CN115656855A - Lithium ion battery health assessment method and system - Google Patents
Lithium ion battery health assessment method and system Download PDFInfo
- Publication number
- CN115656855A CN115656855A CN202210253044.XA CN202210253044A CN115656855A CN 115656855 A CN115656855 A CN 115656855A CN 202210253044 A CN202210253044 A CN 202210253044A CN 115656855 A CN115656855 A CN 115656855A
- Authority
- CN
- China
- Prior art keywords
- battery
- health
- discharge
- lithium ion
- ion battery
- 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
- 230000036541 health Effects 0.000 title claims abstract description 73
- HBBGRARXTFLTSG-UHFFFAOYSA-N Lithium ion Chemical compound [Li+] HBBGRARXTFLTSG-UHFFFAOYSA-N 0.000 title claims abstract description 35
- 229910001416 lithium ion Inorganic materials 0.000 title claims abstract description 35
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000004364 calculation method Methods 0.000 claims abstract description 36
- 230000006870 function Effects 0.000 claims description 31
- 238000005457 optimization Methods 0.000 claims description 21
- 239000002245 particle Substances 0.000 claims description 21
- 238000012544 monitoring process Methods 0.000 claims description 14
- 230000002787 reinforcement Effects 0.000 claims description 14
- 238000007599 discharging Methods 0.000 claims description 12
- 238000010606 normalization Methods 0.000 claims description 10
- 239000011159 matrix material Substances 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 7
- 230000003862 health status Effects 0.000 claims description 5
- 238000004891 communication Methods 0.000 claims description 4
- 230000009466 transformation Effects 0.000 claims description 4
- 230000002708 enhancing effect Effects 0.000 claims description 2
- 238000003487 electrochemical reaction Methods 0.000 abstract description 3
- 238000011156 evaluation Methods 0.000 description 18
- 238000012549 training Methods 0.000 description 13
- 230000008569 process Effects 0.000 description 7
- 238000012360 testing method Methods 0.000 description 7
- WHXSMMKQMYFTQS-UHFFFAOYSA-N Lithium Chemical compound [Li] WHXSMMKQMYFTQS-UHFFFAOYSA-N 0.000 description 6
- 229910052744 lithium Inorganic materials 0.000 description 6
- 239000002253 acid Substances 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- PWHULOQIROXLJO-UHFFFAOYSA-N Manganese Chemical compound [Mn] PWHULOQIROXLJO-UHFFFAOYSA-N 0.000 description 1
- YZCKVEUIGOORGS-IGMARMGPSA-N Protium Chemical compound [1H] YZCKVEUIGOORGS-IGMARMGPSA-N 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 229910017052 cobalt Inorganic materials 0.000 description 1
- 239000010941 cobalt Substances 0.000 description 1
- GUTLYIVDDKVIGB-UHFFFAOYSA-N cobalt atom Chemical compound [Co] GUTLYIVDDKVIGB-UHFFFAOYSA-N 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000004146 energy storage Methods 0.000 description 1
- 238000013210 evaluation model Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 150000002500 ions Chemical class 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
- 229910052748 manganese Inorganic materials 0.000 description 1
- 239000011572 manganese Substances 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 229910052987 metal hydride Inorganic materials 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
Images
Landscapes
- Secondary Cells (AREA)
Abstract
The invention provides a lithium ion battery health assessment method and a system, wherein a battery capacity value in an initial state of a battery is obtained, the battery is discharged for multiple cycles, and the voltage of the battery is reduced to a rated value in each discharge cycle; in each discharge cycle, collecting the discharge voltage and the discharge temperature of the battery at regular time as sample data; and establishing a calculation model, and inputting sample data to obtain an estimated value of the state of health of the battery. The invention provides a method for estimating the state of health of a battery, which is based on the voltage and the temperature of battery discharge as sample data, does not need to consider the internal complex electrochemical reaction of the battery during charge and discharge, does not need to construct an equivalent physical model of the battery, and can quickly estimate the current state of health of the battery.
Description
Technical Field
The invention belongs to the technical field of lithium ion battery health management, and particularly relates to a lithium ion battery health assessment method and system.
Background
Lithium ion batteries have become mainstream energy storage devices due to their advantages of high energy density, good cycle performance, high safety, and the like. The state of health (SOH) of a lithium ion battery reflects the current performance of the battery, and is usually characterized by the current capacity of the battery. The estimation of the state of health of the lithium ion battery is also an important function of a battery management system, and the accuracy of the estimation directly influences the safety and reliability of equipment.
At present, due to a plurality of factors influencing the SOH of the lithium ion battery, the electrochemical parameters of the battery are difficult to measure, and an accurate battery physical model is difficult to construct, and a data-driven method can directly acquire health state information from experiments and monitoring data mining. Compared with other algorithms, the support vector regression model (SVR for short) has high convergence speed and excellent generalization capability, and can effectively estimate the SOH. However, the performance of SVR is highly dependent on model parameters, which are usually determined by manual selection or repeated experiments when building a model, and such a process results in non-ideal parameter selection and affects the estimation result of the model. And when the SVR establishes the parameter solving model based on the structure risk minimization principle, the parameter constraint condition is inequality constraint, and the parameter solving speed is slow.
Therefore, there is a need to provide an improved solution to the above-mentioned deficiencies of the prior art.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and obtain a method for accurately estimating the health state of a lithium battery.
In order to achieve the above purpose, the invention provides the following technical scheme:
a method for evaluating the health of a lithium ion battery,
acquiring a battery capacity value in an initial state of the battery, discharging the battery for multiple cycles, and reducing the voltage of the battery to a rated value in each discharge cycle;
in each discharge cycle, collecting the discharge voltage and the discharge temperature of the battery at regular time as sample data;
and establishing a calculation model, and inputting sample data to obtain an estimated value of the state of health of the battery.
Preferably, the evaluation method further comprises:
performing parameter optimization on the calculation model to obtain a calculation model with an optimal parameter set, and enhancing the calculation model by utilizing reinforcement learning;
and inputting sample data into the calculation model after reinforcement learning to obtain a final value of the health state of the battery.
Preferably, the battery capacity is used to describe the final value of the battery state of health, and the battery state of health after each discharge cycle is output as a tag value.
Preferably, the state of health of the battery is obtained from a ratio of a battery capacity after any discharge to an initial battery capacity.
Preferably, an optimal parameter set is obtained through a particle swarm optimization algorithm so as to optimize parameters of the calculation model.
Preferably, the data normalization processing is performed on the sample data input into the calculation model.
Preferably, a least square support vector regression function is established as a calculation model according to the structure risk minimization principle, lagrange multipliers are introduced, and variables are eliminated by adopting an equation transformation mode to establish an optimal solution solving matrix.
A lithium ion battery health evaluation system evaluates the health state of a battery through the evaluation method.
Preferably, the evaluation system comprises:
the monitoring module is used for collecting the discharge voltage and the discharge temperature of the battery at regular time in each discharge cycle;
and the control center is in communication connection with the monitoring module, evaluates the health state of the battery according to the data acquired by the monitoring module and outputs a result.
Preferably, the evaluation system further comprises:
the charging module is used for charging the battery after each cycle of discharge is finished;
and the discharging module discharges the battery according to constant current and discharges the battery circularly until the voltage of the battery is reduced to a rated value.
Has the beneficial effects that: the invention provides a method for estimating the state of health of a battery, which is based on the voltage and the temperature of battery discharge as sample data, does not need to consider the internal complex electrochemical reaction of the battery during charge and discharge, does not need to construct an equivalent physical model of the battery, and can quickly estimate the current state of health of the battery.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate embodiments of the invention and, together with the description, serve to explain the invention and not to limit the invention. Wherein:
FIG. 1 is a block flow diagram of a battery health assessment method of the present invention;
FIG. 2 is a graph of state of health SOH of a lithium ion battery as a function of cycle number for an embodiment of the present invention;
FIG. 3 is a diagram comparing the SOH value of the lithium ion battery with the actual SOH value in the embodiment of the method of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely below, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived from the embodiments of the present invention by a person skilled in the art, are within the scope of the present invention.
In the description of the present invention, the terms "longitudinal", "lateral", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, which are for convenience of description of the present invention only and do not require that the present invention must be constructed and operated in a specific orientation, and thus, should not be construed as limiting the present invention. The terms "connected" and "connected" used herein should be interpreted broadly, and may include, for example, a fixed connection or a detachable connection; they may be directly connected or indirectly connected through intermediate members, and specific meanings of the above terms will be understood by those skilled in the art as appropriate.
The present invention will be described in detail below with reference to the embodiments with reference to the attached drawings. It should be noted that the embodiments and features of the embodiments of the present invention may be combined with each other without conflict.
As shown in fig. 1 to 3, a method for evaluating health of a lithium ion battery specifically includes: before a test is carried out, a battery capacity value in an initial state of a battery needs to be obtained, then the battery is discharged for multiple cycles, and the voltage of the battery is reduced to a rated value in each discharge cycle; in each discharge cycle, collecting the discharge voltage and the discharge temperature of the battery at regular time, and taking the discharge voltage and the discharge temperature as sample data; and establishing a calculation model for data calculation, inputting sample data into the calculation model, and obtaining an estimated value of the battery health state through calculation. The battery health state can be effectively estimated through the average discharge voltage and the average discharge temperature of the battery discharge cycle obtained through measurement, the cost is low, the precision is high, the estimation speed is high, and therefore the real-time monitoring of the battery health state is achieved. Certainly, the method can be used for health evaluation of lithium batteries, and health status evaluation of nickel-metal hydride batteries, cobalt acid lithium batteries, ternary lithium batteries, iron phosphate lithium batteries, manganese acid lithium batteries and the like can be performed.
Taking 18650 type lithium ion battery as an example, the initial capacity of the battery is set to c 0 And acquiring data of a charging mode, a discharging mode and an impedance test through the test, wherein the discharging mode discharges at a constant current of 2A until the voltage of the battery is reduced to a rated value. During the discharge cycle, the discharge voltage u (t) and the discharge temperature r (t) of the battery are sampled every 2-3s, with the unit being V and A respectively, until when the battery capacity is 30% lower than the initial capacity, eachCollecting R times in discharge cycle, performing T discharge cycles, and recording the battery capacity c after each discharge cycle i (t) in Ah.
In another alternative embodiment, the battery capacity is used to describe the final value of the battery state of health, and the battery state of health after each discharge cycle is output as a tag value. The state of health of the battery is obtained by the ratio of the discharged battery capacity to the initial capacity of the battery. Specifically, the battery state of health is a ratio of a battery capacity after the i-th discharge cycle ends to an initial battery capacity.
Wherein s (t) is a label value of the battery state of health, C i Capacity of the cell for the i-th discharge cycle, C 0 Is the initial capacity of the battery.
In another optional embodiment, to improve the accuracy of the acquired state of health of the battery, the evaluation method further comprises: performing parameter optimization on the calculation model to obtain a calculation model with an optimal parameter set, and performing calculation model reinforcement learning by utilizing reinforcement learning; and inputting sample data into the calculation model after reinforcement learning to obtain a final value of the health state of the battery. The calculation model can be a least square support vector regression model, and the technical model is subjected to reinforcement learning by adopting an Adaboost.R2 reinforcement learning method, so that the obtained calculation model after reinforcement learning has a lower calculation error, the collected sample data is input into the technical model again to obtain a more accurate battery health state, and the estimation precision is improved.
In another optional embodiment, because the collected sample data has different dimensions of the discharge voltage and the discharge temperature, and the evaluation accuracy is affected by directly inputting the calculation model, before the sample data is input into the calculation model, data normalization processing is performed on the collected discharge voltage and the discharge temperature.
In another alternative embodiment, the processing of the sample data is performed by:
before using the sample data, the characteristic value u of the sample is firstly checked 1 (t) and r 1 (t) carrying out normalization treatment, wherein the formula is as follows:
wherein y is an unprocessed sample characteristic value, y max Is the maximum value, y, of the sample feature values min Is the minimum value in the sample characteristic values, and y' is the normalized sample characteristic value. Will u 1 (t)、r 1 (t) is represented by u after normalization treatment by the formula (2) 2 (t)、r 2 (t) and S (t) are expressed as a sample set S = { [ u ]) 2 (t),r 2 (t),s(t)]I T =1,2, \ 8230;, T }, where [ u [, u [ ] 2 (t),r 2 (t),s(t)]Is a sample vector.
In another optional embodiment, based on known data, specifically discharge voltage, discharge temperature and battery capacity data acquired in the discharge process, a least squares support vector regression function is established as a calculation model according to the structural risk minimization principle, lagrangian multipliers are introduced, and variables are eliminated by adopting an equation transformation mode to establish an optimal solution solving matrix. And a least square support vector regression function is used as a model, so that the accuracy is ensured, the solving efficiency is effectively improved, and the battery health state is estimated.
In another optional embodiment, the calculation model is a least squares support vector regression model, LSSVR for short, and the specific steps are as follows:
establishing a regression model function according to the structure risk minimization principle, wherein the regression model function is defined as follows:
f(x)＝w·φ(x)+b (3)
wherein x is an input vector, w is a weight vector, b is a deviation, Φ (x) is a nonlinear function mapping x from a low-dimensional space to a high-dimensional space, and f (x) is a theoretical regression equation;
to find w and b, a minimization function is established to calculate:
where C is a penalty coefficient, ξ i Is an error factor;
taking the formula (4) as an objective function, and introducing a Lagrange multiplier eta i ，
α i ，
Constructing a Lagrangian function to convert the target function into an unconstrained form:
wherein alpha is i ≥0。
Optimization function pair w, b, xi i ，α i The minimum value of (c) is obtained by partial derivative:
the variables w and xi are eliminated by equation (6 a) -equation (6 d) in an equation transformation mode i Establishing the following optimal solution solving matrix to obtain b and alpha i ：
Wherein α = [ α = 1 ；α 2 ；L；α N ]；e 1×N ＝[1；1；L；1](ii) a Kernel function matrix
I N Is an N-dimensional identity matrix, Y = [ Y = [) 1 ；y 2 ；L；y N ]。K(x i ,x j ) Is a kernel function, x i And x j Is the input vector of the kernel function. The gaussian Radial Basis Function (RBF) is typically chosen according to Mercer conditions, defined as follows:
wherein gamma is a kernel function parameter and represents the radius of the RBF kernel.
Firstly, training set S R ＝{[u 2 (t),r 2 (t),s(t)]|t＝1,2,…,T SR Input characteristic signal x = { [ u ]) in 2 (t),r 2 (t)]|t＝1,2,…,T SR And inputting the data into the formula (8) to obtain a kernel function matrix omega. Training set S R The real value of the lithium ion battery state of health (SOH) corresponding to the middle input characteristic signal x is Y = { s (T) | T =1,2, \8230;, T SR }. Substituting omega, C and Y into the formula (7) to obtain values of alpha and b, and obtaining a final regression function model:
normalizing the sample data to obtain the sample data with x (t) = [ u = 2 (t),r 2 (t)]Obtaining an estimate of the state of health of the lithium ion battery as input to equation (9)
Firstly, the methodAnd obtaining an estimated value of the battery health state. The model is trained through the estimated value, the regression function is fitted, and therefore when test data are input again, the final value of the battery health state can be accurately obtained, and the battery health state is monitored.
In another optional embodiment, the particle swarm algorithm is adopted to obtain an optimal parameter set so as to perform parameter optimization on the calculation model, and the establishment of the calculation model is completed after the optimal parameter set is obtained. The model adopts the particle swarm optimization to optimize parameters, so that the problem that the estimation precision is influenced by the fact that the model parameters are determined by means of manual experience in the past is solved, and the estimation precision is further improved; the estimation algorithm can also be generalized to other types of battery state of health estimation problems.
In another optional embodiment, the parameter optimization is performed on the estimation model by adopting a particle swarm optimization according to the mean square error, and the specific steps are as follows:
determining a parameter set P = { C, gamma } to be optimized in a least square support vector regression estimation model, wherein C is a penalty factor and gamma is a kernel function parameter;
and taking the mean square error between the minimum estimated value and the actual value as an optimization objective function:
s.t.0<C<1000 (10b)
0<γ<8 (10c)
equations (10 b) and (10 c) represent the constrained range of the parameters that need to be optimized;
obtaining an optimal parameter set P by adopting a particle swarm algorithm, and updating the speed and the position of the particle through the following formula:
wherein,
representing the velocity vector of the particle i at the kth iteration;
representing the particle position of the particle i at the kth iteration;
representing the optimal solution position of the particle individual i when the particle individual i iterates to the k time; gbest k Representing the position of the optimal solution when the whole population iterates to the kth time; w is an inertial weight coefficient; c. C 1 And c 2 Is a learning factor; r is a radical of hydrogen 1 And r 2 For acceleration factors, in [0,1 ]]Generating randomly;
and after the optimal parameter set P is obtained, determining a least square support vector regression model after particle swarm optimization as a calculation model.
In another optional embodiment, a least square support vector model optimized by a particle swarm optimization is used as a weak learner, the weak learners are aggregated into a strong learner by adopting an Adaboost.R2 reinforcement learning method, and the method comprises the following specific steps:
and determining least square support vector regression optimized by the particle swarm optimization as a weak learner, wherein the maximum iteration number is T'.
The number of initial iterations t =1, and the weight distribution D of each data in the training set t (i) =1/N, wherein i =1,2, \8230;, N. Wherein N is the number of training set samples;
according to weight D t Training weak learner on training set h t (X) and calculating a regression error of the training samples:
wherein L is i For loss of a single training sample, squareThe form:
wherein,
calculating a weight coefficient beta for a weak learner t ＝L t /(1-L t )。
Then, the weight distribution of the sample set is updated:
wherein Z is t For the normalization factor, the expression is:
after the weight is updated, the update iteration number t = t +1.
Repeating the equations (13) - (15) until the average loss coefficient L t Less than 0.5 or the iteration number T is less than or equal to T', and the final strong learner is obtained by:
the more accurate final value of the health state of the lithium ion battery can be output by repeating the formulas (13) to (16) again after the sample data is subjected to the addition, the average and the normalization processing
In another optional embodiment, a lithium ion battery health evaluation system is provided, in which the health status of the battery is evaluated by the evaluation method described in the above embodiments. The evaluation system carries out battery health state evaluation based on a particle swarm optimization least square support vector regression method, firstly, normalization processing is carried out on data by measuring voltage and temperature data in the battery discharging process, a least square support vector regression evaluation model is solved based on the normalized data, then, an optimization objective function is constructed based on the mean square error of the evaluation result and the actual value of the existing regression model, an optimal parameter set can be obtained by adopting a particle swarm optimization, the optimized model is determined, finally, the least square support vector regression model optimized by the particle swarm is subjected to reinforcement learning, sample data is input again, and quick and accurate evaluation of the battery health state is realized.
In some embodiments, the evaluation system comprises: the monitoring module is used for collecting the discharge voltage and the discharge temperature of the battery at regular time in each discharge cycle; and the control center is in communication connection with the monitoring module, evaluates the health state of the battery according to the data acquired by the monitoring module and outputs a result.
In the actual estimation process, in each discharge cycle, the discharge voltage and the discharge temperature of the battery are sampled every 2-3s until the capacity of the battery is reduced by 30% compared with the initial capacity; and the control center is in communication connection with the monitoring module, evaluates the health state of the battery according to the data acquired by the monitoring module and outputs a result, specifically, evaluates the health state of the battery by using the data acquired by the monitoring module as sample data through the evaluation method in the embodiment, and takes the health of the battery after the end of each cycle discharge as a label value. The system is driven based on data, and does not need to consider the complicated internal electrochemical reaction of the lithium ion battery during charging and discharging and construct an equivalent physical model of the battery; the least square supports the vector regression model, can effectively improve the solving efficiency while guaranteeing the precision, when being used for the estimation problem of the health state of the lithium ion battery, can realize the quick accurate estimation.
In another optional embodiment, the evaluation system further comprises: the charging module is used for charging the battery after each cycle of discharge is finished; and the discharging module discharges the battery according to constant current and discharges the battery circularly until the voltage of the battery is reduced to a rated value.
In some embodiments, during charging, the charging module charges the battery in a constant current mode of 1.5A until the battery voltage reaches 4.2V, and then continues charging in a constant voltage mode until the charging current drops to 20mA; the discharging module discharges the battery according to constant current, the battery discharges at constant current of 2A in the process of each discharging cycle until the voltage of the battery is reduced to a rated value, and the battery capacity after each discharging cycle is recorded, wherein the unit of the battery capacity is Ah.
In some embodiments, the battery is discharged 168 times, each discharge cycle reduces the battery capacity by 30% compared to the initial capacity, in estimating the state of health (SOH) of the battery, sample data sampled in each discharge cycle is added and averaged, and then the average discharge voltage u of each discharge cycle is obtained 1 (t) and average discharge temperature r 1 (t) as sample data; after each discharge cycle, the state of health of the battery is described by the battery capacity, and as a label value,
u is to be 1 (t)、r 1 (t) after normalization by the formula, u 2 (t)、r 2 (t) and S (t) are expressed as a sample set S = { [ u ] = { [ 2 (t),r 2 (t),s(t)]L T =1,2, \8230 |, T }, and the number of samples T =168, wherein [ u [/v ] 2 (t),r 2 (t),s(t)]Is a sample vector; the data of the first 100 cycles are selected from the whole sample set as a training set, and are denoted by S R ＝{[u 2 (t),r 2 (t),s(t)]I t =1,2, \ 8230;, 100}, and the remaining 68 samples were selected as the test set, denoted as S E ＝{[u 2 (t),r 2 (t),s(t)]|t＝1,2,…,68}。
Constructing a support vector regression model, and firstly, constructing a training set S in the calculation process R ＝{[u 2 (t),r 2 (t),s(t)]|t＝1,2,…,T SR Input characteristic signal x = { [ u ]) in 2 (t),r 2 (t)]|t＝1,2,…,T SR Inputting the result into equation (8), setting an initial kernel function parameter γ =4, and obtaining a kernel function matrix Ω; setting an initial penalty factor C =100, training set S R Lithium corresponding to middle input characteristic signal xThe true value of the state of health SOH of the ion battery is Y = { s (T) | T =1,2, \ 8230;, T SR }。
Substituting omega, C and Y into the formula (7) to obtain values of alpha and b, and obtaining a final regression function model
Test set S after normalization processing E ＝{[u 2 (t),r 2 (t),s(t)]|t＝1,2,…,T SE X (t) = [ u ] in 2 (t),r 2 (t)]Obtaining an estimated value of the SOH of the lithium ion battery as input data of the formula (9)
In the following, 5 samples are selected as input data to explain, training set S R ＝{[0.7649,0.4829,100],[0.8658,0.5263,99.4492],[0.8713,0.5151,99.4506],[0.8569,0.4780,99.4563],[0.8524,0.4481,99.3865]} inputting a characteristic signal x = { [0.7649,0.4829],[0.8658,0.5263],[0.8713,0.5151],[0.8569,0.4780],[0.8524,0.4481]Inputting the result into equation (8), setting an initial kernel function parameter γ =4, and obtaining a kernel function matrix:
setting an initial penalty factor C =100, training set S R The real value of the state of health SOH of the lithium ion battery corresponding to the medium input characteristic signal x is Y = {100,99.4492,99.4506,99.4563,99.3865}, and the omega, C and Y are substituted into the formula (7), so that α = [7.2552, -2.5725,1.6298, -0.3233, -5.9892]And b =99.6654, and substituting into the formula (9) to obtain a final regression function model.
In some embodiments, two sample data are selected as the test set S E ＝{[0.9435,0.4463,99.4526],[0.9484,0.4686,99.4121]} inputting the characteristic signal x = { [0.9435,0.4463 { [],[0.9484,0.4686]Inputting the parameters into a regression model (9) obtained by solving the parameters to obtain an estimated value of the SOH of the lithium ion battery
Performing parameter optimization on the estimation model of the least square support vector by adopting a particle swarm optimization according to the mean square error, and determining a least square support vector regression model after particle swarm optimization after obtaining an optimal parameter set; the least square support vector model optimized by the particle swarm optimization is used as a weak learner, the weak learners are integrated into a strong learner by adopting an Adaboost.R2 reinforcement learning method,
repeating the reinforcement learning process again to the obtained sample data to obtain more accurate final value of the health state of the lithium ion battery
The above description is only exemplary of the invention and should not be taken as limiting the invention, as any modification, equivalent replacement, or improvement made within the spirit and principle of the invention is intended to be covered by the appended claims.
Claims (10)
1. A lithium ion battery health assessment method is characterized in that,
acquiring a battery capacity value in an initial state of the battery, discharging the battery for multiple cycles, and reducing the voltage of the battery to a rated value in each discharge cycle;
in each discharge cycle, collecting the discharge voltage and the discharge temperature of the battery at regular time as sample data;
and establishing a calculation model, and inputting sample data to obtain an estimated value of the state of health of the battery.
2. The lithium-ion battery health assessment method of claim 1, further comprising:
performing parameter optimization on the calculation model to obtain a calculation model with an optimal parameter set, and enhancing the calculation model by utilizing reinforcement learning;
and inputting sample data into the calculation model after reinforcement learning to obtain the final value of the health state of the battery.
3. The lithium ion battery health assessment method of claim 2, wherein the battery capacity is used to describe the final value of the battery health status, and the battery health status after each discharge cycle is output as a label value.
4. The lithium-ion battery health assessment method according to claim 3, wherein the battery health status is obtained from a ratio of a battery capacity after any discharge to an initial battery capacity.
5. The lithium ion battery health assessment method according to claim 2, wherein an optimal parameter set is obtained by a particle swarm algorithm to perform parameter optimization on the calculation model.
6. The lithium ion battery health assessment method according to claim 1, wherein data normalization processing is performed on sample data input to the calculation model.
7. The lithium ion battery health assessment method according to claim 1, wherein a least squares support vector regression function is established as a calculation model according to a structure risk minimization principle, lagrangian multipliers are introduced, and variables are eliminated by adopting an equation transformation mode to establish an optimal solution matrix.
8. A lithium ion battery health assessment system, characterized in that the state of health of a battery is assessed by the assessment method of claims 1-7.
9. The lithium ion battery health assessment system of claim 8, wherein the assessment system comprises:
the monitoring module is used for collecting the discharge voltage and the discharge temperature of the battery at regular time in each discharge cycle;
and the control center is in communication connection with the monitoring module, evaluates the health state of the battery according to the data acquired by the monitoring module and outputs a result.
10. The lithium-ion battery health assessment system of claim 9, further comprising:
the charging module is used for charging the battery after each cycle of discharge is finished;
and the discharging module discharges the battery according to constant current and discharges the battery circularly until the voltage of the battery is reduced to a rated value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210253044.XA CN115656855A (en) | 2022-03-15 | 2022-03-15 | Lithium ion battery health assessment method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210253044.XA CN115656855A (en) | 2022-03-15 | 2022-03-15 | Lithium ion battery health assessment method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115656855A true CN115656855A (en) | 2023-01-31 |
Family
ID=85024123
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210253044.XA Pending CN115656855A (en) | 2022-03-15 | 2022-03-15 | Lithium ion battery health assessment method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115656855A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117129899A (en) * | 2023-08-31 | 2023-11-28 | 重庆跃达新能源有限公司 | Battery health state prediction management system and method |
| CN118275900A (en) * | 2024-06-03 | 2024-07-02 | 潍柴动力股份有限公司 | Battery health state estimation and modeling method, device, equipment and storage medium |
| WO2025015677A1 (en) * | 2023-07-14 | 2025-01-23 | 深圳先进储能材料国家工程研究中心有限公司 | Hybrid energy storage battery state monitoring method and system based on big data processing |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108445406A (en) * | 2018-03-13 | 2018-08-24 | 桂林电子科技大学 | A kind of power battery health status method of estimation |
| CN108805217A (en) * | 2018-06-20 | 2018-11-13 | 山东大学 | A kind of health state of lithium ion battery method of estimation and system based on support vector machines |
| CN110068774A (en) * | 2019-05-06 | 2019-07-30 | 清华四川能源互联网研究院 | Estimation method, device and the storage medium of lithium battery health status |
| CN111044928A (en) * | 2019-12-31 | 2020-04-21 | 福州大学 | A Lithium Battery State of Health Estimation Method |
| WO2020191800A1 (en) * | 2019-03-27 | 2020-10-01 | 东北大学 | Method for predicting remaining service life of lithium-ion battery employing wde-optimized lstm network |
| US20210055353A1 (en) * | 2017-12-07 | 2021-02-25 | Yazami Ip Pte. Ltd. | Method and system for online assessing state of health of a battery |
| CN113030764A (en) * | 2021-03-04 | 2021-06-25 | 武汉大学 | Battery pack health state estimation method and system |
| CN113419187A (en) * | 2021-06-08 | 2021-09-21 | 上海交通大学 | Lithium ion battery health estimation method |
-
2022
- 2022-03-15 CN CN202210253044.XA patent/CN115656855A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20210055353A1 (en) * | 2017-12-07 | 2021-02-25 | Yazami Ip Pte. Ltd. | Method and system for online assessing state of health of a battery |
| CN108445406A (en) * | 2018-03-13 | 2018-08-24 | 桂林电子科技大学 | A kind of power battery health status method of estimation |
| CN108805217A (en) * | 2018-06-20 | 2018-11-13 | 山东大学 | A kind of health state of lithium ion battery method of estimation and system based on support vector machines |
| WO2020191800A1 (en) * | 2019-03-27 | 2020-10-01 | 东北大学 | Method for predicting remaining service life of lithium-ion battery employing wde-optimized lstm network |
| CN110068774A (en) * | 2019-05-06 | 2019-07-30 | 清华四川能源互联网研究院 | Estimation method, device and the storage medium of lithium battery health status |
| CN111044928A (en) * | 2019-12-31 | 2020-04-21 | 福州大学 | A Lithium Battery State of Health Estimation Method |
| CN113030764A (en) * | 2021-03-04 | 2021-06-25 | 武汉大学 | Battery pack health state estimation method and system |
| CN113419187A (en) * | 2021-06-08 | 2021-09-21 | 上海交通大学 | Lithium ion battery health estimation method |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2025015677A1 (en) * | 2023-07-14 | 2025-01-23 | 深圳先进储能材料国家工程研究中心有限公司 | Hybrid energy storage battery state monitoring method and system based on big data processing |
| CN117129899A (en) * | 2023-08-31 | 2023-11-28 | 重庆跃达新能源有限公司 | Battery health state prediction management system and method |
| CN117129899B (en) * | 2023-08-31 | 2024-05-10 | 重庆跃达新能源有限公司 | Battery health state prediction management system and method |
| CN118275900A (en) * | 2024-06-03 | 2024-07-02 | 潍柴动力股份有限公司 | Battery health state estimation and modeling method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109031153B (en) | An online estimation method for the state of health of lithium-ion batteries | |
| WO2021259196A1 (en) | Battery pack consistency evaluation method and system | |
| Wen et al. | SOH prediction of lithium battery based on IC curve feature and BP neural network | |
| CN115656855A (en) | Lithium ion battery health assessment method and system | |
| CN110888059B (en) | Charge state estimation algorithm based on improved random forest combined volume Kalman | |
| CN113109717B (en) | Lithium battery state of charge estimation method based on characteristic curve optimization | |
| CN111948560A (en) | Lithium battery state of health estimation method based on multi-factor evaluation model | |
| CN110850298B (en) | Data-driven SOH estimation method and system for lithium batteries | |
| CN108519556A (en) | A Lithium-ion Battery SOC Prediction Method Based on Recurrent Neural Network | |
| Zhao et al. | State-of-charge estimation of lithium-ion battery: Joint long short-term memory network and adaptive extended Kalman filter online estimation algorithm | |
| CN112557907A (en) | SOC estimation method of electric vehicle lithium ion battery based on GRU-RNN | |
| CN113128672B (en) | Lithium ion battery pack SOH estimation method based on transfer learning algorithm | |
| CN112782594B (en) | A data-driven algorithm considering internal resistance to estimate the SOC of lithium batteries | |
| CN116643196A (en) | Battery health state estimation method integrating mechanism and data driving model | |
| CN113740735A (en) | Method for estimating SOC of lithium ion battery | |
| Sun et al. | Aging mechanism analysis and capacity estimation of lithium-ion battery pack based on electric vehicle charging data | |
| Hu et al. | State of health estimation for lithium-ion batteries with dynamic time warping and deep kernel learning model | |
| CN113820615A (en) | A kind of battery health detection method and device | |
| CN116626499A (en) | Lithium battery health state estimation method based on peak-pressing feature and improved LSTM | |
| Takyi-Aninakwa et al. | Deep learning framework designed for high-performance lithium-ion batteries state monitoring | |
| Chen et al. | A hybrid DNN-KF model for real-time SOC estimation of lithium-ion batteries under different ambient temperatures | |
| Zhang et al. | A comparative study on state-of-charge estimation for lithium-rich manganese-based battery based on Bayesian filtering and machine learning methods | |
| CN114397578B (en) | A method for estimating the remaining capacity of a lithium-ion battery | |
| Dou et al. | Short term charging data based battery state of health and state of charge estimation using feature pyramid | |
| CN119940112A (en) | A robust and adaptive lithium-ion battery state estimation method guided by deep reinforcement learning |
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 |