CN114492182A - Method for predicting cycle life of battery pack in any topological structure - Google Patents
Method for predicting cycle life of battery pack in any topological structure Download PDFInfo
- Publication number
- CN114492182A CN114492182A CN202210054745.0A CN202210054745A CN114492182A CN 114492182 A CN114492182 A CN 114492182A CN 202210054745 A CN202210054745 A CN 202210054745A CN 114492182 A CN114492182 A CN 114492182A
- Authority
- CN
- China
- Prior art keywords
- parallel
- battery
- cycle life
- battery pack
- topological structure
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000013528 artificial neural network Methods 0.000 claims abstract description 46
- 230000032683 aging Effects 0.000 claims abstract description 26
- 238000012549 training Methods 0.000 claims abstract description 26
- 238000013461 design Methods 0.000 claims abstract description 4
- 239000013598 vector Substances 0.000 claims description 24
- 239000000178 monomer Substances 0.000 claims description 19
- 238000009795 derivation Methods 0.000 claims description 17
- WHXSMMKQMYFTQS-UHFFFAOYSA-N Lithium Chemical compound [Li] WHXSMMKQMYFTQS-UHFFFAOYSA-N 0.000 claims description 14
- 229910052744 lithium Inorganic materials 0.000 claims description 14
- 238000004088 simulation Methods 0.000 claims description 14
- 238000004458 analytical method Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000012544 monitoring process Methods 0.000 claims description 6
- 238000009792 diffusion process Methods 0.000 claims description 5
- 238000009826 distribution Methods 0.000 claims description 4
- 230000036541 health Effects 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 230000002238 attenuated effect Effects 0.000 claims description 3
- 238000005259 measurement Methods 0.000 claims description 3
- 239000002904 solvent Substances 0.000 claims description 3
- 230000032677 cell aging Effects 0.000 claims description 2
- 230000007246 mechanism Effects 0.000 description 7
- 230000008859 change Effects 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000004146 energy storage Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000004806 packaging method and process Methods 0.000 description 3
- 241000135164 Timea Species 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 2
- 125000004122 cyclic group Chemical group 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 239000011149 active material Substances 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000010931 gold Substances 0.000 description 1
- 229910052737 gold Inorganic materials 0.000 description 1
- 239000010439 graphite Substances 0.000 description 1
- 229910002804 graphite Inorganic materials 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- GELKBWJHTRAYNV-UHFFFAOYSA-K lithium iron phosphate Chemical compound [Li+].[Fe+2].[O-]P([O-])([O-])=O GELKBWJHTRAYNV-UHFFFAOYSA-K 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 230000010287 polarization Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 239000007784 solid electrolyte Substances 0.000 description 1
- 230000035882 stress Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000008719 thickening Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4007—Scaling of whole images or parts thereof, e.g. expanding or contracting based on interpolation, e.g. bilinear interpolation
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01M—PROCESSES OR MEANS, e.g. BATTERIES, FOR THE DIRECT CONVERSION OF CHEMICAL ENERGY INTO ELECTRICAL ENERGY
- H01M10/00—Secondary cells; Manufacture thereof
- H01M10/05—Accumulators with non-aqueous electrolyte
- H01M10/052—Li-accumulators
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01M—PROCESSES OR MEANS, e.g. BATTERIES, FOR THE DIRECT CONVERSION OF CHEMICAL ENERGY INTO ELECTRICAL ENERGY
- H01M10/00—Secondary cells; Manufacture thereof
- H01M10/05—Accumulators with non-aqueous electrolyte
- H01M10/058—Construction or manufacture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/04—Ageing analysis or optimisation against ageing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/08—Thermal analysis or thermal optimisation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P70/00—Climate change mitigation technologies in the production process for final industrial or consumer products
- Y02P70/50—Manufacturing or production processes characterised by the final manufactured product
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Business, Economics & Management (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Human Resources & Organizations (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Manufacturing & Machinery (AREA)
- Chemical & Material Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Electrochemistry (AREA)
- General Chemical & Material Sciences (AREA)
- Strategic Management (AREA)
- General Health & Medical Sciences (AREA)
- Economics (AREA)
- Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Computer Hardware Design (AREA)
- Game Theory and Decision Science (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Geometry (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Secondary Cells (AREA)
Abstract
The invention provides a method for predicting the cycle life of any topological structure of a battery pack, which simulates an aging track model of typical structures of all battery packs to generate a training set for describing the topological structure, and trains a first serial-then-parallel topological structure cycle life prediction neural network and a first parallel-then-serial topological structure cycle life prediction neural network by using the training set; and (3) giving a battery pack with design requirements, and inputting all possible topological structures of the battery pack into the trained serial-parallel topological structure cycle life prediction neural network and the trained parallel-serial topological structure cycle life prediction neural network respectively so as to predict the cycle life of the topological structure battery pack. The invention can exert the integral performance of the battery pack to the maximum extent and reduce the operation cost under the unit standard circulation condition.
Description
Technical Field
The invention relates to the technical field of batteries, in particular to a method for deducing and predicting the cycle life of any topological structure of a battery pack.
Background
Under present battery manufacturing technical condition, when the battery cell series-parallel connection use in groups, because the battery parameter that battery monomer appears in links such as production, screening, use, maintenance is inconsistent for battery module after in groups often does not reach the original level that the monocell used, and this inconsistent phenomenon not only can reduce the availability factor, the life of whole group battery, if the mismanagement, can bring uncontrollable safety problem even. How to achieve the balance between the management cost and the service life of the battery module by optimizing the optimal series-parallel connection mode of the battery under the existing condition, thereby exerting the overall cycle life of the battery pack to the maximum extent on the premise of ensuring the safety of the battery, reducing the operation cost under the unit standard cycle condition, realizing the maximum benefit of the energy storage system and having important theoretical and engineering values.
In order to obtain the cycle life of the battery packs with all possible topological structures under the condition of the same battery array, all topological structure battery packs are simulated by establishing all possible topological structure battery pack simulation models, and then the calibration of the optimal topological structure is realized. However, building all possible topologies consumes a lot of manpower and is not universal.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a method for predicting the cycle life of any topological structure of a battery pack, which solves the problem that the topological structure influences the cycle life of a battery pack.
The present invention achieves the above-described object by the following technical means.
A method for predicting cycle life derivation of any topological structure of a battery pack specifically comprises the following steps:
establishing a battery pack serial-to-parallel topological structure cycle life prediction neural network and a battery pack parallel-to-serial topological structure cycle life prediction neural network, wherein the input of the neural network is as follows: the topological structure of the battery pack comprises a 1 multiplied by 4 vector consisting of serial and parallel numbers, and the output of the neural network is as follows: a cycle life index;
simulating aging track models of all battery pack typical structures to obtain cycle times of battery packs with typical structures, wherein the capacity of the battery packs with the typical structures is attenuated to 80% of the rated capacity of the battery packs, and generating a training set which takes a 1 x 4 vector describing a topological structure as an input and takes the cycle times as an output; the training set is used for training a first serial-then-parallel topological structure cycle life prediction neural network and a first parallel-then-serial topological structure cycle life prediction neural network;
and (3) giving a battery pack with design requirements, and inputting all possible topological structures of the battery pack into the trained serial-to-parallel topological structure cycle life prediction neural network and the trained parallel-to-serial topological structure cycle life prediction neural network respectively for predicting the cycle life of the topological structure battery pack.
Further, the aging trajectory model is obtained by:
the battery model parameters and the distribution condition thereof are respectively input into the aging track model of each single battery, the current and the voltage of each single battery at different moments are continuously calculated, the relationship between each model parameter and the capacity loss is utilized to update the capacity of all the single batteries and the numerical value of the model parameter in real time, the measurement of SOH is calculated by combining the monitoring of the extreme voltage in the battery pack and the continuous recursion of the loss of active lithium, the reciprocating charge-discharge simulation of the whole model is completed, and then the aging track model of the battery pack is finally established.
Still further, the cell aging trajectory model is based on analysis of the battery capacity loss and its cell model parameters, including the battery internal resistance increase and the open circuit voltage.
Further, the relationship between the battery capacity loss and the battery internal resistance increase amount is:wherein R isoIs the ohmic internal resistance, alpha, of the battery1、α2Is the model parameter, C, which needs to be determined experimentallylossIs a loss of capacity of the battery, andθ1、θ2、θ3is the model coefficient, T, that needs to be determined experimentallyk、TstdRespectively operating temperature and standard operating temperature, Δ tchg,kThe actual charging time between two monitoring.
Further, the Δ tchg,kAndthe linear relation is specifically as follows:wherein Ea,SEIR is the gas constant, the activation energy of the solvent diffusion in the SEI.
Further, the relationship between the open circuit voltage and the battery capacity loss is: and carrying out interpolation calculation on the open-circuit voltage OCV under different SOC and SOH states.
Further, the SOH is calculated by continuously recurrently calculating the active lithium loss, specifically:
wherein Q is the initial rated capacity of the battery, SOHkThe battery state of health.
Further, the typical structure of the battery pack includes a series-parallel structure, a parallel-series structure, and a series-parallel structure in which the series structure and the parallel structure are combined.
Further, the cycle life indicator is the number of cycles that the battery capacity decays to 80% of its rated capacity.
Further, the vector includes a 1 × 4 vector composed of serial and parallel numbers, specifically:
the 1 × 4 vector of the serial-first parallel sequential topology is: monomer to module series number monomer to module parallel number module to integral series number module to integral parallel number;
the 1 × 4 vector of the parallel-first-then-string sequential topology is: monomer to module parallel number monomer to module series number module to integral parallel number module to integral series number;
if the topological structure has no modules connected in series or in parallel to the whole, the value of the number of the modules connected in series to the whole or the number of the modules connected in parallel to the whole in the third and fourth bits in the vector is the default value 1.
The invention has the beneficial effects that: the method comprises the steps of configuring a BP neural network for regression prediction, selecting a typical structure of the battery pack, establishing an aging track model of the typical structure on the basis of a recyclable lithium loss mechanism, simulating the aging track model of the typical structure under the conditions of given temperature and current multiplying power to obtain the cycle life of the battery pack of the typical structure, generating a training set of the BP neural network, respectively training a first-series-then-parallel topological structure cycle life prediction neural network and a first-parallel-then-series topological structure cycle life prediction neural network through the training set, and finally realizing derivation prediction of the cycle life of any topological structure of the battery pack; the method predicts the cycle life of any topological structure of the battery pack by derivation, selects the topological structure of the battery pack with the optimal cycle life by taking the cycle life as a standard, exerts the overall performance of the battery pack to the maximum extent, reduces the operation cost under the unit standard cycle condition, and is beneficial to realizing the maximum benefit of the energy storage system.
Drawings
FIG. 1 is a flow chart of a method for predicting cycle life derivation of any topology of a battery pack according to the present invention;
FIG. 2 is a schematic diagram of a neural network according to the present invention;
FIG. 3 is a schematic diagram of a first-order R-RC equivalent circuit model according to the present invention;
fig. 4 is a diagram of a process for constructing a battery aging trajectory model according to an embodiment of the present invention;
fig. 5 is a schematic diagram illustrating comparison between the cycle count of the real topology structure and the cycle count predicted by the method and simulated by the modeling method according to the embodiment of the present invention.
Detailed Description
The invention will be further described with reference to the following figures and specific examples, but the scope of the invention is not limited thereto.
As shown in fig. 1, the method for predicting cycle life of any topology structure of a battery pack of the present invention specifically includes the following steps:
in order to analyze the prediction situation of the method of the present invention, an atypical array having a serial number of 6 and a parallel number of 2 is taken as an example. The invention discloses a method for predicting cycle life derivation of any topological structure by taking widely-applied lithium iron phosphate/graphite (LFP/GIC) batteries as experimental samples and working conditions of 1C current multiplying power/25 ℃, wherein the method for predicting cycle life derivation of any topological structure is based on simulation of cycle life of a typical topological structure calculated by continuous recursion of active lithium loss, and specifically comprises the following steps:
step (1), constructing a neural network framework for predicting the cycle life of a topological structure of a battery pack
Step (1.1), determining input and output of neural network
The battery grouping topological structure mainly comprises a series-parallel structure, a parallel-series structure and a series-parallel structure combined with the series-parallel structure and the parallel-series structure, and can be specifically classified into two sequence conditions of series-parallel and parallel-parallel and series-parallel. For this purpose, the present invention describes the topology by using a 1 × 4 vector (e.g., [ 2222 ]) consisting of the serial number and the parallel number, and uses the 1 × 4 vector as the characteristic input quantity of the neural network, and uses the cycle life index (the number of cycles for the battery capacity to decay to 80% of its rated capacity) as the characteristic output quantity of the neural network. As shown in fig. 2.
Wherein, the corresponding positions of the series-parallel numerical values in the 1 × 4 vector describing the serial-parallel sequential topological structure of the battery pack are as follows: monomer to module series number-monomer to module parallel number-module to global series number-module to global parallel number, the vector form is [ monomer to module series number monomer to module parallel number module to global series number module to global parallel number ], the corresponding position of the series-parallel numerical value in the 1 x 4 vector describing the sequential topology of battery pack parallel-to-serial order is: monomer to module parallel number-monomer to module series number-module to global parallel number-module to global series number, vector form is [ monomer to module parallel number monomer to module series number module to global parallel number module to global series number ], if the topology has no module to global series or parallel, the value of the third, fourth digit module to global series number or module to global parallel number in the vector is default value 1. By the vector description method in the form, the digital expression of the topological structure form can be realized, and the realization of the cycle life of the battery pack topological structure through neural network prediction is facilitated.
In order to avoid ambiguity of predicting the cycle life of the battery pack with a serial-first serial-then-parallel and parallel-first serial-then-serial sequential structure by the neural network, the invention respectively constructs: the method comprises the following steps of firstly connecting the topological structure in series and then connecting the topological structure in parallel, predicting the service life of the topological structure in parallel and then connecting the topological structure in series, and predicting the service life of the topological structure in parallel and then connecting the topological structure in parallel.
Step (1.2), designing and configuring a neural network
Because the BP neural network model is superior to the mapping relation for processing complex and fuzzy, and the relation between the distribution form of data and variables is not required to be known, the BP neural network simulation algorithm is adopted for training and predicting.
The prediction problem of the cycle life of the topological structure is a function fitting problem, so that a training function 'rainlm' is selected for neural network training; wherein the network target error 'net.trainparam.gold' is 1e-3, the learning speed 'net.trainparam.lr' is 0.03, and the maximum training step number 'net.trainparam.epochs' is 1000; and setting a BP neural network according to the parameters, and constructing a neural network framework for predicting the cycle life of the topological structure of the battery pack.
Step (2), generating a training set of the neural network
Step (2.1), selecting a typical structure of the battery pack as a training sample
Selecting a typical structure which needs to contain all types of structures (a serial-first structure and a parallel-second structure, a parallel-first structure and a serial-parallel-second structure and a mixed structure combining the serial structure and the parallel-second structure); and because the sample input of the neural network is uniformly distributed as much as possible, the probability of poor prediction capability caused by unobvious training traces in areas with fewer samples due to excessive training at the sample dense position is reduced. Based on the principle, different topological structures with the total battery number of 16, 32, 64 and 128 of the battery pack are selected as typical simulation structures, and the typical simulation structures can be specifically divided into a serial-first parallel-then-parallel typical structure and a parallel-first serial-then-serial typical structure. Wherein, the serial-to-parallel typical structure comprises: 4 strings 4 in parallel, 2 strings 2 in parallel then 2 strings 2 in parallel, 2 strings 4 in parallel then 2 strings, 16 strings 2 in parallel, 4 strings 2 in parallel then 4 strings, 2 strings 2 in parallel then 4 strings 2 in parallel, 32 strings 2 in parallel, 4 strings 2 in parallel then 8 strings, 2 strings 16 in parallel then 2 strings, 4 strings 2 in parallel then 4 strings 2 in parallel, 2 strings 4 in parallel then 4 strings 2 in parallel, 2 strings 64 in parallel, 4 strings 4 in parallel then 8 strings, 4 strings 4 in parallel then 2 strings 4 in parallel, 8 strings 16 in parallel, 2 strings 8 in parallel then 2 strings 4 in parallel; the parallel-first-string typical structure comprises: 4 parallel 4 strings, first 2 parallel 2 strings then 2 parallel 2 strings, first 2 parallel 4 strings then 2 parallel, 16 parallel 2 strings, first 4 parallel 2 strings then 4 parallel, first 2 parallel 2 strings then 4 parallel 2 strings, 32 parallel 2 strings, first 4 parallel 2 strings then 8 parallel, first 2 parallel 16 strings then 2 parallel, first 4 parallel 2 strings then 4 parallel 2 strings, first 2 parallel 4 strings then 4 parallel 2 strings, 2 parallel 64 strings, first 4 parallel 4 strings then 8 parallel, first 4 parallel 4 strings then 2 parallel 4 strings, 8 parallel 16 strings, first 2 parallel 8 strings then 2 parallel 4 strings.
Step (2.2), a typical battery pack aging track model is established by utilizing simulation software
In order to accurately represent the cycle life of a typical battery pack, the aging trajectory model of the battery pack with the typical structure is constructed by utilizing simulation software based on the construction of a single battery aging mechanism and in consideration of the inconsistent influence of parameters in the battery pack.
In the aspect of single batteries, according to the analysis of time constants corresponding to each link in the battery, a first-order R-RC equivalent circuit model is selected as a modeling reference, and the specific model form and the battery terminal voltage mathematical expression are respectively shown as a figure 3 and a formula (1). In the figure, UbatIs the terminal voltage (V), I of the batterybatIs the total current (A) flowing through the battery, OCV is the open circuit voltage of the battery, Ro、RdOhmic and diffused internal resistance (omega), C of the cell, respectivelydIs a diffusion capacitance, ηo、ηdOhmic overpotential and polarization overpotential (V), I respectivelydThe diffusion current (A) of the battery.
On the basis of the selected equivalent circuit model, the invention respectively establishes the mechanism relation between each model parameter and the capacity loss quantity of the equivalent circuit model on the basis of realizing the calculation of the capacity loss of the battery.
During the normal use stage of the battery, the capacity loss is mainly caused by the consumption of the recyclable lithium in the battery caused by the generation and thickening of the solid electrolyte film (SEI film) on the surface of the negative electrode. For this reason, the invention is based on a classical active lithium loss model (as shown in formula (2)) starting from the conventional aging mechanism of the cell. LLIaThe resulting loss of battery capacity is mainly caused by the generation and growth of SEI film, and its mathematical expression is about the charging time t of batterychgAs a function of (c). If the battery is placed under a certain standard stress condition (i.e. the working temperature T is equal to the standard working temperature T)std) (ii) loss of battery capacity LLIa,surfAnd standard charging timeAs a standard curve, the change of T can be considered to be relative to the standard curveWhich acts as an acceleration (or deceleration). Therefore, the active lithium capacity loss LLI under any working temperature T condition can be reducedaRewriting to standard charging timeThe expression of acceleration (or deceleration) of (1) is as shown in equation (3):
in the above formula, ClossAs capacity loss (Ah) of the battery, LLIaLLI is the loss of active lithium capacity (Ah) of a battery during aging0The initial active lithium capacity loss (Ah) of the battery after the factory formation is obtained; t is tchg、Respectively the charging time and the equivalent charging time(s) under the standard condition; t, TstdRespectively, an operating temperature and a standard operating temperature (K), where Tstd=303K;kSEIGenerating a reaction correlation coefficient, E, for the SEI filma,SEIThe activation energy (J/mol) of the solvent diffusing in the SEI, and R is the gas constant (8.314J/(mol. K)).
Further, the battery charging time t under any condition can be setchgConverted to it at the standard temperature T ═ TstdStandard charging timeAs shown in formula (4):
due to the fact thatAnd tchgIn a linear relationship, can be pairedDiscretizing treatment is carried out, as shown in a formula (4),and Δ tchg,kRespectively representing the standard charging time and the actual charging time between two monitoring; meanwhile, the definition of each physical quantity is not changed, and only the discretization form of each physical quantity at the k-th time is rewritten, and k is 1,2, ….
By pairsAccumulating to obtain the accumulated total amount of the equivalent charging time corresponding to the dynamic working condition; finally, a recursive expression of the battery capacity loss and the corresponding equivalent charging time under the time-varying working condition is obtained, as shown in formula (6):
the formula (6) is sorted to obtain a lumped parameter model for capacity loss recursion calculation, wherein theta is shown as a formula (7)1~θ3Model coefficients determined experimentally.
SOH was characterized using a capacity loss recursive calculation as:
wherein Q is the initial rated capacity of the battery, SOHkThe state of health of the battery under the time-varying working condition.
In addition, based on theoretical analysis of the action relationship between the recyclable Lithium Loss (LLI) and the ohmic internal resistance increment of the battery, a functional relationship between the internal resistance increment of the battery and the capacity loss of the battery is obtained, as shown in formula (9)In the formula (9), α1、α2The model parameters that need to be determined experimentally.
Through the experiment of the battery under different temperatures and different aging degrees, the parameter identification is carried out through the least square method, and the parameter theta is determined1Is 8.43 x 106、θ2Is 7.02 x 103、θ3Is 0.59, alpha1Is 4.15 x 10-3、α2Is 2.62 x 10-4。
Meanwhile, the loss of active lithium (LLI) causes the capacity of the battery to be reduced, and the open circuit voltage curve OCV of the battery is also affected correspondingly. Based on analysis of the change rule of the battery under the electrode potential combined coordinate system, the battery OCV curve at the LLI stage presents a definite change rule: as the battery ages, the overall correspondence of the OCV curve of the battery to its state of charge (SOC) does not change, but rather only manifests itself as a gradual disappearance of the high SOC voltage plateau. For this reason, the open-circuit voltages (OCV) of the batteries in all SOC and SOH battery states can be obtained by interpolating the open-circuit voltages (OCV) in different states of charge (SOC) and different states of health (SOH), and the three-dimensional mechanism curve of the OCV-SOC-SOH interpolation result is shown as the partial OCV parameters in fig. 4. The SOC estimation method is an ampere-hour integration method.
In addition, for the RC element in the monomer model, since the value thereof mainly depends on the diffusion rate and particle radius of the battery material itself, and the two are mainly affected by the loss of the active material, the change value with the loss of the battery capacity can be ignored in the conventional aging stage of the battery.
Based on the battery capacity loss and the mechanism analysis and data representation of the monomer model parameters, a single battery aging track model is constructed by using simulation software. On the basis of a single battery aging track model, the embodiment of the invention takes the consideration of the factor of inconsistent model parameters in the battery pack into consideration, and builds upThe aging track model of the battery pack is shown in a specific form in fig. 4. It uses battery model parameter (R)o、Rd、Cd、θ1、θ2、θ3、α1、α2) And the distribution condition is input, the current and voltage conditions of each single at different moments are continuously calculated under the integral framework of the battery pack, the real-time updating of the capacities and parameter values of all the single batteries is realized by utilizing the established mechanism relation between each model parameter and the capacity loss amount on the basis, the measurement of SOH is further calculated by combining the monitoring of extreme value voltage in the battery pack (in the prior art) and the continuous recursion of active lithium loss, the reciprocating charge-discharge simulation of the integral model is completed, and the aging track model of the battery pack is finally established.
And respectively modeling the battery packs with the typical structures through the modeling mode to obtain aging track models of the typical structures of all the battery packs.
Step (2.3), generating a training set based on the aging track model of the typical structure
Under the working conditions of given current multiplying power and temperature, simulating aging track models of all battery pack typical structures to obtain the cycle number of the battery pack with the typical structure, wherein the capacity of the battery pack is attenuated to 80% of the rated capacity of the battery pack with the typical structure, and generating a training set which describes a topological structure and takes a 1 multiplied by 4 vector as input and the cycle number as output.
Step (3), training the neural network through the training set, and further realizing the prediction of the cycle life of any topological structure through the trained neural network
And respectively training a first serial-parallel topological structure cycle life prediction neural network and a first parallel-serial topological structure cycle life prediction neural network based on the obtained training set.
Under the conditions of battery pack rated voltage, rated capacity and battery array required by given design, all possible topological structures are listed through a permutation and combination method, the possible topological structures are described through vectors of 1 x 4 respectively, and the described topological structure vectors are input into a trained first-series-then-parallel topological structure cycle life prediction neural network and a first-parallel-then-series topological structure cycle life prediction neural network respectively and are used for predicting the cycle life of the topological structure battery pack respectively. Because the training samples are not many, the single regression prediction error is large, and in order to reduce the error of the regression prediction, the median of 100 prediction results is taken as the final prediction result; in the embodiment of the invention, the predicted values of the cycle times of the battery pack with the 2-parallel 6-string and 6-string 2-parallel topological structure are obtained.
In order to discuss the accuracy of the battery pack aging trajectory model modeling method provided in the step (2), the embodiment of the invention constructs 2-parallel 6-string and 6-string 2-parallel topological structure battery pack models by the battery pack aging trajectory model modeling method provided in the step (2), and carries out simulation according to the conditions and parameters of the embodiment of the invention to obtain the simulation value of the target topological cycle number.
The cyclic aging life test of the atypical topological structure (2 in parallel and 6 in series and 2 in parallel) is carried out under the working condition of the embodiment of the invention, and the measured value of the cyclic aging cycle number of the atypical topological structure under the working condition is obtained.
The predicted value, the simulation value and the measured value are compared, and the comparison result is shown in fig. 5, so that the relative errors of the prediction method and the modeling simulation method are controlled within 5%, and the effectiveness of the cycle life derivation prediction method for any topological structure in predicting the cycle life of any topological structure is verified. Through the processes of typical structure simulation, neural network prediction and the like, the method can accurately predict the cycle life of any topological structure, and is beneficial to realizing the maximum benefit of the energy storage system.
In summary, the method provided by the embodiment of the invention uses the lumped parameter model of the capacity loss recursion calculation to construct the simulation model of the cycle life of the typical battery pack, uses the simulation result to train the neural network, reduces the workload of constructing the topology model, predicts the cycle life of all possible topology structures under the target battery array based on the trained network, and further realizes the calibration of the optimal topology structure. Compared with the prior art, the method provided by the invention has the advantages of high efficiency, high prediction precision and strong universality of the topology cycle life derivation method.
The present invention is not limited to the above-described embodiments, and any obvious improvements, substitutions or modifications can be made by those skilled in the art without departing from the spirit of the present invention.
Claims (10)
1. A method for predicting cycle life derivation of any topological structure of a battery pack is characterized by comprising the following steps:
establishing a battery pack serial-to-parallel topological structure cycle life prediction neural network and a battery pack parallel-to-serial topological structure cycle life prediction neural network, wherein the input of the neural network is as follows: the topological structure of the battery pack comprises a 1 multiplied by 4 vector consisting of serial and parallel numbers, and the output of the neural network is as follows: a cycle life index;
simulating aging track models of all battery pack typical structures to obtain cycle times of battery packs with typical structures, wherein the capacity of the battery packs with the typical structures is attenuated to 80% of the rated capacity of the battery packs, and generating a training set which takes a 1 x 4 vector describing a topological structure as an input and takes the cycle times as an output; the training set is used for training a first serial-then-parallel topological structure cycle life prediction neural network and a first parallel-then-serial topological structure cycle life prediction neural network;
and (3) giving a battery pack with design requirements, and inputting all possible topological structures of the battery pack into the trained serial-to-parallel topological structure cycle life prediction neural network and the trained parallel-to-serial topological structure cycle life prediction neural network respectively for predicting the cycle life of the topological structure battery pack.
2. The battery pack arbitrary topology cycle life derivation prediction method according to claim 1, wherein the aging trajectory model is obtained by:
the battery model parameters and the distribution condition thereof are respectively input into the aging track model of each single battery, the current and the voltage of each single battery at different moments are continuously calculated, the relationship between each model parameter and the capacity loss is utilized to update the capacity of all the single batteries and the numerical value of the model parameter in real time, the measurement of SOH is calculated by combining the monitoring of the extreme voltage in the battery pack and the continuous recursion of the loss of active lithium, the reciprocating charge-discharge simulation of the whole model is completed, and then the aging track model of the battery pack is finally established.
3. The battery arbitrary topology cycle life derivation prediction method of claim 2, wherein the cell aging trajectory model is based on analysis of cell capacity loss and its cell model parameters, the cell model parameters comprising cell internal resistance increase and open circuit voltage.
4. The battery pack arbitrary topology cycle life derivation prediction method according to claim 3, wherein the relationship between the battery capacity loss and the battery internal resistance increase is as follows:wherein R isoIs the ohmic internal resistance, alpha, of the battery1、α2Is the model parameter, C, which needs to be determined experimentallylossIs a loss of capacity of the battery, andθ1、θ2、θ3is the model coefficient, T, that needs to be determined experimentallyk、TstdRespectively operating temperature and standard operating temperature, Δ tchg,kThe actual charging time between two monitoring.
6. The battery pack arbitrary topology cycle life derivation prediction method of claim 3, wherein the relationship between the open circuit voltage and the battery capacity loss is: and carrying out interpolation calculation on the open-circuit voltage OCV under different SOC and SOH states.
8. The method for prediction of cycle life derivation for any topology of battery packs according to claim 1, wherein the typical configuration of the battery pack comprises a series-first-parallel configuration, a parallel-first-series-second-series configuration, and a hybrid configuration of the two configurations.
9. The battery arbitrary topology cycle life derivation prediction method of claim 1, wherein the cycle life indicator is a number of cycles that the battery capacity decays to 80% of its rated capacity.
10. The method for estimating and predicting the cycle life of the battery pack in any topology structure according to claim 1, wherein the method comprises a 1 x 4 vector consisting of serial and parallel numbers, and specifically comprises the following steps:
the 1 × 4 vector of the serial-first parallel sequential topology is: monomer to module series number monomer to module parallel number module to integral series number module to integral parallel number;
the 1 × 4 vectors of the parallel-first and serial-second sequential topology are: monomer to module parallel number monomer to module series number module to integral parallel number module to integral series number;
if the topological structure has no modules connected in series or in parallel to the whole, the value of the number of the modules connected in series to the whole or the number of the modules connected in parallel to the whole in the third and fourth bits in the vector is the default value 1.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210054745.0A CN114492182A (en) | 2022-01-18 | 2022-01-18 | Method for predicting cycle life of battery pack in any topological structure |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210054745.0A CN114492182A (en) | 2022-01-18 | 2022-01-18 | Method for predicting cycle life of battery pack in any topological structure |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114492182A true CN114492182A (en) | 2022-05-13 |
Family
ID=81512654
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210054745.0A Pending CN114492182A (en) | 2022-01-18 | 2022-01-18 | Method for predicting cycle life of battery pack in any topological structure |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114492182A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117151201A (en) * | 2023-08-24 | 2023-12-01 | 广芯微电子(广州)股份有限公司 | Battery pack sample organization method and device for neural network training |
-
2022
- 2022-01-18 CN CN202210054745.0A patent/CN114492182A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117151201A (en) * | 2023-08-24 | 2023-12-01 | 广芯微电子(广州)股份有限公司 | Battery pack sample organization method and device for neural network training |
CN117151201B (en) * | 2023-08-24 | 2024-03-15 | 广芯微电子(广州)股份有限公司 | Battery pack sample organization method and device for neural network training |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Zhao et al. | A compact methodology via a recurrent neural network for accurate equivalent circuit type modeling of lithium-ion batteries | |
KR20200140093A (en) | Prediction Method and Prediction System for predicting Capacity Change according to Charging / Discharging Cycle of Battery | |
CN111007401A (en) | Electric vehicle battery fault diagnosis method and device based on artificial intelligence | |
CN108519556A (en) | A kind of lithium ion battery SOC prediction techniques based on Recognition with Recurrent Neural Network | |
CN106055775B (en) | A kind of service life of secondary cell prediction technique that particle filter is combined with mechanism model | |
CN110045292A (en) | Lithium ion battery SOC prediction technique based on big data and bp neural network | |
CN109001640B (en) | Data processing method and device for power battery | |
CN111426957B (en) | SOC estimation optimization method for power battery under simulated vehicle working condition | |
Zhang et al. | Intelligent state of charge estimation of battery pack based on particle swarm optimization algorithm improved radical basis function neural network | |
CN110568359A (en) | lithium battery residual life prediction method | |
CN112711892A (en) | Cloud battery management system and method based on digital twin and block chain technology | |
CN113064093A (en) | Energy storage battery state of charge and state of health joint estimation method and system | |
KR102090476B1 (en) | a BMS optimizing system using a cloud system and big data | |
CN108009585B (en) | Lead-acid battery health state prediction method based on local information fusion | |
CN108804800A (en) | Lithium ion battery SOC on-line prediction methods based on echo state network | |
CN110082682A (en) | A kind of lithium battery charge state estimation method | |
CN113125960A (en) | Vehicle-mounted lithium ion battery charge state prediction method based on random forest model | |
CN111707956A (en) | Method for predicting health state and residual life of multi-type lithium ion battery pack management system | |
CN114781176B (en) | Equivalent circuit parameter identification method for lumped parameters of lithium ion battery energy storage system | |
CN114636948A (en) | Energy storage system service life assessment method and device, electronic equipment and storage medium | |
CN114492182A (en) | Method for predicting cycle life of battery pack in any topological structure | |
Geng et al. | SOC Prediction of power lithium battery using BP neural network theory based on keras | |
CN115114843A (en) | Battery capacity prediction method, related device, equipment and storage medium | |
Zhang et al. | Lithium battery SOC prediction based on mproved BP eural etwork algorithm | |
CN113495214A (en) | Super capacitor charge state estimation method based on temperature change model |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |