviation value and the second deviation value with a set evaluation threshold range, and determining whether a fault single battery exists in the target battery pack or not and positioning the fault single battery. According to the invention, detection of low-SOC and low-capacity mixed faults, internal resistance faults, connection faults and external short-circuit faults of the battery can be realized.

Description

Technical Field

[0001] The present invention relates to the field of battery application technology, and in particular to a lithium battery pack multi-fault diagnosis method, system, device and medium based on local weighted Manhattan distance. Background Art

[0002] At present, the research on lithium-ion battery fault diagnosis methods mainly includes three categories: knowledge-based fault diagnosis method, data-driven fault diagnosis method, and model-based fault diagnosis method. Among them, the knowledge-based fault diagnosis method is a method that uses expert knowledge and rules to identify and locate system faults, which can be subdivided into expert systems, graph theory, and fuzzy logic. Expert system is a knowledge-based computer program that aims to simulate and apply the knowledge and reasoning ability of experts in the field, but it has limitations such as difficulty in knowledge acquisition and poor adaptability. Graph theory generates fault diagnosis pattern diagrams based on fault propagation and impact in the battery system. The results are easy to understand, but there are problems such as difficulty in knowledge acquisition and modeling, complex calculations, and data accuracy. Fuzzy logic can handle information containing uncertainty and ambiguity in battery fault diagnosis, but due to the complex fault mechanism of the power battery system and the difficulty in expressing and acquiring knowledge, it is a huge challenge to establish an effective knowledge system. Both data-driven fault diagnosis methods and model-based fault diagnosis methods have problems such as only studying a single charging stage and no mixed faults. Summary of the invention

[0003] The purpose of the present invention is to provide a lithium battery pack multi-fault diagnosis method, system, device and medium based on local weighted Manhattan distance, which can detect mixed faults of low SOC and low capacity of the battery, internal resistance faults, connection faults and external short circuit faults.

[0004] To achieve the above object, the present invention provides the following solutions:

[0005] A multi-fault diagnosis method for lithium battery pack based on local weighted Manhattan distance, comprising:

[0006] Obtaining voltage curves of each single cell in the target battery pack during the charging and discharging stages;

[0007] Using a conventional Manhattan distance algorithm to calculate each voltage curve in the charging stage, a first deviation value is obtained; the first deviation value is the deviation data between each voltage curve in the charging stage and a normal voltage curve;

[0008] Using a local weighted Manhattan distance algorithm to calculate each voltage curve in the discharge stage, a second deviation value is obtained; the second deviation value is the deviation data between each voltage curve and a normal voltage curve in the discharge stage;

[0009] The first deviation value and the second deviation value are respectively compared with a set evaluation threshold range to determine whether there is a faulty single cell in the target battery group and the location of the faulty single cell.

[0010] Optionally, the expression of the traditional Manhattan distance algorithm is:

[0011] d (i,j) =|V 1,i -V 1,j |+|V 2,i -V 2,j |+…+|V n,i -V n,j

[0012] Among them, d (i,j) is the Manhattan distance between single cell i (i=1,2,…,m) and single cell j (j=1,2,…,m) in the battery pack, n represents the first voltage sampling time, and m represents the total number of single cells.

[0013] Optionally, the expression of the weighting coefficient in the local weighted Manhattan distance algorithm is:

[0014]

[0015] Among them, C i Indicates the rated capacity of single cell i in the battery pack, C j represents the rated capacity of single cell j, α represents the discharge rate factor affecting the discharge rate, β represents the temperature coefficient factor affecting the temperature, I d Represents the discharge current under dynamic conditions, represents the temperature of the single cell i of the battery pack at time t, represents the temperature of the single cell j of the battery pack at time t, n represents the first voltage sampling time, and k represents the second voltage sampling time.

[0016] Optionally, the expression of the local weighted Manhattan distance algorithm is:

[0017] d dis(i,j) =w (i,j) |V n,i -V n,j |+…+w (i,j) |V k,i -V k,j

[0018] Among them, d dis(i,j) represents the Manhattan distance between single cell i and single cell j in the battery pack, w (i,j) Represents the weight relationship between single cell i and single cell j in the battery pack, V n,iRepresents the sampled voltage of single cell i at time n, V n,j represents the sampled voltage of single cell j at time n, V k,i represents the sampled voltage of single cell i at time k, V k,j Represents the sampled voltage of single cell j at time k.

[0019] Optionally, the set evaluation threshold range is (0, 0.17).

[0020] The present invention also provides a lithium battery pack multi-fault diagnosis system based on local weighted Manhattan distance, comprising:

[0021] A data sampling unit, used to obtain the voltage curve of each single battery in the target battery pack during the charging stage and the discharging stage;

[0022] A charging stage deviation calculation unit, used to calculate each voltage curve in the charging stage using a traditional Manhattan distance algorithm to obtain a first deviation value; the first deviation value is deviation data between each voltage curve and a normal voltage curve in the charging stage;

[0023] A discharge phase deviation calculation unit, used to calculate each voltage curve in the discharge phase using a local weighted Manhattan distance algorithm to obtain a second deviation value; the second deviation value is deviation data between each voltage curve and a normal voltage curve in the discharge phase;

[0024] The fault diagnosis unit is used to compare the first deviation value and the second deviation value with a set evaluation threshold range respectively, to determine whether there is a faulty single cell in the target battery pack and the location of the faulty single cell.

[0025] The present invention also provides an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the above-mentioned lithium battery pack multi-fault diagnosis method based on local weighted Manhattan distance.

[0026] The present invention also provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the lithium battery pack multi-fault diagnosis method based on local weighted Manhattan distance as described above.

[0027] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0028] The present invention discloses a method, system, device and medium for multi-fault diagnosis of lithium battery packs based on local weighted Manhattan distance, the method comprising obtaining voltage curves of each single cell in the target battery pack during the charging stage and the discharging stage; calculating each voltage curve in the charging stage using the traditional Manhattan distance algorithm to obtain a first deviation value; calculating each voltage curve in the discharging stage using the local weighted Manhattan distance algorithm to obtain a second deviation value; wherein the deviation value is the deviation data between each voltage curve and the normal voltage curve in the charging or discharging stage; comparing the first deviation value and the second deviation value with the set evaluation threshold range respectively to determine whether there is a faulty single cell in the target battery pack and the location of the faulty single cell. The present invention can realize the detection of mixed faults, internal resistance faults, connection faults and external short circuit faults of low SOC and low capacity of batteries. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0030] Figure 1 It is a schematic diagram of the flow chart of the multi-fault diagnosis method of lithium battery pack of the present invention;

[0031] Figure 2 Schematic diagram of Euclidean distance and Manhattan distance in this embodiment;

[0032] Figure 3 This is a Manhattan distance calculation diagram for the charging stage curve in this embodiment. DETAILED DESCRIPTION

[0033] 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.

[0034] The purpose of the present invention is to provide a lithium battery pack multi-fault diagnosis method, system, device and medium based on local weighted Manhattan distance, which can detect mixed faults of low SOC and low capacity of the battery, internal resistance faults, connection faults and external short circuit faults.

[0035] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0036] like Figure 1 As shown, the present invention provides a lithium battery pack multi-fault diagnosis method based on local weighted Manhattan distance, comprising:

[0037] Step 100: Obtain voltage curves of each single battery in the target battery pack during the charging stage and the discharging stage.

[0038] Step 200: Calculate each voltage curve in the charging stage using a conventional Manhattan distance algorithm to obtain a first deviation value; the first deviation value is deviation data between each voltage curve in the charging stage and a normal voltage curve.

[0039] Step 300: Calculate each voltage curve in the discharge stage using a local weighted Manhattan distance algorithm to obtain a second deviation value; the second deviation value is deviation data between each voltage curve in the discharge stage and a normal voltage curve.

[0040] Step 400: Compare the first deviation value and the second deviation value with a set evaluation threshold range respectively to determine whether there is a faulty single cell in the target battery pack and the location of the faulty single cell.

[0041] Based on the above technical solution, the following specific implementation methods are provided.

[0042] Constructing a multi-fault diagnosis strategy for lithium battery packs: Based on the Manhattan distance evaluation of the battery pack voltage sequence, a local weighted Manhattan distance evaluation method is proposed, and the fault types of single faults and mixed faults are detected through a voltage ratio-based outlier comprehensive analysis method. This example will specifically explain its theory and give a diagnostic idea, and propose the overall process of the multi-fault diagnosis method for lithium-ion battery packs.

[0043] As a more specific processing step, the curve Manhattan distance calculation is first performed.

[0044] Manhattan distance is the distance between two points expressed by the sum of the absolute values ​​of the coordinate differences of the relevant points in two-dimensional space. The larger the distance, the greater the difference between the samples and the less similar the samples are; conversely, the smaller the distance, the smaller the difference between the samples and the higher the similarity between the samples. The Manhattan distance between two points a and b in two-dimensional space is calculated by formula (1):

[0045] d (a,b) =|x a -x b |+|y a -y b | (1)

[0046] Among them, d (a,b) Represents a point (x a ,ya ) and point (x b ,y b ) is the Manhattan distance between .

[0047] Manhattan distance indicates the position of two points relative to the coordinate axis. It does not compress the original distance and reflects the real difference between the two points. Figure 2 As shown, A represents Manhattan distance, B represents Euclidean distance, and C represents equivalent Manhattan distance.

[0048] In order to quantify the distance between two curves of the same length in two-dimensional space, the traditional Manhattan distance is improved to quantify the Manhattan distance between two curves of equal length.

[0049] Assume that l1 and l2 are two equal-length curves in two-dimensional space. The distance between l1 and l2 is calculated by the curve Manhattan distance, and the formula is: d = |x1-x′1|+|y1-y′1|+…+|x n -x′ n |+|y n -y′ n |.

[0050] Where (x n ,y n ) and (x' n ,y' n ) are the n points that constitute the curves l1 and l2 respectively; d represents the Manhattan distance between the curves l1 and l2. For multi-fault diagnosis of lithium-ion battery packs, assuming that the terminal voltage curve data of each monomer in the charging and discharging stage of the lithium-ion battery pack is obtained as shown in formula (2), the specific detection and positioning steps are as follows:

[0051]

[0052] In the formula [V 1,m ,…V n,m ] T is the mth battery voltage sequence in the charging stage of the lithium-ion battery pack, [V n,m ,…V k,m ] T is the mth battery voltage sequence in the discharge phase of the lithium-ion battery pack; V n,m is the sampled voltage of the mth battery at time n during the charging phase, V k,m is the sampled voltage of the mth battery at time k during the discharge phase. m is the total number of batteries in the battery pack.

[0053] According to the terminal voltage curve data of each battery, the Manhattan distance between battery cells during the charging stage is calculated:

[0054] d (i,j) =|V 1,i -V1,j |+|V 2,i -V 2,j |+…+|V n,i -V n,j | (3)

[0055] Among them, d (i,j) is the Manhattan distance between single cell i (i = 1, 2, ..., m) and single cell j (j = 1, 2, ..., m) in the battery pack, n represents the first voltage sampling time, and m represents the total number of single cells. The Manhattan matrix of the lithium battery pack in the charging stage is established as shown in formula (4).

[0056]

[0057] In formula (4), d' (i,j) is the normalized Manhattan distance between cells i and j in a lithium-ion battery pack, where cells i and j are both cells in the battery pack, and d' (i,j) =d' (j,i) .

[0058] like Figure 3 As shown, assuming that curve a, curve b, and curve c are three charging stage voltage curves, the Manhattan distance D between curve a and curve b is D = ΔU ab ; The Manhattan distance D′ between curve b and curve c can be expressed as D′=ΔU bc1 +|ΔU bc2 |+|ΔU bc3 |.

[0059] Then, the local weighted Manhattan distance is constructed: because the battery terminal voltage under dynamic conditions is affected by the discharge rate and temperature, there are large differences between the single-cell voltages in the battery pack, which makes the normal battery present an abnormality similar to that of the faulty battery during the discharge stage. Therefore, this embodiment adopts the local weighted Manhattan distance method during the discharge stage, and reduces the voltage anomaly caused by the discharge rate and temperature by assigning different weights to different battery cells, and performs accurate fault diagnosis through the curve Manhattan distance.

[0060] Assume that α and β are the discharge rate factor and temperature coefficient factor that affect the discharge rate and temperature respectively. i , C j Indicates the rated capacity of the single cell in the battery pack, I d represents the discharge current under dynamic conditions; t represents the duration of the discharge process, represents the temperature of the single cell i of the battery pack at time t, as shown in formula (7):

[0061]

[0062] In the formula, the capacity difference and temperature difference of the discharge rate between the single cells are quantified into the form of distance difference, and then the weighting coefficient W is derived, where α and β must satisfy 0≤α, β≤1. From the above, it can be seen that the weighted Manhattan distance of the discharge stage is obtained by combining the discharge rate and temperature factors:

[0063] d dis(i,j) =w (i,j) |V n,i -V n,j |+…+w (i,j) |V k,i -V k,j | (8)

[0064] Among them, d dis(i,j) represents the Manhattan distance between single cell i and single cell j in the battery pack, w (i,j) Represents the weight relationship between single cell i and single cell j in the battery pack, V n,i Represents the sampled voltage of single cell i at time n, V n,j represents the sampled voltage of single cell j at time n, V k,i represents the sampled voltage of single cell i at time k, V k,j Represents the sampled voltage of single cell j at time k.

[0065] The local weighted Manhattan matrix of the lithium-ion battery pack is established as shown in formula (9):

[0066]

[0067] where d' dis(i,j) is the normalized local weighted Manhattan distance between cells i and j in a lithium-ion battery pack. In addition to the above voltage distance evaluation for charging and discharging, the following steps need to be clarified:

[0068] For batteries in a lithium-ion battery pack, the normalized distance between two cells is mapped into a smaller range, and the normalized Manhattan distance is usually between (0,1).

[0069] (1) During the charging phase, the Manhattan distance of a normal battery is usually around 0.1. The selection of the threshold is related to the standardized Manhattan distance of the normal battery. In order to reduce missed faults and ensure timely warning, the threshold must be no less than the Manhattan distance of the normal battery. After multiple calculations and tests, the threshold range is set between (0, 0.17).

[0070] (2) Due to the influence of discharge rate, temperature and other factors, the Manhattan distance range in the discharge stage is much larger than that of a normal battery. The local weighted Manhattan distance method is used to evaluate and ensure that the threshold in the discharge stage is consistent with that in the charging stage.

[0071] For batteries in a lithium-ion battery pack, the voltage curve trends during the charging phase are similar. Due to factors such as temperature and production process, the characteristics of battery data are inconsistent, which will be further exacerbated under dynamic conditions.

[0072] Therefore, this embodiment proposes an evaluation method based on local weighted Manhattan distance to solve the voltage discharge rate difference of the battery in the discharge stage, so as to achieve the same threshold evaluation standard as the charging stage. The integrity of the Manhattan distance evaluation is guaranteed by integrating the charging and discharging stages. The deviation between the complete charge and discharge voltage curve of the faulty battery cell and the normal curve is quantified to locate and detect the faulty cell in the battery pack.

[0073] When there are different faults in the battery pack, the faulty cell will show a voltage deviation different from the normal voltage. By using different Manhattan distance methods to evaluate the battery pack during the charge and discharge phase and quantify its deviation, the fault can be easily, quickly and reliably located and detected.

[0074] This embodiment improves on the Manhattan distance of the curve in the charging stage and innovatively adds the local weighted Manhattan distance in the discharging stage. While accurately quantifying the changes in the voltage curves between batteries, it solves the limitation of only using the voltage in the charging stage in fault diagnosis and uses the complete charging and discharging process to locate multiple faults of the lithium-ion battery pack.

[0075] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0076] This article uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only used to help understand the core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A multi-fault diagnosis method for lithium battery pack based on local weighted Manhattan distance, characterized in that: include: Obtaining voltage curves of each single cell in the target battery pack during the charging and discharging stages; Using a conventional Manhattan distance algorithm to calculate each voltage curve in the charging stage, a first deviation value is obtained; the first deviation value is the deviation data between each voltage curve in the charging stage and a normal voltage curve; Using a local weighted Manhattan distance algorithm to calculate each voltage curve in the discharge stage, a second deviation value is obtained; the second deviation value is the deviation data between each voltage curve and a normal voltage curve in the discharge stage; The first deviation value and the second deviation value are respectively compared with a set evaluation threshold range to determine whether there is a faulty single cell in the target battery group and the location of the faulty single cell.

2. The lithium battery pack multi-fault diagnosis method based on local weighted Manhattan distance according to claim 1, characterized in that: The expression of the traditional Manhattan distance algorithm is: d (i,j) =|V 1,i -V 1,j |+|V 2,i -V 2,j |+…+|V n,i -V n,j | Among them, d (i,j) is the Manhattan distance between single cell i (i=1,2,…,m) and single cell j (j=1,2,…,m) in the battery pack, n represents the first voltage sampling time, and m represents the total number of single cells.

3. The lithium battery pack multi-fault diagnosis method based on local weighted Manhattan distance according to claim 1, characterized in that: The expression of the weighted coefficient in the local weighted Manhattan distance algorithm is: Among them, C i Indicates the rated capacity of single cell i in the battery pack, C j represents the rated capacity of single cell j, α represents the discharge rate factor affecting the discharge rate, β represents the temperature coefficient factor affecting the temperature, I d Represents the discharge current under dynamic conditions, represents the temperature of the single cell i of the battery pack at time t, represents the temperature of the single cell j of the battery pack at time t, n represents the first voltage sampling time, and k represents the second voltage sampling time.

4. The lithium battery pack multi-fault diagnosis method based on local weighted Manhattan distance according to claim 3 is characterized in that: The expression of the local weighted Manhattan distance algorithm is: d dis(i,j) =w (i,j) |V n,i -V n,j |+…+w (i,j) |V k,i -V k,j | Among them, d dis(i,j) represents the Manhattan distance between single cell i and single cell j in the battery pack, w (i,j) Represents the weight relationship between single cell i and single cell j in the battery pack, V n,i Represents the sampled voltage of single cell i at time n, V n,j represents the sampled voltage of single cell j at time n, V k,i represents the sampled voltage of single cell i at time k, V k,j Represents the sampled voltage of single cell j at time k.

5. The lithium battery pack multi-fault diagnosis method based on local weighted Manhattan distance according to claim 1, characterized in that: The evaluation threshold range is set to (0, 0.17).

6. A lithium battery pack multi-fault diagnosis system based on local weighted Manhattan distance, characterized in that: include: A data sampling unit, used to obtain the voltage curve of each single battery in the target battery pack during the charging stage and the discharging stage; A charging stage deviation calculation unit, used to calculate each voltage curve in the charging stage using a traditional Manhattan distance algorithm to obtain a first deviation value; the first deviation value is deviation data between each voltage curve and a normal voltage curve in the charging stage; A discharge phase deviation calculation unit, used to calculate each voltage curve in the discharge phase using a local weighted Manhattan distance algorithm to obtain a second deviation value; the second deviation value is deviation data between each voltage curve and a normal voltage curve in the discharge phase; The fault diagnosis unit is used to compare the first deviation value and the second deviation value with a set evaluation threshold range respectively, to determine whether there is a faulty single cell in the target battery pack and the location of the faulty single cell.

7. An electronic device, characterized in that: It comprises a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the lithium battery pack multi-fault diagnosis method based on local weighted Manhattan distance according to any one of claims 1-5.

8. A computer-readable storage medium, characterized in that: It stores a computer program, and when the computer program is executed by a processor, it implements the lithium battery pack multi-fault diagnosis method based on local weighted Manhattan distance as described in any one of claims 1 to 5.

Images